Recent Changes - Search:

G3 Wiki Read First

* HomePage

* What is Gauge 3?

How To Contribute Content

Contact Administrator


MOST RECENT UPDATES:

* Updated Paint / Colour Matching Article

* OpenSCAD and Sample 3D Print Files

* Extended Guide to 3D Printing

Rep Rap

The Bits in the Box.

The name Rep Rap comes from the Repeatable Rapid Prototype project of the University of Bath. The idea was to produce a machine that was simple and cheap to make that could produce its own parts -or more machines. The modern range of RepRap machines have been produced from the GPL designs of the Czech Josef Prusa. Most machines are now based on the Iteration 3 machine design - Prusa i3

My Prusa I3 RepRap machine arrived as a box with two layers each of foamed plastic. Each assembly was clearly marked and logically arranged in the layers. The instructions came on a CD and were in PDF format. The documentation is however in "Chinglish" but everything was understandable with only one exception. This was the "plug out" for the main PCB -however the RepRap Wiki provided an English wording for it. The kit is made from laser cut 6mm plywood and slots and bolts together with ease. It is a two person job as it often requires a "third hand" to hold the part whilst bolting up. The kit comes complete with all the tools for assembly -a range of Allen keys and a screwdriver. The assembly of the main part took about six hours for two people, wiring and "plugging" it took a further two hours. I would advise you to do the wiring and "plugging" as you progress the assembly as some of the wires require feeding through holes. Careful wiring discipline is essential as the cables move as the bed plate and printer arm move and elevate. Ensure that the cables when installed do not pass through loops of other cables.

Spaghetti is allowable but Macrame is (K)NOT...

Armour the cables with "spiral wrap" or "spider wrap". The former is easier to apply the latter is far more robust.

There are several "after market" parts that you might consider at this stage. I have bought; An additional five micro switches for movement detection (23p each), a set of five PTFE top bushes for the Z columns to damp down vibration of the screws (10p each), two CNC grade aluminium flexible couplers for the Z axis motors (£2.06p each), and plugs for the additional micro switches (£1.49p for a bag of 50).

The Build....

The first pic shows the contents of two layers of the box. The main upright comes with a special screw fitted on the RHS of the frame. This is a dome head to prevent the Y carriage from bashing into it, the rest of the fixing are M3 caphead allens of 6mm 12mm and 16mm sizes.

After about 2 hours of slotting parts together and tightening up the allens you arrive at the basic frame. The X axis is the heated bed (red square) the Y axis carries the printer head carriage and the two vertical M8 threaded rods fit, with shrink wrap tubes, to the motors.

The next step is "plugging up" the motors and sensors etc. Fortunately the supplied cables are (seemingly) over long but once threaded through the holes in the frame they are correct. At this point I would recommend that you use "spider wrap" to armour the sensor cables as they are somewhat fragile. This is a black nylon tape "tube" rather like an orange net. In that the mesh expands to encompass the cables and snaps back once released.

The "power" cables are best fitted with "spiral wrap". To ensure good working curvature thread a couple of pieces of filament into the spiral wrap. This will take the working strain -rather than the cables.

The next shot is the most difficult part of the construction... The control PCB is PLASTERED with sockets. They all have individual functions but nearly all use the same connector(!) So a short guided tour.

Top LHS is the main DC feed to the PCB. Top centre is the feed to the heated bed. Top RHS is the feed to the Printer heater.

The two ribbon cables are the feeds to the control display panel.

The four small PCBs are the motor drivers. The black green red blue cabled plugs are the feeds to the four motors.

The bottom row is the heater fan and cooler fan.

Orange and Black are the Z axis sensors. Green and Red are the Y axis sensors. Brown and Blue are the X axis sensors.

The next shot shows the RepRap after construction and before the "after market" parts have been installed. You can see the stepper motor that draws the filament through the heater assembly and squirts it out of the brass nozzle below. The heater block has a thermistor to regulate the temperature of the block. This can be set from the control knob on the display. panel.

