The RepRap Polymer Extruder Head
Version 1, 30th August 2005
Adrian Bowyer
University of Bath Department of Mechanical Engineering Technical Report 19/05
Summary
This document describes in detail how I made the first FDM extruder head for the RepRap project. It is intended to allow other researchers to reproduce the results. The head was designed to be made by the RepRap rapid prototyper itself, and thus consists of standard parts plus rapid-prototyped components. There were also a few parts that required other tools to make; the intent of the design was to make these parts as simple as possible, to minimise the tools needed, and to keep their number as small as possible. The head extruded a stream of Polymorph polymer about 0.5 mm in diameter at a temperature of 180 oC and a rate of up to 4 mms-1. The extruder is experimental, in that many parameters for it (particularly in the software) are user-set and controlled, whereas in a finished RepRap machine these would permenantly be set to defaults.
Contents:
1. Introduction
Clicking on all but the simplest photographs in this document will give a high-resolution version of each one.
Initially I have decided that the RepRap machine should use FDM rapid prototyping as its means of building objects. In the future, people (including me) will doubtless run the machine to make other RepRaps that use different RP technologies, in particular employing the wide range of techniques in powder sintering and gluing that are the basis of many machines. Indeed powder gluing may well be both simpler and faster than FDM, so the question immediately arises: why not start with that? There is one overriding reason: powder-gluing technology requires the use of an ink-jet print head to project the glue droplets onto the powder, and such heads would be very hard to make in the RepRap machine themselves. It may be argued that the RepRap machine can also not make the microcontroller chips that it uses, so what is the difference? The answer is that microcontrollers are widely available from multiple sources, whereas a given model of ink-jet print head is only available from one supplier, and such suppliers are increasingly including technology in their print heads to prevent multiple use.
I made some parts of the extuder head on a lathe, but it is also possible to make them using a much cheaper alternative. See How to turn on a lathe without one by Vik Olliver for details.
At many places in the design, parts made by rapid prototyping are held by conventional nuts and bolts of either M3 or M4 size. It is important to use washers both under the nuts and under the bolt heads at all such places. The washers spread the load from tightening, and also minimise damage to the RP plastic from the rotation of the nut or bolt head.
The extruder is experimental. In particular it is designed to be directly controllable by a person (or piece of software) sending it commands on a serial interface. In the final RepRap machine this control will almost certainly be exercised by another part of the RepRap machine itself via an I2C interface or a token ring.
Here is a picture of the finished FDM extruder head that this document describes.
2. High-temperature nozzle
Here is a close up of the heated nozzle (D in the picture above) with a 3 mm Polymorph polymer rod being fed into it.
To build the high-temperature nozzle, start with the brass tube. This is made from 6 mm diameter brass rod (all the brass parts of the RepRap machine can be made from this size of rod), but could alternatively be made from a length of M6 threaded brass studding, or from an M6 brass screw from which the head has been hacksawed off.
Here is a drawing of the brass tube as it should be when finished. (This drawing was created with the open-source version of QCad, and is in the Mechanics folder of the download.)
The brass part of the nozzle is a tube 50mm long and 6mm in diameter. The internal hole diameter is 3mm. The left-hand end is threaded M6 for a length of 15 mm. The right-hand end is a fine nozzle formed by a small diameter hole in the end. This is made by drilling the hole, turning a cone on the outside, and then drilling the 3mm hole down the middle from the left. The nozzle is formed by the conical end of the drill, the fine hole, and the external cone.
It was found to be very important to keep the length of the fine-diameter hole that forms the nozzle as short as possible. This is because the resistance to the flow of a viscous fluid (and molten polymer is highly viscous) is very dependent on the length of any tube through which it has to flow.
The diameter of the nozzle is a compromise: a large hole allows material to be deposited fast, but reduces the resolution of the machine; a small hole reduces material deposition rate but allows greater resolution. Most commercial machines use a diameter just under 0.5 mm. Of course, a RepRap machine could have two nozzles: one for coarse infill, the other for surfaces and fine detail. Note that the extrudate will end up slightly larger than the diameter choosen because of die swell (molten polymer is not a Newtonian fluid, it is viscoelastic; this means that when it is momentarily extruded through a fine nozzle it 'remembers' its former diameter and swells a bit larger than the nozzle diameter from which it has emerged). I have not listed tools in the parts section, as all of them are completely standard and widely available. However, fine drill bits are sometimes hard to find, so I have included suppliers of 0.5 mm (and other small) drills.
Start by turning the right-hand end of the rod square, then mark the end with a fine centre drill in the lathe's tailstock. If a fine centre drill is not available, a 1 mm drill projecting just a couple of mm from the tailstock chuck would work just as well. Make only a shallow dent. Next drill the fine nozzle hole using an 0.5 mm drill (other diameters could be used, both finer and coarser) to a depth of about 3 mm. This could be done in the lathe, but I found it easier to use a minidrill in a drill press, with the brass rod held vertically in a vice. Fine drills are very delicate - you need to use minimum force, to make sure you don't bend the drill, and to back off repeatedly to clear the swarf (i.e. use a woodpecker cycle).
Next turn a cone on the end of the rod to form the nozzle. The angle of the cone needs to be rather shallower than the cone created by the 3 mm drill bit that will be used to create the large hole, as in the diagram above. This allows the nozzle's fine-diameter hole to be as short as possible, while giving the greatest strength. When the cone has been turned, the nozzle hole may need to be cleaned again with the fine drill used to make it.
Next turn the left end of the rod flat, and mark a centre on that. Remove the rod from the lathe chuck and offer the 3 mm drill to be used to make the large hole up to the side of it. Position it so that the tip is about 0.5 mm inwards from the tip of the turned cone, then wrap some sticky tape round the other end of the drill to mark the length of the rod. Put the rod back in the lathe and drill it until the tape just touches the end of the rod. That way the hole will be as deep as possible without breaking through at the nozzle end. Once again, it may be necessary to clean the nozzle hole with the fine drill used to make it.
I found it worth wrecking one or two trial nozzles to end up with one where the fine nozzle hole was as short as possible to maximise flow.
If studding or a screw is being used to make the brass part of the nozzle (as opposed to plain rod), this paragraph can be ignored. Cut an M6 thread to a length of about 15mm on the rod on the left end. If a lathe is being used, a good way to get the M6 die square for this operation is to put the die in the lathe chuck squaring it up with a rule across the face of the chuck (unplug the power first...) and to put the rod in the tailstock. Then put the lathe in its fastest gear to make the chuck easy to turn by hand, and release the tailstock so that it will slide along the bed. Then push the rod against the die and - keeping a slight pressure on - rotate the chuck by hand to cut the thread. Rotate it anti-clockwise occasionally to clear the swarf.
Next make the PTFE tube. Here is a drawing of it (again this is in the Mechanics folder of the download).
The PTFE tube is 55 mm long and 10.5 mm in diameter. It has a 3 mm hole right through it, and the last 15 mm are drilled and tapped M6. A small cone is cut at the left-hand end so the polymer rod to be extruded can engage in the 3 mm hole more easily. You may find it easier to get 10 mm diameter rod than 10.5 mm; if so only one change needs to be made to the design - this is discussed in the section on the nozzle clamp below.
Cut a length of 10.5 mm diameter PTFE rod and turn the ends square to a length of 55 mm. Drill a 3mm hole right through, and then a larger hole (5 mm) to a depth of 15 mm. Tap this hole M6. (Again, line things up with the lathe: unplug the power, put the tap in the tailstock, set it free to slide, press it against the 5mm hole, and turn the chuck to cut the thread. Do half a turn anti-clockwise every so often to clear the swarf.)
Use a large drill bit to open out the left-hand end of the 3 mm hole into a cone by hand as shown above. This makes it easier for the end of the polymer rod to locate when it is fed into the hole.
Thoroughly clean the brass and PTFE parts. Any swarf that remains in them will get swept down by the flowing polymer and will block the nozzle. Screw the parts together tightly, but not so tightly that it strips the threads in the PTFE.
Now wind the heating coil round the brass. I used 0.2 mm diameter nichrome wire, but slightly different diameters would work just as well. Use a multimeter to measure off a 12 ohm length of the wire and cut it (it will be about 200 mm long).
Hold the PTFE tube in a vice with the brass rod projecting horizontally.
Then wrap a helix of PTFE tape round the brass rod so that the helix overlaps itself to give two thicknesses, one on top of the other. This tape is to electrically insulate the nichrome from the brass. If a threded rod is used the PTFE will pull into the valleys of the thread; this is okay.
Use a blob of Blu Tack to stick one end of the length of nichrome wire to the PTFE tube where the brass rod is screwed into it and wind the wire in a helix down the brass. If a threaded rod is used you can simply wind it in the valleys of the thread, otherwise you must try to get the pitch reasonably even. Stop about 5 mm from the nozzle end of the brass and wrap a piece of PTFE tape round the last few turns to secure them. Wrap another helix of PTFE tape round the nichrome coil, again overlapping to get two thicknesses. Lead the free bottom end of the nichrome wire straight back up the outside of this to the top of the brass, and wrap more PTFE tape round that to secure it. You should now have something that looks like the photograph above, but without the ribbon cable and without the thermistor wires on the left.
Next, the thermistor. I used the glass-bead thermistor in the Parts List below. It had a beta value of 3500, a room-temperature resistance of about 6 kilohms, and a resistance of about 100 ohms at 170 oC. You could use a different one if need be (there's a spreadsheet in the Electronics section below to help with the calculations), but it needs to be rated up to 250 oC. Carefully wrap the thermistor's leads in PTFE tape to insulate them, then placed the thermistor against the bottom of the brass in the 5 mm gap left below the nichrome coil. Wrap yet more PTFE tape round it and its leads to hold them and to insulate the device. I found that one doesn't need too much insulation - the nichrome is going to have about an amp put through it, and 12 watts will heat things pretty quickly - but the more one has, the more efficient the device will be.
Solder a 150 mm length of four-way ribbon cable to the thermistor leads, using fine heat-shrink sleeving to insulate the join. Soldering to the nichrome is a bit more problematic - solder doesn't wet nichrome... However I found that wrapping both the nichrome and the bared end of the ribbon cable round a single piece of thicker wire (a cut-off resistor lead worked well) would allow solder to be run over the lot (use a little extra flux if it is reluctant to do so) and made a good solid connection. Once again, use heat shrink to insulate them both, but leave all the unsoldered nichrome above the coil in free air - it gets hot.
Attach the ribbon cable to the PTFE tube with a cable tie, taking care not to attach it too near the hot end...
Solder connectors onto the free ends of the ribbon cable. A good cheap way to do this is to get a pack of crimp connectors for unshrouded shells like this:
They are about 12 mm long and 2 mm across. Solder them
to the
ends of the wires, then put heatshrink sleve flush with the open end
round
them to insulate them. You finish up with each wire having a neat
small individual connector on the end. The connectors are listed
in the Parts Section below.
Finally - measuring from the connectors you've just created - check that the resistance of the heater is still about 12 ohms (that is, check that it's not shorting on the brass tube), and check that the resistance of the thermistor is its room-temperature value.
3. Drive mechanism
I had hoped to do the CAD designs for the RP parts using the open-source program Art of Illusion, but at the time of designing there were one or two improvements that needed to be made to that package. I shall return to it and re-design the parts using it as soon as possible, and post the results on this site. In the meantime I did the designs using the academic version of the commercial Solid Edge program. All the Solid Edge .par files and their corresponding rapid-prototyping STL files for these parts are in the Mechanics directory of the download.
Here is the stackable pinch-wheel holder:
It should be possible to make the part in any RP machine from the STL file, including, of course, the RepRap machine itself when its design is finished.
Here is the motor mounting and the worm drive:
The photograph on the right shows the worm drive fitted. This will be described below.
Here is the nozzle clamp:
The large circular hole is where the PTFE rod fits. As distributed, this is set for a rod 10.5 mm in diameter. If you have a different sized PTFE rod this is the only dimension you need to change.
Here is a picture of the parts of the drive mechanism with two of the four stacking pinch wheels:
When making the RP parts, they were formed with the 'up' direction as shown in this photograph to get maximum strength (the graph paper was a flat horizontal surface).
The nozzle clamp is part A. The clamp mechanism is completed by the 25 mm M4 screw, two washers, and nut shown at B.
Parts C through H are the components of the motor and worm drive holder. D is the worm gear, and beside it are two M3 washers that will be fitted at either end of it. The two parts labelled E are brass bearing bushes; they are 6 mm long, have a 3mm hole down their centre, and are 6 mm in diameter (so they can be made from the same brass rod used for the heater nozzle; alternatively just cut two 6 mm lengths from a brass tube with 3 mm i.d. and 6 mm o.d.). Part F is an ordinary 35 mm M3 screw with a flat filed on it (the silvery section towards the left end) to take the grub screw in the worm gear. This will drive the gear. It, in turn is driven by part G, which is a 15 mm length cut from an Allen key that fits the head of the M3 screw. This fits into the collar (H) with a grub screw at either end (again made from 6 mm brass bar). The collar connects to the motor (not shown in this picture).
The device was held together with three lengths of M4 studding (I) and six M4 nuts and washers. The lengths of these pieces of studding are (26 + N × 14.14 + Z) mm, where N is the number of stacking pinch wheels (two in the picture immediately above; four in the device I made) and Z is the number of mm needed to accommodate whatever holder or bracket the extruder is to be attached to.
Two of the pinch wheel assemblies are shown at J. The bottom one has the drive showing that engages with the worm gear.
3.1 The Stacking Pinch Wheel Assemblies
Each pinch wheel assembly is identical, with the exception of the top one, which has an extra gear wheel that drives the whole machine.
The rapid-prototyped part is A. Part B is one of three 20 mm long brass bearing bushes with a 4 mm internal diameter and a 6 mm external diameter (the other two are shown fitted in the close up of the front of the device on the right). Three steel spur gears (C) form the drive, together with the three 40 mm M4 Allen bolts E. Two of the bolt heads make the actual pinch wheels; as they are knurled on the outside this gives a good grip. The M4 washers (D) go between the gear wheels and the brass bushes.
The best source I have found for the gears (see Parts List and Suppliers below) is a very helpful company, but their gears have a through hole of 3 mm diameter and no grub screws. You therefore need to drill their gears out to 4mm and to drill and tap the gears' bosses M3 for grub screws. These gears are completely standard (10 mm pitch-circle diameter, 20-tooth, 0.5 module steel gears) so you may well be able to find alternative suppliers that have a 4 mm internal diameter and a ready-fitted grub screw. The design would work equally well with brass gears, and probably the harder plastic ones too, though fully-plastic gears (as opposed to plastic gears with brass hubs) cannot hold a grub screw very strongly.
Start by push-fitting the two internal bushes into the RP part as shown on the right. They should be pushed in until they are just flush with the bottom of the cavities in which the pinch-wheel screw caps will fit.
Next cut two of the M4 screws to a shaft length (i.e. ignoring the head) of 30 mm (if you can get 30 mm M4 bolts with a 20 mm unthreaded top section this would obviously be unnecessary) and file flats on their ends for the grub screws in the gears.
Then put a light film of grease on the screws and gears and assemble the device as shown above. It is important not to use too much grease; in particular it must not get on the screw heads that form the pinch wheels and hence onto the Polymorph. It is also important to keep the gears clean; particles of swarf or RP-polymer in the teeth will cause them to jam.
As can be seen, I used two M4 washers between the top gear and the back bush, and one each between the pinch-wheel gears and their bushes.
The top gear comes in two forms:
The gear on the left is the general one that is fitted between all but the top two stacked pinch wheels. The one on the right has the worm wheel that meshes with the worm gear on its right-hand end, and it goes with the top pinch wheel pair. As with the other gears, the worm wheel needs to be drilled out to 4 mm internal diameter and to have a grub screw fitted.
The general (left) gear is an idler. In fact it doesn't need the grub screw that you can see, and could be greased and left to turn on the M4 screw to which it is fitted. Note the two lock-nuts on the right hand end that hold the entire assembly together. These need to be placed so that everything can rotate freely.
The drive gear (right) needs to be fixed with its grub screw (and the grub screw on the gear that engages with the worm). Therefore file two flats on the M4 screw to facilitate this.
Finally cut a 45 mm length of M3 studding and fit it with nuts and washers between the posts on the front of the device at A above. Tightening the M3 nuts causes the device to bend at B, and moves the two pinch wheels together. This is how the force on the Polymorph rod is generated.
Design Improvement (20 October 2005)
Make the M3 studding 12 mm longer and add in a 12 mm cut length of plastic tubing (I used 6 mm diameter tube intended for bubbling air in aquariums). The tube acts as a stiff spring. Here's a picture:
This has the effect of adding compliance to the pinch force applied to the rod, so that - if the rod has slight irregularities - the pinch wheels can move in and out a small distance to accommodate them. The plastic tube seems to work perfectly well, but if you want to use real springs there are some listed in the parts section below.
3.2 The Worm Drive and Motor Mount
The rapid-prototyped part (A) should be manufactured in the orientation shown. This gives it the maximum bending strength for its drive column.
Start by making the two bushes B. These are 6 mm long, with an internal diameter of 3 mm, and - like all other such parts - can be made from 6 mm rod. The bushes push-fit into the RP part, one at the inside of the left-hand cavity, and the other in a through-hole at A. I used cyanoacrylate (super glue) to fix the second bush as - because it is free at both ends - it had a tendency to slip. Be careful not to get the glue on the inside of the bush.
Then file a flat for the worm grub screw on a 35 mm M3 Allen-headed bolt (E), then lightly greased the inside of the bushes, the faces of the washers, and the worm gear.
The worm gear (C) has the right internal diameter (3 mm) this time, but it still needs the addition of a grub screw in its plain end.
Assemble an M3 washer (D), the worm (C), and another washer (D) in the left-hand cavity, then put the M3 bolt through with its head on the right. By rotating the bolt using an Allen key inserted from the right-hand end of the device and adjusting the grub screw on the worm, it is possible to feel when the grub screw is lined up with the flat on the bolt, and then to tighten it. Tighten it in a position so that the left-hand end of the M3 bolt is just clear of the bottom of the blind hole in the left-hand end of the RP part. The worm and the left-hand washer should bear on the top of the left-hand bush.
Next cut a 15 mm length from the Allen key (F) and file a flat on it for about half its length. Then turn a 16 mm long collar (G) from the 6 mm diameter brass rod. Drill this right through with a drill the same diameter as the distance across the diametrically opposite corners of the Allen key, and then drill the right-hand end out to 4 mm internal diameter for a depth of 8 mm to accomodate the shaft of the motor. Drill two side holes, tap them M3, and fit grub screws.
Then clamp the motor's shaft in a vice and file a flat on that, slide the collar G onto it, and slide the cut Allen key F into the other end. Put a small dab of grease in the Allen hole in the top of the M3 bolt - this is a simple way to make a self-aligning drive. Then put the motor on the right-hand end of the RP part and use a pair of twezers to slide parts F and G until the cut Allen key engaged with the M3 bolt head. Tighten the grub screws on the collar G. Then secure the motor with four 15 mm M3 screws, four M3 nuts and eight M3 washers.
Solder 120-mm-long twin wires onto the motor terminals, cover them with heat shrink, and make two connectors on the free ends of the wires as described at the end of the high-temperature nozzle section above.
3.3 Final Assembly of the Drive
Put the three lengths of M4 studding through the holes in the worm gear holder, put three M4 washers on them, and thread on three M4 nuts. Slide the first stacking pinch-wheel holder over the studding, then add the drive bolt with the worm wheel.
Then add the next pinch-wheel holder to the stack. I found that I had to pack the outside of the brass bush round the drive bolt to get it to locate firmly in the two half-cylindrical holes in the two stacking components. A small piece of folded paper worked as a perfectly good shim for this.
Carry on adding idler bolts and gears and stacking pinch-wheel holders until you have located all four. Two fewer idler gears and bolts are needed than the total number of stacking pinch-wheel holders. (That is, if there are four stacking pinch-wheel holders one needs two idlers plus one drive gear and bolt.)
Finally add the nozzle clamp, put three M4 washers onto the free ends of the M4 studding, and thread three M4 nuts on finger tight. I found that slight distortions in the RP parts meant that the worm gear engaged too hard when the nuts were tightened. This was easily fixed with more paper shim between the top pinch-wheel holder and the worm and motor holder. See the top of this section for pictures of how the final assembly should look.
Here is a top-view template for the centres of the M4 studding holes (radius 2 mm) to allow the whole device to be attached to other parts of the RepRap machine:
4. Electronics
Ultimately, the RepRap machine will be building its own circuitry automatically - you'll just have to plug in the chips. But, while waiting for that, I made the circuit for driving the extrusion head conventionally on 2.54 mm stripboard.
The circuit was designed with the free (but not open source) Eagle PCB designer. Eagle includes a neat routing tool for automatically creating PCBs if you want to etch one, rather than building on stripboard. If you do that, put two 3 mm mounting holes on one edge of the PCB 32 mm apart. All the files for this circuit are in the Electronics directory of the download.
The circuit is based round the PIC 16F628 microcontroller from Microchip Inc.
AN0 is an analogue input. It measures the voltage at the top of R3, which forms a potential divider with the thermistor which is plugged into pins 1 and 2 of of the pin-strip PCB header SV1. C3 smooths the signal. As the temperature of the brass extruder increases the thermistor's resistance falls, so the voltage on AN0 increases.
RA1 is an output. When RA1 is a logical 1 it drives the base of Q1 via R1, turning it on. Q1 is a darlington pair, and so has very high current gain. This means that a quite modest base-current of a couple of milliamps is enough to saturate it. The heater coil is connected across pins 3 and 4 of the header SV1, so - when on - the darlington grounds it, putting almost 12v across it. As the resistance of the heater is about 12 ohms, this gives a current of about 1 amp through the heater and a power output of about 12 watts. LED1 and its current-limiting resistor R2 are in parallel with the heater, so the LED glows whenever the PIC turns the heater on. Q1 is always fully on or fully off, and so doesn't dissipate a lot of heat. However it is worth mounting it on a small heatsink.