This shot shows the Y axis motor and the other Z axis motor. The microswitch "stop sensors" are visible on the top of the Z motor and side of the Y carriage. The connection between the Z motor and the M8 threaded bar is by the supplied "shrink fit" flexible coupling. It does work very well -I just don't like it...

Is it hard to build? The answer has to be no. Are the "after market" parts essential? The answer has to be no -but they do fix a few "niggles I have and I think they should be bought. What will it print? It prints Polymerised Lactic Acid (PLA) and Acrylonitrile Butadiene Styrene (ABS), The former is reckoned to be the easier to use and takes CA very easily, whilst ABS is stronger it requires careful setting of temperatures.


The Control Knob.

This is actually the only control on a RepRap and consists of a multi turn switch and with plunger. Turning the knob presents the user with a menu, pressing the knob selects the item on the menu. There are LOTS of menus...


You will have to use the menu options to setup your printer and align all the motors to the stop points. It took me a week of faffing with the thing to get the settings perfect. Now that they are, the only option on the menus I use is "Print from SD card". This afternoon it has printed six stop lamps and one station gate. It is now as boringly functional as the washing machine or tumble drier. Load it up, select the program and go and do other things whilst it works...

Some of the things not noted in the manuals are the height of the print nozzle above the bed and the tape used to cover the bed. The nozzle height is about a playing cards thickness from the bed. This allows the plastic to "pour" onto the bed and not "squirt out" between the nozzle and the cold build. I have found that of the two common filament media PLA is far easier to use then ABS. It has a lower melting point, is very hard and is slightly translucent. This is not really a problem as I always prime my models. So it is not important if they are pink, purple, blue or red... If you are building a large >6cm long print then use "Masking Tape" or "The Blue Tape" (oil protective painters tape) on the glass bed. This will stop the print sliding about on the glass bed whilst it prints!!! If you have a "big" print then rubbing a small amount of "PRITT" across the bed improves adhesion.

The quality that is acceptable is up to the user. My initial test print of a stop lamp was done at coarse resolution 0.5mm layers. It printed in 85 seconds and simply proved the printer worked. It looked like a very rough model of a stop lamp. But I was happy with it. High resolution of 0.1mm layers produced an acceptable print of a stop lamp in 13 minutes but the layer marks are visible, but these will disappear under primer. Ensure that your feed filament has plenty of slack off the reel. The feed motor to the printer head is fine with at least a loop or two off the reel, what it cannot do is to haul around a 1kg reel of filament.

Is a RepRap the perfect modellers tool? I would say that it gets me to 95% of what I want, while I do other things. The final 5% is down to me anyway.


Materials that are printed

Myriad materials are available, such as Acrylonitrile Butadiene Styrene (ABS), Poly Lactic acid (PLA), Poly Carbonate (PC), Poly Amide (PA), Poly Styrene (PS), "Lignin" (a wood fibre & nylon composite), rubber, and chocolate among many others, with different trade-offs between strength and temperature properties. In addition, even the colour of a given thermoplastic material may affect the strength of the printed object. Clear and Green are thought to be the the strongest. Recently a German company demonstrated for the first time the technical possibility of processing granular PEEK into filament form and 3D printing parts from the filament material using this technique.

The type of material that is printed depends on the capacity of your print head. I use a 0.4mm nozzle and I have a 0.2mm and 0.7mm nozzles. After some experimentation I have found that the 0.2mm nozzle requires very high temperatures to keep the feed at a fluid temperature and it does clog easily because of this. The 0.7mm nozzle will produces acceptable replacement parts for my Rep Rap as they wear out and things like brackets and holders etc. The original RepRaps only used PLA printed parts and they have lasted for several years. PLA is actually a plastic known in Roman times and they made theirs from boiled milk and vinegar and they squirted it between joints of hot wood to make a glue, which proves there is nothing new....