RB3 and RB4 are outputs connected to the H-bridge motor controller IC3. The motor itself is connected across pins 5 and 6 of the pin-strip header SV1. When RB3 is 1 and RB4 is 0 the motor turns anti-clockwise; when RB3 is 0 and RB4 is 1 the motor turns clockwise. When both outputs are 0 the motor is isolated and stops. The PIC software drives RB3 and RB4 with a variable mark-space square wave to control the motor's speed in both directions using pulse-width modulation. IC3 contains shorting diodes to damp the motor's back EMF when its power is cut.
LED2 is driven by software. It is on continuously when the extruder is powerd up but idle, and it flashes when the motor is turning. It doesn't have to be blue, but make the two LEDs different colours so you can tell what's going on more easily.
The pin-strip PCB header SV2 is the power and data connector. The circuit needs a single 12v supply on pin 4 (regulated down by IC2 and its smothing capacitors to 5v for the PIC). Serial data comes in on pin 3 and is transmitted on pin 2. Pin 1 is ground. The serial data is at TTL levels, so if you want to convert it to RS-232 (for example to plug it into a serial port on a PC or a USB-to-serial converter) you need to add a voltage converter chip like the MAX233. Put this on a separate circuit board - the final RepRap machine will do all its internal communication using TTL levels, so ultimately the MAX233 will not be connected to this device.
Header SV3 is an alternative method of control. The software is written so you can turn the extruder on by menu command or by grounding pin 1 of SV3. In the latter case the extruder can act as a dumb device that can be turned on or off with just one logic signal.
Use two 3 mm self-tapping screws to mount the circuit on the spacers sticking out of the side of the nozzle clamp. The location holes in the spacers are 32 mm apart.
The voltage at the top of R3 changes non-linearly with temperature, so I have written a spreadsheet to simplify the calculations. It is in the Electronics directory of the download. There are two files: one in .sxc format for OpenOffice (free), and one in .xls format for Excel (not...).
You fill in the five values top-left. The beta value should come from the thermistor's datasheet, and T0 and R0 are the thermistor's room-temperature resistance, which you should be able to measure with a thermometer and a multimeter. The series R value is the value of R3. A good digital multimeter will give you a better value for this than the colour code... The rail voltage is the voltage driving the circuit - in this case 5v from IC2. The T column is temperature in oC, and the R column is the corresponding resistance of the thermistor. The voltage column is the voltage at the top of R3 measured by the PIC. The two columns on the right are the numbers the PIC's analogue to digital converter generates. The PIC has two reference voltage settings - low and high - hence the two columns. You can select which you want at run time (see below). The A-to-D is only four bits (we're not talking precision here), so values are between 0 and 15 (other values in the spreadsheet are spurious - anything below 0 comes out 0; anything above 15 on the high scale is too hot, but it'll stick at 15...). This coarse measurement is, however, good enough to control the extruder's temperature to within 10 oC or so, which is quite adequate.
A different thermistor to the one I used should be easy to accomodate simply by changing R3 so that the A-to-D values around 170 oC come out to about the same numbers. Remember to get a thermistor that is rated up to at least 250 oC.
5. Software
The software for the extruder is stored in the directory called Software in the download.
The software is for the PIC and is written in C. It is intended to be compiled using SDCC, and GPUTILS which produce a file called extruder.hex that contains all the information that the PIC needs. How you load extruder.hex into the PIC depends on what sort of PIC programmer you have. If your programmer is compatible with the Picstart Plus programmer from Microchip themselves you can use gpicp. I used SDCC version #1055 (Jul 4 2005).
The program (in src/extruder.c) is divided into sections.
The first section saves the set temperature and speed into the PIC's EEPROM, and restores them on power-up. This means that the PIC remembers them even when it is powerd off then on again. The idea is that you can experiment with values by operating the extruder interactively via its serial interface, and then - once you have got the values right - use it in dumb mode where it is turned on (or off) by grounding (or not) pin 1 of SV3.
The second section contains functions that handle communications with the outside world via the PIC's serial interface. If you have the extruder plugged into a PC (via a level shifter such as the MAX233) you can talk and listen to it with programs like minicom (Unix/Linux) or Hyperterminal (Microsoft Windows). The baud rate needs to be set to 2400. Details of the commands you can give and their meaning are given in the Operation section below.
The third section contains functions that control the heater, the motor, and the blue LED.
The final section contains the initialization function and the main program. The main program runs an infinite loop listening for commands and controlling the peripherals by repeated function calls - i.e. crude process timesharing without interrupts. Note that menu items that need further input stop the loop until that input is complete, so once you have typed such a menu-item command letter don't sit scratching your head thinking what hex value to type when the heater is on, for example...
There is a short header file in include/extruder.h that defines a few useful items and some default values.
Under Linux and other Unix variants, you simply type "make" in the Software directory and - if SDCC and the other software are installed correctly - the program should compile and link to produce the extruder.hex file.
There is also a copy of the extruder.hex file in the directory Software/Saved-hex-file, in case you have trouble with the compiler.
At some point I'll publish instructions here for getting everything working under Microsoft Windows and on Macs (if you beat me to it, let me have details...)
6. Material preparation
The extruder needs a supply of 3 mm diameter polymorph rods as input. If you can buy these cheaply, so much the better. But it is quite simple to make your own from Polymorph granules - the usual form in which the polymer is supplied.
To make your own rods first drill a 3 mm hole in - well, just about anything rigid; a scrap piece of aluminium about 10 mm thick is good. You will use this to pass the finished rod through to test that it is not too fat.
Next gather together the following:
Keep everything from here on very clean. Any bits of dust and dirt will become incorporated in the rod you make, and will then clog the extruder nozzle.
Measure out about 3 g of polymorph granules. If you can't weigh 3 g, then this is what 3 g of Polymorph looks like:
The individual granules in the picture are about 3 or 4 mm across.
Bend the wire to form a right angle at the (insulation-stripped) end about 15 mm long. You will use this to hook the Polymorph out of the boiling water in the beaker.
Boil about 1 litre of water in the kettle, put the granules in the beaker, and pour boiling water over them. Fill the beaker about two-thirds full.
The Polymorph will melt and coagulate. Move the resulting lump about underwater with the hook to gather up any stray granules, then pull the lump out.
Using your fingers, squeeze out the hot water, and make the lump into a sausage about 50 mm long and 10 mm in diameter like this:
OBVIOUS WARNING: getting boiling water on your fingers from the Polymorph lump will scald you. Don't do it. Wait till things have cooled a little.
Now roll the Polymorph between the glass sheet and a flat surface such as a table top. When the Polymorph is hot it will squash under the glass very easily, and you will only need the lightest pressure. As it cools it will become more solid and need a greater pressure. The glass lets you see what you are doing.
ANOTHER OBVIOUS WARNING: If you push down too hard on the glass it will break, cut your hand, and spoil the rod you are making with a big red stain. Don't do that either.
As the rod cools it will reach a point where it cannot be worked any further. It will probably look something like this:
Bend it, put it back in the beaker, reboil the kettle, and add new hot water. Take care not to let the Polymorph ends touch each other in the beaker as they will stick together. When the Polymorph is soft again take it out with the hook and repeat the rolling process. You may have to do this several times. When the diameter gets near to 3 mm, put the reference 3 mm rod (B) under the glass like this:
The Polymorph rod is labeled A. You will find that it naturally tends to roll next to the reference rod. With the fact that the glass slopes very slightly towards its bottom edge against the table, this gives a Polymorph rod slightly smaller in diameter than the reference rod (it will end up about 2.9 mm in diameter) - this is exactly what you want. (The reference rod in the picture is not bent - that's a distortion created by the wide-angle camera lens...)
Test the rod when it is cold by putting it through your reference 3 mm hole. It should not be lose, but neither should it jam. If it jams, reheat it and roll it down a bit more. If it's too lose, you've gone too far...
7. Operation
Place the rod between the pinch-wheels in the extruder head and push it down into the PTFE and brass tubes. It should move smoothly but tightly all the way to the bottom without jamming.
Tighten the pinch-wheel-closing M3 nuts. I found that fingernail-tight plus half-a-turn gave about the right pressure, but your fingernails may be different to mine when it comes to using them as a torque wrench. Experiment.
If everything is working properly and the PC serial interface is connected, when you turn the 12v power on to the electronics you should get a prompt like this:
RepRap
>
Note that when the PIC is first turned on after programming it the set temperature and speed are not defined. This means that the heater will probably turn on. Turn it off immediately first time by typing "T00" (without the quotes) at the prompt. This will put the whole machine in idle mode.
The PIC is then in its main control loop thermostatically setting the temperature and switching on or off the motor and the blue LED. After you've typed "T00" the heater should be off (i.e. the red LED should be off) as the default temperature setting is 0. The motor should also be off, and the blue LED should be on continuously, all of which indicate that the device is idle and awaiting your instructions.
At the "> " prompt you can type a variety of single-letter commands, some of which you need to follow with a single hexadecimal number. Such numbers need to be given as two characters (i.e. they represent one byte), with the letters in lower case. Thus "07" (without the quotes) is 7 and "a5" is 165 and so on. When the PIC gives you a hexadecimal number it preceeds it with 0x. You don't need to do this when you type one.
Menu items that need further input stop the main loop until that input is complete, so, as I said above, don't sit scratching your head thinking what to type when the heater is on, for example... Updates to the software will be interrupt-driven soon.
When the PIC is turned on subsequent to the first time it will recover the temperature and speed setting that you last used before you last turned it off.
The software contains an interlock that stops the motor turning unless the temperature is up to the set temperature minus 1, whatever command you type. Generally you don't want the motor running in either direction unless the Polymorph in the brass tube is melted, as it will be jammed. But (especially when you are setting things up) it is sometimes useful to be able to turn the motor on when the device is cold and there is no rod in it at all. This you can do when the set temperature is 0, as that condition will be satisfied when the device is cold.
If you want to run the motor when there's a Polymorph rod in the machine without moving the rod, simply slaken off all the pinch-wheel-tightening nuts. The pinch wheels then won't drive the rod.
Note also that if you ask the motor to turn on and the device is not up to temperature, the motor will come on later automatically when the set temperature is reached. So don't turn the motor on, see nothing happening, then go for coffee...
As explained in the Electronics section above, the analogue-to-digital convereter in the PIC software is fairly crude, with only 16 possible values from 00 to 0f. The range of temperatures that can be measured and controlled can be further extended by selecting either the low or the high range for the PIC's internal voltage reference. See the spreadsheet for the actual temperatures that the values represent.
Run through the lower-case commands in the table. These report back values rather than changing things.
Then set the temperature to about 170oC (if you use the thermistor I used that's "H" followed by "T0c"; get the numbers from the spreadsheet for a different thermistor). Set the speed to "S80".
When you first turn the extruder on and set a temperature, the brass tube will come up to that temperature quite quickly. Nonetheless, don't start the motor right away - give the machine a minute or so to stabilise first.
Then set the motor running (the "F" command) and check that it is going in the right direction. If it isn't, stop it with the "m" command and swap the motor leads over.
The Polymorph should extrude well at temperatures around 170oC. Broadly speaking, the hotter it gets, the easier it will extrude.
The machine should extrude polymer whenever the motor is on. There are a couple of seconds delay between the motor starting and stopping and the flow starting and stopping. These are pretty consistent, and so should be easy to allow for in the RepRap driving software.
If you set the motor speed too fast, the pinch wheels will start slipping. You can tighten them a bit more, of course, but there is a limit to how far you want to go as the greater the stresses the machine is working under, the less reliable it will be long-term. Increasing the temperature will also speed up the flow, of course.
Once you have the device working, you can switch it off, disconnect the PC and RS232, and just drive it by sending 0 or 1 to SV3 pin 1. Or you can continue to use the serial interface of course. The effects of the "F" command and a 0 on SV3 pin 1 are effectively ORed by the software; either will turn the device on and keep it on. Note that if you turn the device on with the "F" or "R" commands it will continue to run until you type "m" regardless of the state of SV3 pin 1. Similarly, a 0 on SV3 pin 1 will keep the device on even if you type "m".
8. Bugs, improvements, and future work
Doubtless this section will get longer with time...
You will notice that the RP pinch-wheel holder has an unused through-hole at the back opposite one of the gear wheels. This is intended to takje a shaft out the back from the top pinch-wheel (from which the gear would have been removed). This shaft would drive a toothed wheel rotating in a slotted opto-switch. The resulting signal would be fed into the PIC. As the polymer rod was pulled between the wheels it would cause the un-geared wheel to rotate, driving the toothed wheel. When this stopped it would be an indication that the rod had run out and that it was time to fetch another. This will happen automatically in the final RepRap machine. A slowing of the train of opto-switch pulses would also indicate that the machine was jammed and that the other pinch wheels were slipping on the rod. However, at the moment the through-hole is obscured by the worm drive - design change needed...
Serial input should be interrupt driven.
9. Parts list and sources
If you find a good supplier for a part whom I haven't listed, please let me know. Please give me the URL for the supplier and their part number if you do, though.
Numbers in brackets after links to suppliers are the supplier's part number for the part.
The rapid-prototyped parts obviously have no supplier. The common parts, such as M4 cap screws, can also be obtained at any good hardware or DIY shop, or online engineering suppliers.
Note that many suppliers will only supply much larger quantities than you need of such things as ribbon cable (for which one supplier's minimum is 30 m when all you need is 200 mm...).
9.1 Mechanics
9.2 Electronics
9.3 Material
Polymorph granules can be obtained from the following suppliers:
UK:
USA (Polymorph is called Friendly Plastic in the USA):
10. Design and software downloads
To download this webpage and all the associated design files in gzipped .tar format as a single file (20 MB), click here: extruder.tgz.
On Unix/Linux run the file through gunzip and then tar.
With GNU tar these steps can be conflated:
Under Microsoft Windows, Winzip will unpack the archive.
The result will be a single directory called extruder which contains:
The Electronics directory contains the circuit diagram and the thermistor calculation spreadsheet.
The LICENCE file is the GNU Public Licence.
The Mechanics directory contains the CAD and STL files for the RP parts, and the DXF files for the drawings.
The Pix directory contains all the images used in this page.
The README file is, I hope, reasonably self explanatory.
The Software directory contains the program that needs to be downloaded into the PIC.
The extruder.html file is this page.
11. Licence
RepRap is copyright © 2005 University of Bath, the RepRap researchers (see the project's People page), and other contributors.
This web page and all the other files in its download are covered by that copyright.
Principal author:
Adrian Bowyer
Department of Mechanical Engineering
Faculty of Engineering and Design
University of Bath
Bath BA2 7AY
U.K.
e-mail: A.Bowyer@bath.ac.uk
RepRap is free; you can redistribute it and/or modify it under the terms of the GNU Library General Public Licence as published by the Free Software Foundation; either version 2 of the Licence, or (at your option) any later version.
RepRap is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Library General Public Licence for more details.
For this purpose the words "software" and "library" in the GNU Library General Public Licence are taken to mean any and all computer programs computer files data results documents and other copyright information available from the RepRap project.
You should have received a copy of the GNU Library General Public Licence along with RepRap; if not, write to the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA, or see
http://www.gnu.org/
12. Here it is working
I mounted the extruder on the old Cartesian robot described briefly the the RepRap Blog of 14 May 2005. I set it running to build a 20 mm x 10 mm x 2 mm rectangular block by outlining it and then alternately zig-zag filling it in layers. Here's a picture of the first attempt at making that:
The extruder speed was set to 0x70 and the temperature was 0x08 on the high scale. If you click on the image it links to a Quicktime movie of it working. N.B. This is 20 MB long, and so is not included in the download...
This document describes in detail how I made the first FDM extruder head for the RepRap project. It is intended to allow other researchers to reproduce the results. The head was designed to be made by the RepRap rapid prototyper itself, and thus consists of standard parts plus rapid-prototyped components. There were also a few parts that required other tools to make; the intent of the design was to make these parts as simple as possible, to minimise the tools needed, and to keep their number as small as possible. The head extruded a stream of Polymorph polymer about 0.5 mm in diameter at a temperature of 180 oC and a rate of up to 4 mms-1. The extruder is experimental, in that many parameters for it (particularly in the software) are user-set and controlled, whereas in a finished RepRap machine these would permenantly be set to defaults.
Contents:
1. Introduction
2. High-temperature nozzle
3. Drive mechanism
4. Electronics
5. Software
6. Material preparation
7. Operation
8. Bugs, improvements, and future work
9. Parts list and sources
10. Downloads
11. Licence
12. Here it is working
2. High-temperature nozzle
3. Drive mechanism
4. Electronics
5. Software
6. Material preparation
7. Operation
8. Bugs, improvements, and future work
9. Parts list and sources
10. Downloads
11. Licence
12. Here it is working
1. Introduction
Clicking on all but the simplest photographs in this document will give a high-resolution version of each one.
Initially I have decided that the RepRap machine should use FDM rapid prototyping as its means of building objects. In the future, people (including me) will doubtless run the machine to make other RepRaps that use different RP technologies, in particular employing the wide range of techniques in powder sintering and gluing that are the basis of many machines. Indeed powder gluing may well be both simpler and faster than FDM, so the question immediately arises: why not start with that? There is one overriding reason: powder-gluing technology requires the use of an ink-jet print head to project the glue droplets onto the powder, and such heads would be very hard to make in the RepRap machine themselves. It may be argued that the RepRap machine can also not make the microcontroller chips that it uses, so what is the difference? The answer is that microcontrollers are widely available from multiple sources, whereas a given model of ink-jet print head is only available from one supplier, and such suppliers are increasingly including technology in their print heads to prevent multiple use.
I made some parts of the extuder head on a lathe, but it is also possible to make them using a much cheaper alternative. See How to turn on a lathe without one by Vik Olliver for details.
At many places in the design, parts made by rapid prototyping are held by conventional nuts and bolts of either M3 or M4 size. It is important to use washers both under the nuts and under the bolt heads at all such places. The washers spread the load from tightening, and also minimise damage to the RP plastic from the rotation of the nut or bolt head.
The extruder is experimental. In particular it is designed to be directly controllable by a person (or piece of software) sending it commands on a serial interface. In the final RepRap machine this control will almost certainly be exercised by another part of the RepRap machine itself via an I2C interface or a token ring.
Here is a picture of the finished FDM extruder head that this document describes.
 |
The large grid squares on the image are 10 mm across. A 12v geared electric motor (A) drives a 3 mm diameter rod of the polymer to be extruded (B) by means of a stack of pinch-wheels (C) into a heated nozzle (D). The heated nozzle has a small hole in the bottom end out of which a stream of molten polymer emerges. The electronics for controlling the heater and the motor can be partly seen at (E). The polymer that the RepRap FDM extruder head uses is called Polymorph (see the Materials section below for suppliers). This has a low melting point of 62 oC, but the heater is capable of being controlled to temperatures up to 200 oC. At around 170 oC the Polymorph becomes inviscid enough to extrude easily. The pinch wheel stack (there are four sets of pinch wheels in it here) is designed to have as many or as few sets of pinch wheels as needed. The idea is that each pair of wheels forces the polymer rod downwards, so increasing the stacked number increases the total force. A single-wheel pair can only be tightened on the polymer rod being fed so much before they start to squash it permenantly, so the force they exert is limited by this and by the coefficient of friction between the wheels and the rod. The stack allows the force to be multiplied. The horizontal threaded rods you can see in front of each pair of pinch wheels are the means by which the pinch is tightened. OBVIOUS WARNING: the heated nozzle D gets very hot (almost as hot as a soldering iron). It is a bad idea to touch it or the wires leading to it when the machine is working or has just recently been switched off. LESS OBVIOUS WARNING: The motor is geared down to drive the pinch wheels very slowly. This means that, even though it is a fairly weedy motor, they are turned with a considerable torque. It is well-nigh impossible to get one's fingers in the works, but if one manages to do so, the motor will just keep going and injuries will occur. WARNING FOR NON-CHEMISTS: Nozzle D is partially made from PTFE. PTFE is stable up to 250 oC. Above that it breaks down into various gasses, some of which are quite toxic. The extruder should not be operated above 250 oC. |
2. High-temperature nozzle
Here is a close up of the heated nozzle (D in the picture above) with a 3 mm Polymorph polymer rod being fed into it.
 |
The 3 mm Polymorph rod to be extruded (A) enters a PTFE tube
(B). A smaller brass tube (C) is screwed into the bottom of the
PTFE tube. The bottom end of the brass tube forms a fine
nozzle. The brass tube is wrapped in PTFE tape, on top of which
is wound a coil of nichrome heater wire. The ends of the heater wire
can be seen at D. More PTFE tape is wound over the wire to keep
it in place and to act as a thermal insulator. A glass-bead thermistor (the slight bump on the bottom left of the brass tube C) is held against the brass tube by more PTFE tape; the wires from the thermistor can be seen at E. A short four-way ribbon cable is connected to both the heater and the thermistor and terminates in a connector at F. This connector interfaces with the control electronics. The ribbon cable is held to the PTFE tube by the green cable tie. The nichrome wire heats the brass (C), melting the polymer within it. The PTFE tube (B) guides the solid polymer rod (A) into the brass and insulates the rest of the machine from the high temperature of the brass. As the rod (A) is forced into the top of the PTFE it acts as a piston driving molten polymer out of the nozzle. The thermistor monitors the temperature of the brass and allows it to be thermostatically controlled. Note the bare flying nichrome wires just below D. These need to be kept clear of the rest of the assembly as shown. They get very hot when the heater is first turned on, and must be in the open air. |
To build the high-temperature nozzle, start with the brass tube. This is made from 6 mm diameter brass rod (all the brass parts of the RepRap machine can be made from this size of rod), but could alternatively be made from a length of M6 threaded brass studding, or from an M6 brass screw from which the head has been hacksawed off.
Here is a drawing of the brass tube as it should be when finished. (This drawing was created with the open-source version of QCad, and is in the Mechanics folder of the download.)
The brass part of the nozzle is a tube 50mm long and 6mm in diameter. The internal hole diameter is 3mm. The left-hand end is threaded M6 for a length of 15 mm. The right-hand end is a fine nozzle formed by a small diameter hole in the end. This is made by drilling the hole, turning a cone on the outside, and then drilling the 3mm hole down the middle from the left. The nozzle is formed by the conical end of the drill, the fine hole, and the external cone.