Filament feed material is available in two standard sizes 1.75mm dia and 3mm dia. For 1.75mm filament expect to pay about £11 for a 1Kg reel of PLA and about £25 for a 1kg reel of ABS.


Operational checklist and "what to buy next"...

The main thing to check is that the print head is securely fixed. As the nozzle is Brass, the heater block is Aluminium and the feed mounting tube Steel -these expand and contract at differing rates. The only way to be sure is to tighten everything up while it is hot -with the caveat that everything is going to be at a minimum temperature of 200C... The eventual cheat used by me was to drill tap and fit M3 grub screws that held everything tight -whether it wanted to come loose or not.

Make sure that the feed reel has plenty of slack to it and give it the occasional twirl to free more. I have taken to mounting my feed reel on its side and having the feed filament "spiral" off it.

Check frequently on the state of the PTFE feed pipe after the feed motor to the print head. This can split and you will have molten PLA etc oozing out of the top of your print head which when it cools embeds the assembly in plastic! You have to "preheat" the head and then take it apart and scrape out the muck.

The PTFE pipe is available in 1m lengths at 99p per metre. NEW print heads are very cheap, buy a batch of 10 from eBay and then you can swap out the clogged ones very easily. There are micro drills to help unclog the nozzles. I would recommend that for normal work a 0.4mm nozzle is fine and for detail work a 0.2mm nozzle. I have a couple of 0.7mm nozzles for printing brackets and other non visually critical things. The resultant prints from them are not pretty but brackets and hangers do not have to be that aesthetic...


3D Printing. A guide for beginners.

How it works.

I will only deal with one of the two types of 3D printing. This is called Fused Deposition Modelling (FDM). If you remember from school pottery work, the first pot you would have made would have been a “coil pot”. The clay was rolled into a thin string and then wound around the base, spiralling higher and higher until the pot was formed. The clay coils were stuck together with a mixture of clay and water (slip). In much the same principle an FDM printer produces its output. The filament is heated until it melts and then is squirted out of a fine nozzle onto the layer below. The heat of the liquid fuses this to the previous layer in a manner similar to laying on a run of weld metal from a MIG plant. The printer “head” contains a heating element and a thermal detector. Different materials require differing temperatures for fluidity. The head works as follows. The solid filament is pushed down by the stepper motor gear teeth meshed into it, the filament is force cooled by a fan to prevent it melting, this is the “cold end”. This forms a “piston” as the filament moves down the PTFE tube to the “hot end”. The filament melts and the moving, still solid filament, pushes the melt out of the nozzle whilst itself melting…

The “hot end” is typically over 200 deg C for PLA (Polymerised Lactic Acid).

There are two design schools for the hot end. The Mark 8 hot end has the fan, stepper motor, gear feed etc all mounted directly above the hot end. The J Type hot end has the stepper motor remote from the hot end and the filament is fed by a very long length of PTFE tube to the hot end. The filament is cooled by the J Type head having cooling fins not unlike a motor cycle cylinder. The two camps both have their supporters. The Mark 8 gives better control of filament speed at the expense of having a far heavier lump to move around. The J Type is lighter, thus capable of very high speeds of movement and changes of direction between deposition runs. The PTFE pipe connects to the J Type via an M5 compressed air fitting, (the blue ringed end).

The glass bed of the printer is normally heated to a temperature of around 60 deg C for PLA. Depending on the size of the object to be printed, the glass build plate may have a layer of masking tape to enable the object to be removed from the glass plate…

In the US masking tape is called painter tape and is normally blue(?)


Levelling the Bed.