It was found to be very important to keep the length of the fine-diameter hole that forms the nozzle as short as possible. This is because the resistance to the flow of a viscous fluid (and molten polymer is highly viscous) is very dependent on the length of any tube through which it has to flow.
The diameter of the nozzle is a compromise: a large hole allows material to be deposited fast, but reduces the resolution of the machine; a small hole reduces material deposition rate but allows greater resolution. Most commercial machines use a diameter just under 0.5 mm. Of course, a RepRap machine could have two nozzles: one for coarse infill, the other for surfaces and fine detail. Note that the extrudate will end up slightly larger than the diameter choosen because of die swell (molten polymer is not a Newtonian fluid, it is viscoelastic; this means that when it is momentarily extruded through a fine nozzle it 'remembers' its former diameter and swells a bit larger than the nozzle diameter from which it has emerged). I have not listed tools in the parts section, as all of them are completely standard and widely available. However, fine drill bits are sometimes hard to find, so I have included suppliers of 0.5 mm (and other small) drills.
Start by turning the right-hand end of the rod square, then mark the end with a fine centre drill in the lathe's tailstock. If a fine centre drill is not available, a 1 mm drill projecting just a couple of mm from the tailstock chuck would work just as well. Make only a shallow dent. Next drill the fine nozzle hole using an 0.5 mm drill (other diameters could be used, both finer and coarser) to a depth of about 3 mm. This could be done in the lathe, but I found it easier to use a minidrill in a drill press, with the brass rod held vertically in a vice. Fine drills are very delicate - you need to use minimum force, to make sure you don't bend the drill, and to back off repeatedly to clear the swarf (i.e. use a woodpecker cycle).
Next turn a cone on the end of the rod to form the nozzle. The angle of the cone needs to be rather shallower than the cone created by the 3 mm drill bit that will be used to create the large hole, as in the diagram above. This allows the nozzle's fine-diameter hole to be as short as possible, while giving the greatest strength. When the cone has been turned, the nozzle hole may need to be cleaned again with the fine drill used to make it.
Next turn the left end of the rod flat, and mark a centre on that. Remove the rod from the lathe chuck and offer the 3 mm drill to be used to make the large hole up to the side of it. Position it so that the tip is about 0.5 mm inwards from the tip of the turned cone, then wrap some sticky tape round the other end of the drill to mark the length of the rod. Put the rod back in the lathe and drill it until the tape just touches the end of the rod. That way the hole will be as deep as possible without breaking through at the nozzle end. Once again, it may be necessary to clean the nozzle hole with the fine drill used to make it.
I found it worth wrecking one or two trial nozzles to end up with one where the fine nozzle hole was as short as possible to maximise flow.
If studding or a screw is being used to make the brass part of the nozzle (as opposed to plain rod), this paragraph can be ignored. Cut an M6 thread to a length of about 15mm on the rod on the left end. If a lathe is being used, a good way to get the M6 die square for this operation is to put the die in the lathe chuck squaring it up with a rule across the face of the chuck (unplug the power first...) and to put the rod in the tailstock. Then put the lathe in its fastest gear to make the chuck easy to turn by hand, and release the tailstock so that it will slide along the bed. Then push the rod against the die and - keeping a slight pressure on - rotate the chuck by hand to cut the thread. Rotate it anti-clockwise occasionally to clear the swarf.
Next make the PTFE tube. Here is a drawing of it (again this is in the Mechanics folder of the download).
The PTFE tube is 55 mm long and 10.5 mm in diameter. It has a 3 mm hole right through it, and the last 15 mm are drilled and tapped M6. A small cone is cut at the left-hand end so the polymer rod to be extruded can engage in the 3 mm hole more easily. You may find it easier to get 10 mm diameter rod than 10.5 mm; if so only one change needs to be made to the design - this is discussed in the section on the nozzle clamp below.
Cut a length of 10.5 mm diameter PTFE rod and turn the ends square to a length of 55 mm. Drill a 3mm hole right through, and then a larger hole (5 mm) to a depth of 15 mm. Tap this hole M6. (Again, line things up with the lathe: unplug the power, put the tap in the tailstock, set it free to slide, press it against the 5mm hole, and turn the chuck to cut the thread. Do half a turn anti-clockwise every so often to clear the swarf.)
Use a large drill bit to open out the left-hand end of the 3 mm hole into a cone by hand as shown above. This makes it easier for the end of the polymer rod to locate when it is fed into the hole.
Thoroughly clean the brass and PTFE parts. Any swarf that remains in them will get swept down by the flowing polymer and will block the nozzle. Screw the parts together tightly, but not so tightly that it strips the threads in the PTFE.
Now wind the heating coil round the brass. I used 0.2 mm diameter nichrome wire, but slightly different diameters would work just as well. Use a multimeter to measure off a 12 ohm length of the wire and cut it (it will be about 200 mm long).
Hold the PTFE tube in a vice with the brass rod projecting horizontally.
Then wrap a helix of PTFE tape round the brass rod so that the helix overlaps itself to give two thicknesses, one on top of the other. This tape is to electrically insulate the nichrome from the brass. If a threded rod is used the PTFE will pull into the valleys of the thread; this is okay.
Use a blob of Blu Tack to stick one end of the length of nichrome wire to the PTFE tube where the brass rod is screwed into it and wind the wire in a helix down the brass. If a threaded rod is used you can simply wind it in the valleys of the thread, otherwise you must try to get the pitch reasonably even. Stop about 5 mm from the nozzle end of the brass and wrap a piece of PTFE tape round the last few turns to secure them. Wrap another helix of PTFE tape round the nichrome coil, again overlapping to get two thicknesses. Lead the free bottom end of the nichrome wire straight back up the outside of this to the top of the brass, and wrap more PTFE tape round that to secure it. You should now have something that looks like the photograph above, but without the ribbon cable and without the thermistor wires on the left.
Next, the thermistor. I used the glass-bead thermistor in the Parts List below. It had a beta value of 3500, a room-temperature resistance of about 6 kilohms, and a resistance of about 100 ohms at 170 oC. You could use a different one if need be (there's a spreadsheet in the Electronics section below to help with the calculations), but it needs to be rated up to 250 oC. Carefully wrap the thermistor's leads in PTFE tape to insulate them, then placed the thermistor against the bottom of the brass in the 5 mm gap left below the nichrome coil. Wrap yet more PTFE tape round it and its leads to hold them and to insulate the device. I found that one doesn't need too much insulation - the nichrome is going to have about an amp put through it, and 12 watts will heat things pretty quickly - but the more one has, the more efficient the device will be.
Solder a 150 mm length of four-way ribbon cable to the thermistor leads, using fine heat-shrink sleeving to insulate the join. Soldering to the nichrome is a bit more problematic - solder doesn't wet nichrome... However I found that wrapping both the nichrome and the bared end of the ribbon cable round a single piece of thicker wire (a cut-off resistor lead worked well) would allow solder to be run over the lot (use a little extra flux if it is reluctant to do so) and made a good solid connection. Once again, use heat shrink to insulate them both, but leave all the unsoldered nichrome above the coil in free air - it gets hot.
Attach the ribbon cable to the PTFE tube with a cable tie, taking care not to attach it too near the hot end...
Solder connectors onto the free ends of the ribbon cable. A good cheap way to do this is to get a pack of crimp connectors for unshrouded shells like this:
Finally - measuring from the connectors you've just created - check that the resistance of the heater is still about 12 ohms (that is, check that it's not shorting on the brass tube), and check that the resistance of the thermistor is its room-temperature value.
3. Drive mechanism
 Back view of the drive
 Front view of the drive
|
Here are two views of the drive mechanism. A 12v
electric motor (A) with an integral gearbox (B) drives a worm gear
(C). This gear in turn drives a stack of pinch wheels (Ds) to
force the Polymorph rod into the high-temperature nozzle described
above; this is clamped at E. There are three different RP designs in the drive mechanism: 1. The motor and worm gear holder,
2. The stackable pinch-wheel drives (there were four in my version), and 3. The high-temperature nozzle clamp. The whole assembly is held together by three lengths of M4 studding, the ends of which can be seen at F. This studding can also be used to attach the whole device to the rest of the RepRap machine that will drive it around in 3D space to deposit material. It is possible to include as many pinch-wheel pairs as required. Four seemed to produce a reliable machine. In general, the more one has, the higher the force that can be generated to push the Polymorph rod into the nozzle. This is because the pinch wheels can only be done up so tight before they start permenantly to compress the Polymorph rod (especially when the machine is left standing for a long time unused). As Polymorph is quite slippery, the coefficient of friction between the pinch wheels and the Polymorph rod is not very high, and so an individual set of pinch wheels can exert a reasonable - but not a large - force. Stacking them up multiplies the force. |
I had hoped to do the CAD designs for the RP parts using the open-source program Art of Illusion, but at the time of designing there were one or two improvements that needed to be made to that package. I shall return to it and re-design the parts using it as soon as possible, and post the results on this site. In the meantime I did the designs using the academic version of the commercial Solid Edge program. All the Solid Edge .par files and their corresponding rapid-prototyping STL files for these parts are in the Mechanics directory of the download.
Here is the stackable pinch-wheel holder:
It should be possible to make the part in any RP machine from the STL file, including, of course, the RepRap machine itself when its design is finished.
Here is the motor mounting and the worm drive:
The photograph on the right shows the worm drive fitted. This will be described below.
Here is the nozzle clamp:
The large circular hole is where the PTFE rod fits. As distributed, this is set for a rod 10.5 mm in diameter. If you have a different sized PTFE rod this is the only dimension you need to change.
Here is a picture of the parts of the drive mechanism with two of the four stacking pinch wheels:
When making the RP parts, they were formed with the 'up' direction as shown in this photograph to get maximum strength (the graph paper was a flat horizontal surface).
The nozzle clamp is part A. The clamp mechanism is completed by the 25 mm M4 screw, two washers, and nut shown at B.
Parts C through H are the components of the motor and worm drive holder. D is the worm gear, and beside it are two M3 washers that will be fitted at either end of it. The two parts labelled E are brass bearing bushes; they are 6 mm long, have a 3mm hole down their centre, and are 6 mm in diameter (so they can be made from the same brass rod used for the heater nozzle; alternatively just cut two 6 mm lengths from a brass tube with 3 mm i.d. and 6 mm o.d.). Part F is an ordinary 35 mm M3 screw with a flat filed on it (the silvery section towards the left end) to take the grub screw in the worm gear. This will drive the gear. It, in turn is driven by part G, which is a 15 mm length cut from an Allen key that fits the head of the M3 screw. This fits into the collar (H) with a grub screw at either end (again made from 6 mm brass bar). The collar connects to the motor (not shown in this picture).
The device was held together with three lengths of M4 studding (I) and six M4 nuts and washers. The lengths of these pieces of studding are (26 + N × 14.14 + Z) mm, where N is the number of stacking pinch wheels (two in the picture immediately above; four in the device I made) and Z is the number of mm needed to accommodate whatever holder or bracket the extruder is to be attached to.
Two of the pinch wheel assemblies are shown at J. The bottom one has the drive showing that engages with the worm gear.
3.1 The Stacking Pinch Wheel Assemblies
Each pinch wheel assembly is identical, with the exception of the top one, which has an extra gear wheel that drives the whole machine.
The rapid-prototyped part is A. Part B is one of three 20 mm long brass bearing bushes with a 4 mm internal diameter and a 6 mm external diameter (the other two are shown fitted in the close up of the front of the device on the right). Three steel spur gears (C) form the drive, together with the three 40 mm M4 Allen bolts E. Two of the bolt heads make the actual pinch wheels; as they are knurled on the outside this gives a good grip. The M4 washers (D) go between the gear wheels and the brass bushes.
The best source I have found for the gears (see Parts List and Suppliers below) is a very helpful company, but their gears have a through hole of 3 mm diameter and no grub screws. You therefore need to drill their gears out to 4mm and to drill and tap the gears' bosses M3 for grub screws. These gears are completely standard (10 mm pitch-circle diameter, 20-tooth, 0.5 module steel gears) so you may well be able to find alternative suppliers that have a 4 mm internal diameter and a ready-fitted grub screw. The design would work equally well with brass gears, and probably the harder plastic ones too, though fully-plastic gears (as opposed to plastic gears with brass hubs) cannot hold a grub screw very strongly.
Start by push-fitting the two internal bushes into the RP part as shown on the right. They should be pushed in until they are just flush with the bottom of the cavities in which the pinch-wheel screw caps will fit.
Next cut two of the M4 screws to a shaft length (i.e. ignoring the head) of 30 mm (if you can get 30 mm M4 bolts with a 20 mm unthreaded top section this would obviously be unnecessary) and file flats on their ends for the grub screws in the gears.
Then put a light film of grease on the screws and gears and assemble the device as shown above. It is important not to use too much grease; in particular it must not get on the screw heads that form the pinch wheels and hence onto the Polymorph. It is also important to keep the gears clean; particles of swarf or RP-polymer in the teeth will cause them to jam.
As can be seen, I used two M4 washers between the top gear and the back bush, and one each between the pinch-wheel gears and their bushes.
The top gear comes in two forms:
The gear on the left is the general one that is fitted between all but the top two stacked pinch wheels. The one on the right has the worm wheel that meshes with the worm gear on its right-hand end, and it goes with the top pinch wheel pair. As with the other gears, the worm wheel needs to be drilled out to 4 mm internal diameter and to have a grub screw fitted.
The general (left) gear is an idler. In fact it doesn't need the grub screw that you can see, and could be greased and left to turn on the M4 screw to which it is fitted. Note the two lock-nuts on the right hand end that hold the entire assembly together. These need to be placed so that everything can rotate freely.
The drive gear (right) needs to be fixed with its grub screw (and the grub screw on the gear that engages with the worm). Therefore file two flats on the M4 screw to facilitate this.
Finally cut a 45 mm length of M3 studding and fit it with nuts and washers between the posts on the front of the device at A above. Tightening the M3 nuts causes the device to bend at B, and moves the two pinch wheels together. This is how the force on the Polymorph rod is generated.
Design Improvement (20 October 2005)
Make the M3 studding 12 mm longer and add in a 12 mm cut length of plastic tubing (I used 6 mm diameter tube intended for bubbling air in aquariums). The tube acts as a stiff spring. Here's a picture:
This has the effect of adding compliance to the pinch force applied to the rod, so that - if the rod has slight irregularities - the pinch wheels can move in and out a small distance to accommodate them. The plastic tube seems to work perfectly well, but if you want to use real springs there are some listed in the parts section below.
3.2 The Worm Drive and Motor Mount
The rapid-prototyped part (A) should be manufactured in the orientation shown. This gives it the maximum bending strength for its drive column.
Start by making the two bushes B. These are 6 mm long, with an internal diameter of 3 mm, and - like all other such parts - can be made from 6 mm rod. The bushes push-fit into the RP part, one at the inside of the left-hand cavity, and the other in a through-hole at A. I used cyanoacrylate (super glue) to fix the second bush as - because it is free at both ends - it had a tendency to slip. Be careful not to get the glue on the inside of the bush.
Then file a flat for the worm grub screw on a 35 mm M3 Allen-headed bolt (E), then lightly greased the inside of the bushes, the faces of the washers, and the worm gear.
The worm gear (C) has the right internal diameter (3 mm) this time, but it still needs the addition of a grub screw in its plain end.
Assemble an M3 washer (D), the worm (C), and another washer (D) in the left-hand cavity, then put the M3 bolt through with its head on the right. By rotating the bolt using an Allen key inserted from the right-hand end of the device and adjusting the grub screw on the worm, it is possible to feel when the grub screw is lined up with the flat on the bolt, and then to tighten it. Tighten it in a position so that the left-hand end of the M3 bolt is just clear of the bottom of the blind hole in the left-hand end of the RP part. The worm and the left-hand washer should bear on the top of the left-hand bush.
Next cut a 15 mm length from the Allen key (F) and file a flat on it for about half its length. Then turn a 16 mm long collar (G) from the 6 mm diameter brass rod. Drill this right through with a drill the same diameter as the distance across the diametrically opposite corners of the Allen key, and then drill the right-hand end out to 4 mm internal diameter for a depth of 8 mm to accomodate the shaft of the motor. Drill two side holes, tap them M3, and fit grub screws.
Then clamp the motor's shaft in a vice and file a flat on that, slide the collar G onto it, and slide the cut Allen key F into the other end. Put a small dab of grease in the Allen hole in the top of the M3 bolt - this is a simple way to make a self-aligning drive. Then put the motor on the right-hand end of the RP part and use a pair of twezers to slide parts F and G until the cut Allen key engaged with the M3 bolt head. Tighten the grub screws on the collar G. Then secure the motor with four 15 mm M3 screws, four M3 nuts and eight M3 washers.
Solder 120-mm-long twin wires onto the motor terminals, cover them with heat shrink, and make two connectors on the free ends of the wires as described at the end of the high-temperature nozzle section above.
3.3 Final Assembly of the Drive
Put the three lengths of M4 studding through the holes in the worm gear holder, put three M4 washers on them, and thread on three M4 nuts. Slide the first stacking pinch-wheel holder over the studding, then add the drive bolt with the worm wheel.
Then add the next pinch-wheel holder to the stack. I found that I had to pack the outside of the brass bush round the drive bolt to get it to locate firmly in the two half-cylindrical holes in the two stacking components. A small piece of folded paper worked as a perfectly good shim for this.
Carry on adding idler bolts and gears and stacking pinch-wheel holders until you have located all four. Two fewer idler gears and bolts are needed than the total number of stacking pinch-wheel holders. (That is, if there are four stacking pinch-wheel holders one needs two idlers plus one drive gear and bolt.)
Finally add the nozzle clamp, put three M4 washers onto the free ends of the M4 studding, and thread three M4 nuts on finger tight. I found that slight distortions in the RP parts meant that the worm gear engaged too hard when the nuts were tightened. This was easily fixed with more paper shim between the top pinch-wheel holder and the worm and motor holder. See the top of this section for pictures of how the final assembly should look.
Here is a top-view template for the centres of the M4 studding holes (radius 2 mm) to allow the whole device to be attached to other parts of the RepRap machine:
4. Electronics
Ultimately, the RepRap machine will be building its own circuitry automatically - you'll just have to plug in the chips. But, while waiting for that, I made the circuit for driving the extrusion head conventionally on 2.54 mm stripboard.
The circuit was designed with the free (but not open source) Eagle PCB designer. Eagle includes a neat routing tool for automatically creating PCBs if you want to etch one, rather than building on stripboard. If you do that, put two 3 mm mounting holes on one edge of the PCB 32 mm apart. All the files for this circuit are in the Electronics directory of the download.
The circuit is based round the PIC 16F628 microcontroller from Microchip Inc.
AN0 is an analogue input. It measures the voltage at the top of R3, which forms a potential divider with the thermistor which is plugged into pins 1 and 2 of of the pin-strip PCB header SV1. C3 smooths the signal. As the temperature of the brass extruder increases the thermistor's resistance falls, so the voltage on AN0 increases.
RA1 is an output. When RA1 is a logical 1 it drives the base of Q1 via R1, turning it on. Q1 is a darlington pair, and so has very high current gain. This means that a quite modest base-current of a couple of milliamps is enough to saturate it. The heater coil is connected across pins 3 and 4 of the header SV1, so - when on - the darlington grounds it, putting almost 12v across it. As the resistance of the heater is about 12 ohms, this gives a current of about 1 amp through the heater and a power output of about 12 watts. LED1 and its current-limiting resistor R2 are in parallel with the heater, so the LED glows whenever the PIC turns the heater on. Q1 is always fully on or fully off, and so doesn't dissipate a lot of heat. However it is worth mounting it on a small heatsink.
RB3 and RB4 are outputs connected to the H-bridge motor controller IC3. The motor itself is connected across pins 5 and 6 of the pin-strip header SV1. When RB3 is 1 and RB4 is 0 the motor turns anti-clockwise; when RB3 is 0 and RB4 is 1 the motor turns clockwise. When both outputs are 0 the motor is isolated and stops. The PIC software drives RB3 and RB4 with a variable mark-space square wave to control the motor's speed in both directions using pulse-width modulation. IC3 contains shorting diodes to damp the motor's back EMF when its power is cut.
LED2 is driven by software. It is on continuously when the extruder is powerd up but idle, and it flashes when the motor is turning. It doesn't have to be blue, but make the two LEDs different colours so you can tell what's going on more easily.
The pin-strip PCB header SV2 is the power and data connector. The circuit needs a single 12v supply on pin 4 (regulated down by IC2 and its smothing capacitors to 5v for the PIC). Serial data comes in on pin 3 and is transmitted on pin 2. Pin 1 is ground. The serial data is at TTL levels, so if you want to convert it to RS-232 (for example to plug it into a serial port on a PC or a USB-to-serial converter) you need to add a voltage converter chip like the MAX233. Put this on a separate circuit board - the final RepRap machine will do all its internal communication using TTL levels, so ultimately the MAX233 will not be connected to this device.
Header SV3 is an alternative method of control. The software is written so you can turn the extruder on by menu command or by grounding pin 1 of SV3. In the latter case the extruder can act as a dumb device that can be turned on or off with just one logic signal.
Use two 3 mm self-tapping screws to mount the circuit on the spacers sticking out of the side of the nozzle clamp. The location holes in the spacers are 32 mm apart.
The voltage at the top of R3 changes non-linearly with temperature, so I have written a spreadsheet to simplify the calculations. It is in the Electronics directory of the download. There are two files: one in .sxc format for OpenOffice (free), and one in .xls format for Excel (not...).
You fill in the five values top-left. The beta value should come from the thermistor's datasheet, and T0 and R0 are the thermistor's room-temperature resistance, which you should be able to measure with a thermometer and a multimeter. The series R value is the value of R3. A good digital multimeter will give you a better value for this than the colour code... The rail voltage is the voltage driving the circuit - in this case 5v from IC2. The T column is temperature in oC, and the R column is the corresponding resistance of the thermistor. The voltage column is the voltage at the top of R3 measured by the PIC. The two columns on the right are the numbers the PIC's analogue to digital converter generates. The PIC has two reference voltage settings - low and high - hence the two columns. You can select which you want at run time (see below). The A-to-D is only four bits (we're not talking precision here), so values are between 0 and 15 (other values in the spreadsheet are spurious - anything below 0 comes out 0; anything above 15 on the high scale is too hot, but it'll stick at 15...). This coarse measurement is, however, good enough to control the extruder's temperature to within 10 oC or so, which is quite adequate.