The print bed must lie in the same plane as the print head to ensure a vertical print. This is done by setting the head to “autohome” and then hitting “disable steppers”. The nozzle is the moved to each of the four corners of the build plate. A “feeler gauge”, (I use a length of 1mm thick steel sheet), is used to test the height of the print head above the build plate. The height is adjusted by winding or unwinding the wing nuts that compress the springs supporting the build plate. This takes about 2 minutes for all four corners. You can purchase “electronic sensors” that work by adding or subtracting the distance from the nozzle at each corner to the build height.

If your machine comes fitted with them then this is fine -but don’t bother with the expense of them as an after market purchase.


Typical design types.

The X-Y-Z type.

This is perhaps the simplest type to understand the operation of, as each part of the printing process involves moving the print head through a distinct number of steps and then squirting material from the nozzle. These are very simple to build from home designs or kits. The print head moves on an arm in the Y axis which moves either up and down in the Z axis and across the build plate the X axis. How you get it to do this is rather variable…

The OMEROD design has a mobile plate and the arm moves with everything on it. This relies on the base of the arm being very rigidly fixed. The system normally uses a Mk 8 print head. The OMEROD design is common.

The PRUSA i3 design has a movable base plate to provide the X axis and the Y arm moves horizontally on a belt and vertically on two screws for the Z axis. The Prusa designs normally uses a Mk8 print head. This design is extremely common.

The Delta type.

This uses three stepper motors and can define a position in “Three Space” by the intersection of the three lengths of the vertical arms -to the base... These normally have a Type J printer head and the build area is round rather than square. These are popular with people who produce the “artistic” 3D prints for vases and statues. Shown below is the Kossel design. This is the commonest of all the Delta designs. They work extremely well and produce beautiful prints -even if they are really strange to watch!!!

The J type head is used to good effect on the Kossel as the fins cool the head prior to it being pumped through the nozzle. The brass block of the hot end can be seen at the basal point of the three arms with the black finned cooler above it. The gear feed drive is mounted at the top with the PTFE tube connecting them. The arms have rose joints on the ends to enable the pivoting action.


Prusa i3 -an owners perspective.

I have owned a PRUSA i3 for over two and a half years as of June 2019. Has it been a happy relationship(?) -on the whole I have to say yes. There is however a steep learning curve as to the techniques and methods used. The first believe or not -was how to switch it on… The power module and the control modules are of differing sides of the machine. Flipping the power switch on the RHS lit the blue information screen. The control for the machine is actually the small “click select” knob on the RHS provides the list of menus. Turning and clicking the menu option is how the machine is controlled.

I have found that the machine will perform a better print if the system has been set to “Pre Heat PLA1” and reached its temperature before you select “Print from SD”. This does not stop you simply flipping the switch and telling it just to print, but it operates better from “steady state”.

The “poly fuses” on the main PCB can be a source of frustration. If you “blow” a poly fuse it is supposed to reset once you have switched the machine off. But sometimes you have REALLY totalled the fuse! These sit next the main power input lines from the power module to the PCB. They are small square and very fiddly. I get mine from RS components -they are dirt cheap and a packet of 10 is pennies.

The machine can be running for a long time -put it somewhere where the conducted vibration will not drive the family insane. Mine lives on the top of the family Baby Grande piano in the living room. Some cork padding on the six feet will help stop it sliding around on the polished surfaces.

Changing the nozzle is quite often painful!!! They are cheap (circa 15p) and easy to replace -however the head is at 200C and you have to do this while the head is “hot” due to the differing expansion rates of the head. Just tightening the nozzle to the head whilst it is cold will produce a continuous dribble… I only print PLA with one colour at a time. I use a pipe wrench to hold the heater block and a pair of pliers to unscrew the nozzle. Typically this is where I short out the heater wires and blow the poly fuse… Why do I need to replace a nozzle? They are made from brass and the continuous high temperature causes oxidation and the nozzle shape changes with it. You can get Stainless Steel nozzles or even ones made of synthetic ruby(!)