A different thermistor to the one I used should be easy to accomodate simply by changing R3 so that the A-to-D values around 170 oC come out to about the same numbers. Remember to get a thermistor that is rated up to at least 250 oC.
5. Software
The software for the extruder is stored in the directory called Software in the download.
The software is for the PIC and is written in C. It is intended to be compiled using SDCC, and GPUTILS which produce a file called extruder.hex that contains all the information that the PIC needs. How you load extruder.hex into the PIC depends on what sort of PIC programmer you have. If your programmer is compatible with the Picstart Plus programmer from Microchip themselves you can use gpicp. I used SDCC version #1055 (Jul 4 2005).
The program (in src/extruder.c) is divided into sections.
The first section saves the set temperature and speed into the PIC's EEPROM, and restores them on power-up. This means that the PIC remembers them even when it is powerd off then on again. The idea is that you can experiment with values by operating the extruder interactively via its serial interface, and then - once you have got the values right - use it in dumb mode where it is turned on (or off) by grounding (or not) pin 1 of SV3.
The second section contains functions that handle communications with the outside world via the PIC's serial interface. If you have the extruder plugged into a PC (via a level shifter such as the MAX233) you can talk and listen to it with programs like minicom (Unix/Linux) or Hyperterminal (Microsoft Windows). The baud rate needs to be set to 2400. Details of the commands you can give and their meaning are given in the Operation section below.
The third section contains functions that control the heater, the motor, and the blue LED.
The final section contains the initialization function and the main program. The main program runs an infinite loop listening for commands and controlling the peripherals by repeated function calls - i.e. crude process timesharing without interrupts. Note that menu items that need further input stop the loop until that input is complete, so once you have typed such a menu-item command letter don't sit scratching your head thinking what hex value to type when the heater is on, for example...
There is a short header file in include/extruder.h that defines a few useful items and some default values.
Under Linux and other Unix variants, you simply type "make" in the Software directory and - if SDCC and the other software are installed correctly - the program should compile and link to produce the extruder.hex file.
There is also a copy of the extruder.hex file in the directory Software/Saved-hex-file, in case you have trouble with the compiler.
At some point I'll publish instructions here for getting everything working under Microsoft Windows and on Macs (if you beat me to it, let me have details...)
6. Material preparation
The extruder needs a supply of 3 mm diameter polymorph rods as input. If you can buy these cheaply, so much the better. But it is quite simple to make your own from Polymorph granules - the usual form in which the polymer is supplied.
To make your own rods first drill a 3 mm hole in - well, just about anything rigid; a scrap piece of aluminium about 10 mm thick is good. You will use this to pass the finished rod through to test that it is not too fat.
Next gather together the following:
An electric kettle
A 300 ml beaker (a clean coffee mug will do fine)
A 3 mm diameter rod about 300 mm long to act as a reference. I used a silver-steel one, but again any rigid material will do.
A sheet of glass about 300 mm x 250 mm (the size is not critical, but don't make the large dimension much less than 300 mm).
A 150 mm length of stiff wire (the single-strand copper used in house wiring is fine - in that case leave the insulation on all but the last 30 mm)
If the glass is freshly
cut, smooth the edges with fine sandpaper so they will not cut you in
turn.A 300 ml beaker (a clean coffee mug will do fine)
A 3 mm diameter rod about 300 mm long to act as a reference. I used a silver-steel one, but again any rigid material will do.
A sheet of glass about 300 mm x 250 mm (the size is not critical, but don't make the large dimension much less than 300 mm).
A 150 mm length of stiff wire (the single-strand copper used in house wiring is fine - in that case leave the insulation on all but the last 30 mm)
Keep everything from here on very clean. Any bits of dust and dirt will become incorporated in the rod you make, and will then clog the extruder nozzle.
Measure out about 3 g of polymorph granules. If you can't weigh 3 g, then this is what 3 g of Polymorph looks like:
The individual granules in the picture are about 3 or 4 mm across.
Bend the wire to form a right angle at the (insulation-stripped) end about 15 mm long. You will use this to hook the Polymorph out of the boiling water in the beaker.
Boil about 1 litre of water in the kettle, put the granules in the beaker, and pour boiling water over them. Fill the beaker about two-thirds full.
The Polymorph will melt and coagulate. Move the resulting lump about underwater with the hook to gather up any stray granules, then pull the lump out.
Using your fingers, squeeze out the hot water, and make the lump into a sausage about 50 mm long and 10 mm in diameter like this:
OBVIOUS WARNING: getting boiling water on your fingers from the Polymorph lump will scald you. Don't do it. Wait till things have cooled a little.
Now roll the Polymorph between the glass sheet and a flat surface such as a table top. When the Polymorph is hot it will squash under the glass very easily, and you will only need the lightest pressure. As it cools it will become more solid and need a greater pressure. The glass lets you see what you are doing.
ANOTHER OBVIOUS WARNING: If you push down too hard on the glass it will break, cut your hand, and spoil the rod you are making with a big red stain. Don't do that either.
As the rod cools it will reach a point where it cannot be worked any further. It will probably look something like this:
Bend it, put it back in the beaker, reboil the kettle, and add new hot water. Take care not to let the Polymorph ends touch each other in the beaker as they will stick together. When the Polymorph is soft again take it out with the hook and repeat the rolling process. You may have to do this several times. When the diameter gets near to 3 mm, put the reference 3 mm rod (B) under the glass like this:
The Polymorph rod is labeled A. You will find that it naturally tends to roll next to the reference rod. With the fact that the glass slopes very slightly towards its bottom edge against the table, this gives a Polymorph rod slightly smaller in diameter than the reference rod (it will end up about 2.9 mm in diameter) - this is exactly what you want. (The reference rod in the picture is not bent - that's a distortion created by the wide-angle camera lens...)
Test the rod when it is cold by putting it through your reference 3 mm hole. It should not be lose, but neither should it jam. If it jams, reheat it and roll it down a bit more. If it's too lose, you've gone too far...
7. Operation
Place the rod between the pinch-wheels in the extruder head and push it down into the PTFE and brass tubes. It should move smoothly but tightly all the way to the bottom without jamming.
Tighten the pinch-wheel-closing M3 nuts. I found that fingernail-tight plus half-a-turn gave about the right pressure, but your fingernails may be different to mine when it comes to using them as a torque wrench. Experiment.
If everything is working properly and the PC serial interface is connected, when you turn the 12v power on to the electronics you should get a prompt like this:
RepRap
>
Note that when the PIC is first turned on after programming it the set temperature and speed are not defined. This means that the heater will probably turn on. Turn it off immediately first time by typing "T00" (without the quotes) at the prompt. This will put the whole machine in idle mode.
The PIC is then in its main control loop thermostatically setting the temperature and switching on or off the motor and the blue LED. After you've typed "T00" the heater should be off (i.e. the red LED should be off) as the default temperature setting is 0. The motor should also be off, and the blue LED should be on continuously, all of which indicate that the device is idle and awaiting your instructions.
At the "> " prompt you can type a variety of single-letter commands, some of which you need to follow with a single hexadecimal number. Such numbers need to be given as two characters (i.e. they represent one byte), with the letters in lower case. Thus "07" (without the quotes) is 7 and "a5" is 165 and so on. When the PIC gives you a hexadecimal number it preceeds it with 0x. You don't need to do this when you type one.
Here are the commands:
| Letter(s) you type |
Meaning |
| t |
Return the temperature that has been set as the target
temperature |
| T0c |
Set the target temperature to 0c (or
whatever). Valid numbers are: 00 to 0f |
| v |
Return the current temperature measured by the thermistor |
| m |
Turn the motor off |
| F |
Turn the motor on forwards |
| R |
Put the motor in reverse |
| s |
Return the currently-set motor speed |
| S80 |
Set the motor speed to 80 (or whatever). Valid
numbers are: 00 to ff |
| L |
Set the analogue-to-digital range low |
| H |
Set the analogue-to-digital range high |
| l (lower case L) |
Return the analogue-to-digital range
(0x01 == low, 0x00 == high) |
Menu items that need further input stop the main loop until that input is complete, so, as I said above, don't sit scratching your head thinking what to type when the heater is on, for example... Updates to the software will be interrupt-driven soon.
When the PIC is turned on subsequent to the first time it will recover the temperature and speed setting that you last used before you last turned it off.
The software contains an interlock that stops the motor turning unless the temperature is up to the set temperature minus 1, whatever command you type. Generally you don't want the motor running in either direction unless the Polymorph in the brass tube is melted, as it will be jammed. But (especially when you are setting things up) it is sometimes useful to be able to turn the motor on when the device is cold and there is no rod in it at all. This you can do when the set temperature is 0, as that condition will be satisfied when the device is cold.
If you want to run the motor when there's a Polymorph rod in the machine without moving the rod, simply slaken off all the pinch-wheel-tightening nuts. The pinch wheels then won't drive the rod.
Note also that if you ask the motor to turn on and the device is not up to temperature, the motor will come on later automatically when the set temperature is reached. So don't turn the motor on, see nothing happening, then go for coffee...
As explained in the Electronics section above, the analogue-to-digital convereter in the PIC software is fairly crude, with only 16 possible values from 00 to 0f. The range of temperatures that can be measured and controlled can be further extended by selecting either the low or the high range for the PIC's internal voltage reference. See the spreadsheet for the actual temperatures that the values represent.
Run through the lower-case commands in the table. These report back values rather than changing things.
Then set the temperature to about 170oC (if you use the thermistor I used that's "H" followed by "T0c"; get the numbers from the spreadsheet for a different thermistor). Set the speed to "S80".
When you first turn the extruder on and set a temperature, the brass tube will come up to that temperature quite quickly. Nonetheless, don't start the motor right away - give the machine a minute or so to stabilise first.
Then set the motor running (the "F" command) and check that it is going in the right direction. If it isn't, stop it with the "m" command and swap the motor leads over.
The Polymorph should extrude well at temperatures around 170oC. Broadly speaking, the hotter it gets, the easier it will extrude.
The machine should extrude polymer whenever the motor is on. There are a couple of seconds delay between the motor starting and stopping and the flow starting and stopping. These are pretty consistent, and so should be easy to allow for in the RepRap driving software.
If you set the motor speed too fast, the pinch wheels will start slipping. You can tighten them a bit more, of course, but there is a limit to how far you want to go as the greater the stresses the machine is working under, the less reliable it will be long-term. Increasing the temperature will also speed up the flow, of course.
Once you have the device working, you can switch it off, disconnect the PC and RS232, and just drive it by sending 0 or 1 to SV3 pin 1. Or you can continue to use the serial interface of course. The effects of the "F" command and a 0 on SV3 pin 1 are effectively ORed by the software; either will turn the device on and keep it on. Note that if you turn the device on with the "F" or "R" commands it will continue to run until you type "m" regardless of the state of SV3 pin 1. Similarly, a 0 on SV3 pin 1 will keep the device on even if you type "m".
8. Bugs, improvements, and future work
Doubtless this section will get longer with time...
You will notice that the RP pinch-wheel holder has an unused through-hole at the back opposite one of the gear wheels. This is intended to takje a shaft out the back from the top pinch-wheel (from which the gear would have been removed). This shaft would drive a toothed wheel rotating in a slotted opto-switch. The resulting signal would be fed into the PIC. As the polymer rod was pulled between the wheels it would cause the un-geared wheel to rotate, driving the toothed wheel. When this stopped it would be an indication that the rod had run out and that it was time to fetch another. This will happen automatically in the final RepRap machine. A slowing of the train of opto-switch pulses would also indicate that the machine was jammed and that the other pinch wheels were slipping on the rod. However, at the moment the through-hole is obscured by the worm drive - design change needed...
Serial input should be interrupt driven.
9. Parts list and sources
If you find a good supplier for a part whom I haven't listed, please let me know. Please give me the URL for the supplier and their part number if you do, though.
Numbers in brackets after links to suppliers are the supplier's part number for the part.
The rapid-prototyped parts obviously have no supplier. The common parts, such as M4 cap screws, can also be obtained at any good hardware or DIY shop, or online engineering suppliers.
Note that many suppliers will only supply much larger quantities than you need of such things as ribbon cable (for which one supplier's minimum is 30 m when all you need is 200 mm...).
9.1 Mechanics
| Number off/quantity |
Description |
Supplier |
| 4 |
stackable pinch-wheel holders | RP part |
| 1 |
motor mounting and worm drive | RP part |
| 1 |
nozzle clamp | RP part |
| 200 mm |
steel M3 studding |
RS (530-292), Farnell
(517320) |
| 12 |
steel M3 nuts |
RS (837-206), Farnell (758796) |
| 18 |
steel or brass M3 washers |
RS (560-338), Farnell (149687) |
| 4 |
steel 16 mm M3 cap screws |
RS (376-4555), Farnell (100165) |
| 1 |
steel 35 mm M3 cap screw | Toolfastdirect (0150M390035) |
| 13 |
steel M3 grub/set screws | RS (431-965), Farnell (757330) |
| 1 |
M3 Allen key |
any DIY shop, Farnell (108662) |
| 1 |
steel 25 mm M4 cap screw |
RS (467-9919), Farnell (103080B) |
| 11 |
steel 40 mm M4 Allen bolts (must
have 20 mm unthreaded at the top of the shaft) |
Toolfastdirect (150M450040) |
| 11 |
steel M4 nuts |
RS (525-896), Farnell (149682) |
| 20 |
steel or brass M4 washers |
RS (525-925), Farnell (149689) |
| 300 mm |
steel M4 studding | RS (530-309), Farnell (517331) |
| 9 |
10 mm pitch-circle 20-tooth 0.5 module steel gears |
Huco (101211020) |
| 1 |
0.5 module 7 mm pitch diameter steel worm gear | Huco (103300000) |
| 1 |
0.5 module 20-tooth bronze worm wheel | Huco (103000007) |
| 60 mm |
10.5 mm diameter PTFE rod (you may find it easier to get 10
mm diameter, in which case change the diameter of the corresponding
hole in the nozzle clamp; no other modifications are necessary) |
Flurocarbon, RS (10 mm: 680-628), Farnell (10 mm: 7174445) |
| 300 mm |
6 mm diameter brass rod |
RS (682-630), Farnell (7096173) |
| 1 |
0.5 mm drill bit |
Modelcraft Collection (N08BN), Farnell (203778) |
| 1 roll |
PTFE plumber's tape |
any DIY shop, RS (512-238), Farnell (110374) |
| A dab |
grease |
any DIY shop |
| A smear |
soldering flux |
any DIY shop |
| 4 | 6mm diameter 12mm long springs (15 N/mm) | Springmasters (D11860) [not needed if you use plastic tube instead] |
9.2 Electronics
| Number off/quantity | Description | Supplier |
| 300 mm |
0.2 mm diameter nichrome wire | Electronics Plus, Inc. (24 AWG =
0.207 mm dia: 24BNC-1/4), Sciencestore (0.27 mm dia: CB0535), MUTR (0.1 mm dia: EW2 024) |
| 200 mm |
4-way ribbon cable |
RS (9-way - divide it: 436-2548), Farnell (3295114) |
| 1 |
Thermistor |
RS (484-0149) |
| 6 |
crimp connectrors |
RS (233-1889) |
| 1 |
geared electric motor, 12v, 60 RPM |
Farnell (147873) |
| 1 |
6-way PCB header pin connector (SV1) | RS (479-181), Farnell (32 way - buy one and cut it: 312241) |
| 1 |
4-way PCB header pin connector (SV2) | RS (467-560), Farnell (32 way - buy one and cut it: 312241) |
| 1 |
red LED (LED1) |
RS (826-515), Farnell (3350861) |
| 1 |
blue LED (LED2) |
RS (247-1561), Farnell (233481) |
| 1 |
180R resistor (R1) | RS (144-150), Farnell (513842) |
| 1 |
560R resistor (R2) | RS (144-217), Farnell (513969 |
| 1 |
220R Resistor (R3) | RS (144-166), Farnell (513866) |
| 1 |
270R Resistor (R4) |
RS (144-172), Farnell (513880) |
| 1 |
10 uF electrolytic capacitor (C1) |
RS (768-728), Farnell (4311267) |
| 1 |
1000 uF electrolytic capacitor (C2) | RS (440-6597), Farnell (3036420) |
| 1 |
0.1 uF capacitor (C3) | RS (829-615), Farnell (3549811) |
| 1 |
TIP110 darlington pair (Q1) | RS (485-9648), Farnell (4347353) |
| 1 |
PIC16F628-P microcontroller (IC1) |
RS (379-2869), Farnell (3154786) |
| 1 |
78L05 voltage regulator (IC2) |
RS (810-295), Farnell (412430) |
| 1 |
BA6286 motor-control H bridge(IC3) |
RS (245-6045) |
| 1 |
18 pin DIL chip socket (for IC1) |
RS (197-2647), Farnell (4242373) |
| 100 mm |
2 mm dia. heatshrink sleve |
RS (366-2847), Farnell (236482) |
9.3 Material
Polymorph granules can be obtained from the following suppliers:
UK:
Maplin Electronics Ltd
National Distribution Centre
Valley Road
Wombwell
Barnsley
South Yorkshire
S73 0BS
Sales: +44 (0)870 4296000
Link to the Maplin Polymorph page.
National Distribution Centre
Valley Road
Wombwell
Barnsley
South Yorkshire
S73 0BS
Sales: +44 (0)870 4296000
Link to the Maplin Polymorph page.
Middlesex University Teaching Resources (MUTR) Ltd.
Unit 10, The IO Centre
Lea Road
Waltham Cross
Herts, EN9 1AS
Link to the MUTR Polymorph page.
Unit 10, The IO Centre
Lea Road
Waltham Cross
Herts, EN9 1AS
Link to the MUTR Polymorph page.
USA (Polymorph is called Friendly Plastic in the USA):
The Compleat Sculptor
90 Vandam Street
West SoHo
N.Y.C.
Phone: 800-9-SCULPT
Link to The Compleat Sculptor Friendly Plastic page.
90 Vandam Street
West SoHo
N.Y.C.
Phone: 800-9-SCULPT
Link to The Compleat Sculptor Friendly Plastic page.
Sunshine Discount Crafts
12335 62nd St N
Largo, FL 33773
Phone: 1-800-729-2878
Link to the Sunshine Discount Crafts Friendly Plastic page.
This company also supplies Friendly Plastic in a wide range of colours.
12335 62nd St N
Largo, FL 33773
Phone: 1-800-729-2878
Link to the Sunshine Discount Crafts Friendly Plastic page.
This company also supplies Friendly Plastic in a wide range of colours.
10. Design and software downloads
To download this webpage and all the associated design files in gzipped .tar format as a single file (20 MB), click here: extruder.tgz.
On Unix/Linux run the file through gunzip and then tar.
gunzip extruder.tgz
tar -xvf extruder.tar
tar -xvf extruder.tar
With GNU tar these steps can be conflated:
tar -xvzf extruder.tgz
Under Microsoft Windows, Winzip will unpack the archive.
The result will be a single directory called extruder which contains:
drwxr-xr-x
2 ab users 4096 Sep 5 15:38 Electronics
-rw-r--r-- 1 ab users 25267 Sep 5 15:46 LICENCE
drwxr-xr-x 2 ab users 4096 Sep 5 15:42 Mechanics
drwxr-xr-x 2 ab users 4096 Sep 6 18:04 Pix
-rw-r--r-- 1 ab users 238 Sep 5 15:29 README
drwxr-xr-x 8 ab users 4096 Sep 5 17:48 Software
-rw-r--r-- 1 ab users 183414 Sep 9 18:08 extruder.html
-rw-r--r-- 1 ab users 25267 Sep 5 15:46 LICENCE
drwxr-xr-x 2 ab users 4096 Sep 5 15:42 Mechanics
drwxr-xr-x 2 ab users 4096 Sep 6 18:04 Pix
-rw-r--r-- 1 ab users 238 Sep 5 15:29 README
drwxr-xr-x 8 ab users 4096 Sep 5 17:48 Software
-rw-r--r-- 1 ab users 183414 Sep 9 18:08 extruder.html
The Electronics directory contains the circuit diagram and the thermistor calculation spreadsheet.
The LICENCE file is the GNU Public Licence.
The Mechanics directory contains the CAD and STL files for the RP parts, and the DXF files for the drawings.
The Pix directory contains all the images used in this page.
The README file is, I hope, reasonably self explanatory.
The Software directory contains the program that needs to be downloaded into the PIC.
The extruder.html file is this page.
11. Licence
RepRap is copyright © 2005 University of Bath, the RepRap researchers (see the project's People page), and other contributors.
This web page and all the other files in its download are covered by that copyright.
Principal author:
Adrian Bowyer
Department of Mechanical Engineering
Faculty of Engineering and Design
University of Bath
Bath BA2 7AY
U.K.
e-mail: A.Bowyer@bath.ac.uk
RepRap is free; you can redistribute it and/or modify it under the terms of the GNU Library General Public Licence as published by the Free Software Foundation; either version 2 of the Licence, or (at your option) any later version.
RepRap is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Library General Public Licence for more details.
For this purpose the words "software" and "library" in the GNU Library General Public Licence are taken to mean any and all computer programs computer files data results documents and other copyright information available from the RepRap project.
You should have received a copy of the GNU Library General Public Licence along with RepRap; if not, write to the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA, or see
http://www.gnu.org/
12. Here it is working
I mounted the extruder on the old Cartesian robot described briefly the the RepRap Blog of 14 May 2005. I set it running to build a 20 mm x 10 mm x 2 mm rectangular block by outlining it and then alternately zig-zag filling it in layers. Here's a picture of the first attempt at making that:
The extruder speed was set to 0x70 and the temperature was 0x08 on the high scale. If you click on the image it links to a Quicktime movie of it working. N.B. This is 20 MB long, and so is not included in the download...