The system will “default” to a Frame rate of 100% with a PLA setting of 200C for the Head and 60C for the Bed. This is a good starting place to begin experiments with a new reel. I often need to do a few small test prints of simple items, (a BR OLEO is a good one). The “Tune” option on the menu gives access to the settings to play with. Once set up it can be left to print with just the occasional “look” to see if it is “happy”. Print times of days for an object have been done from my machine -to produce parts for the next one.

Setting the nozzle height is again best done when the machine is “hot” I use a “feeler gauge” of a piece of 1mm sheet steel and wind the bed plate up or down on the springs at each corner. The first layer, “the skirt”, is 0.3mm thick regardless of what you have set in the slicing program and this is used to see if the nozzle height is correct.

It is probably nowadays an obsolete design and the build plate is only 200x200x180mm. But it is still a great design and I would recommend it as a first machine.


Layer Height, Frame Rate and Nozzle Size.

There is a lot of confusion on this subject. The three interact so much there is no answer -other than empirical guidelines...

Layer Height

Is the thickness of one pass of the printer. The thinnest one is normally somewhere in the region of 0.05mm or 50microns. Does this improve the precision of the print?

The answer is -MAYBE...

It will certainly increase the print time and some experimentation with frame rates, nozzle sizes, nozzle temperatures and bed temperatures will be required. Thin layers will cool rapidly and have very little “heat” in them -thus it can cause problems with welding to the layer that it is being deposited on. I normally print at 0.1 or 0.2mm layers.

Frame Rate

Is the amount of time that is used to complete a layer. This is a variable that can be used to alter both the finish and time required to print the object. A High frame rate (200%) will produce a roughish finish, which might be exactly what you require. A Slower frame rate (50%) will produce a smoother finish. In the former a stone wall, in the latter a steel sheet. High frame rates will cause the heater element to strain in the hot end, as these are normally only rated at 40W and best used with a print temperature below 250 deg C with narrower nozzles, (see below). There are after market heaters rated at 50W that will reach 350deg C for under £4. The stone walls for Brassica station were printed at a frame rate of 200% with a nozzle size of 0.4mm to produce a rough finish. I have printed a K2 telephone box at 50% to make it seem more like smooth metal. The Tardis at 100% -like the K2 both were printed with a 0.4mm nozzle.

Nozzle Size

Is the size of the output nozzle of the printer and thus dictates the amount of filament that can be printed in one pass. To print very large objects at high frame rates then a large nozzle is required. However if you fit a small nozzle the amount of filament that passes through the nozzle is low -thus the nozzle is not very well heated and the filament can set and clog the nozzle solid. With small sizes it is an advantage to “lag” the nozzle with a stuck on piece of cork sheet -or a beer mat(!) Large nozzles will give some visible layer lines, it is up to you to decide if they are objectionable. High build primer or a light sand/scrape will remove them. Small nozzles will give a finer finish but you may not actually need it for what you are printing. I only use 0.4 and 0.2mm dia nozzles

The 1st example above was printed at 50% frame rate and used a 0.2mm nozzle. The 1mm bore holes on the shuttle plate and the end of the lever arm did require a quick twist with a reamer -but that was all. Nozzles are normally made of Brass but increasingly of Stainless Steel. I have seen a nozzle made from Titanium and there is at least one supplier of them in synthetic Sapphire...

You can obtain multi feed print heads -but these are best used in the Delta designs due to the “plumbing” problems of the pipes cables and feeds motors.

If you are making a Mould or Master then it makes sense to print at as high a finish as possible. But do Models need to be at that sort of resolution when they will be covered with primer and paint?


Printing materials.

This is the “options screen” from CURA. It lists various types of filament types from specific suppliers or generics. It is from here that the program develops the GCODE for the object to be printed.

Selecting the options pre-loads the specs for the printer to use.


Printing problems and how to cure them.

N.B. The faults have been specifically engineered to show them.

1: Cold Build Plate.

This is characterised by the base layer peeling away from the build plate and it resembles “ribbons” as the print build base layer progresses. This is not a major problem unless the build is tall and thin… The next layer will normally weld the ribbons together and the fault will disappear after about two layers. However the “grip” of the model on the baseplate is low. Next time, increase the bed temperature, increase the temperature of the nozzle and slow down the frame rate.

2: Step.

This when the printer head has jumped a cog on the drive belts or screws and continued printing from this new offset position... Or something has whacked the printer to induce this, the family cat is a prime culprit -as are small inquisitive fingers…

3: Collapse.

This is the build not having enough support during the print process. Several slicing programs have the option to include “support structures” for your print. Another source of collapse is the fact that the fill rate is either too little, (structurally weak), or too high, (physically heavy). The last problem is the physical shape of the print. This can make the Centre of Gravity far too high or too offset from the centre line for the adhesion of the base on the baseplate to hold. The solution is to either slow down the frame rate so that the bed movement producing the Kinetic Energy is below the adhesion failure point of the base -or weight the internal structure of the print lowering the Centre of Gravity with sand or salt. This will alter the leverage moment from the Centre of Gravity to the baseplate and shorten it. The Delta design has a fixed bedplate which enables it to produce tall thin structures. But there does come a point when the physical leverage from the Delta’s J type head can induce failure -due to mechanical leverage of the end point to the vertical axis of the Delta. Thus like a pot spinning too fast on the wheel -it folds...

4: Spaghetti.

This is when the filament does not adhere to the build. This can be due to two main reasons. The frame rate is too high and it doesn’t have enough time to weld to the layer below or the nozzle temperature is too low. Slow the frame rate and increase the nozzle temperature. Note the layer separation in the shot below caused by poor welding of the melt and it has become strands.

5: Fill Rate % and types of fill “patten”

This is the amount of fill and type of structure that is used between the “skins” of the print. Different slicing programs have differing structural “honeycombs”. If you are printing a large non structural model, (such as a Tardis), then the fill rate can be extremely low (say 20%). However if you are printing a structural part for a tool -then 50% to 100% would be reasonable. The normal fill pattern is a longitudinal diamond -the more advanced slicing programs can produce longitudinal hexagons or a series of interlocking cubes to combat stress points. Although producing a too dense fill rate is just slow and expensive it can produce a model that is too heavy and thus unstable as it prints on the bed plate.

Just a few of the fill pattens in the CURA program that can be specified for a print. I normally use a simple cross hatch diamond, but some of them are useful for “elastic” prints -like bangles or radio controlled car tyres.

6: Jitters.

These can be heard coming from the head and felt in the feed filament as a “tack tack tack” sound or sensation. This is caused by the geared tooth cog slipping on the filament. The reason is simple, there is no melt moving through the print nozzle. There are three possible reasons. The PTFE pipe that links the gear drive at the Cold end to the Hot end has either broken or become blocked/folded -thus starving the head of filament. If this is the cause then you will have to dismantle the head whilst it is still hot and replace the PTFE pipe. The next cause is EITHER the frame rate is too high OR the nozzle temperature is too low. Lower the frame rate or increase the nozzle temperature.

When you dismantle the head to replace the PTFE tube ensure that the feed hole down through the heater block is free of any melt materials. These will conduct heat and cause the filament to be “hot” by the time it reaches the Hot end. Melt material can then be forced back up the PTFE tube and cover everything with melt. The only way remove the solidified melt -is to chisel it off…

7: Hairy Prints.

The filament should retract as the nozzle moves to a new position on the print. This is down to the filament still dripping down the nozzle from the previous print run. If the hairs are internal there is nothing really to worry about. If the hairs are drawn between close parts of the print, (or another print on the same bed plate), then these can be annoying -but in no way damaging. If it is a major problem then reduce the temperature on the nozzle.


Edit - History - Print - Recent Changes - Search
Page last modified on June 26, 2019, at 06:20 PM