Skip navigation

3D Printing

12 Posts authored by: Andy Clark (Workshopshed) Top Member

During the Enchanted Objects design challenge I learnt quite a few things about 3D printing and OpenSCAD, I thought I'd summarise and share them here. My posts in this challenge were all themed with a running story hence the strange titles. My printer is the Cel Robox which I received earlier in the year for a roadtest, this was my first exposure to 3D printing.


Failure to making parts that clip together, self tapping screw to make moving joints

Enchanted Objects Design Challenge - The Woodcutter and the Blacksmith


The above parts in action

Servo and mechanics in action


A fibre optic adapter for the Arduino Yún

Enchanted Objects Design Challenge - The Crystal Cave


Simple window frame

Enchanted Objects Design Challenge - Windows 10


Standoffs, printing first time, Wifi and Power buttons with embossed logos

A bit of 3D Printing


From earlier in the year, my review of the Robox


All the models I created and the source OpenSCAD script can be seen here

I've been designing a magnifier light and one of the parts I thought I'd 3D print. It's right on the limits of size that the Robox can manage but it just fits onto the bed. I started a test print to check the sizes and it seems to be correct. I did however increase the speed and layer thickness and that meant that the edges went quite wavy. I'll give this another go when I've got some time as I believe it could be a very long print job although I think a few model changes and adjusting a couple of settings could improve print time.


Challenge Print

I also gave the temple challenge print another go. This time I simply loaded it up with the Normal settings and printed. I think that the needle valves really help with this kind of model as you could see them in action as each of the slender columns was built up. As it reached the point nearing the top of the arches the model seemed a little unstable but these were knitted together successfully in the next couple of layers. However I'd say that was a feature of the model rather than the printer. If you had problems you could reduce the speed of the print and that should help.

2015-02-20 22.52.12.jpg2015-02-20 23.45.12.jpg


Software update

Whilst on the CEL Robox forum, I spotted that the 1.1 version of Automaker was shipped with the auto-update turned off so I needed to download the latest 1.1.2 manually. I noticed improvements described for filament load and eject, I've not had any issues with those with this model, the preproduction one was quite tricky in that regard. There is also an update to the pre-print purge to ensure that the nozzles are clean before printing, I've not actually had any issues with that step. The door open was giving me some issues with the production printer but seems to be a bit softer with this software update and now reliably unlocks.




As mentioned in my original plan of action, I hoped to print some gears as part of my road test. There's a couple of reasons for this. It should test out the accuracy of the printer with fine detail. It should also check it's ability to print small and thin things without warping.


Given that I'd already had some success with OpenSCAD I decided to continue using it for this test.

I found a Parametric Involute Bevel and Spur Gears script by Greg Frost but I could not work out how to use it. Luckily there was also Spur Gear Fitter Script by Cliff L. Biffle which uses that library to "fit" gears together. I tweeked that to space the gears slightly and generated the above image and an STL file for the printer.




The print was scaled to 25% giving a large gear of approx 20 mm and a small gear of 8 mm, it should be possible to get OpenSCAD to print out the sizes so I'll do that for next time.


My first attempt at printing was a fail in that it did not stick to the bed. I removed the failed prints and gave the bed a wipe down with a rub. For good measure I also moved the gears to a different part of the bed. I also changed the print speed to 50%. This did not help. I checked all the settings and it was at this point that I spotted that my white filament had been ABS not PLA all along.


I tried again in the middle of the bed at fine resolution and normal speed, I also added a large brim to the print. This seemed to work a lot better with the gears sticking to the bed although there seemed to be some curling up at the edges. The print completed without error and I waited for the bed to cool before removing it. The brim broke off easily so the gears should be good to clean up. The over all shape of the teeth and the inner holes is very well reproduced. However when we look at the gears from the side the results are not so good. There is significant distortion in the teeth shape rendering it unusable as a gear. I've not bothered with scanning this one as you can tell all you need from the photos.


I hope to get some advice for how to print better gears and I'm wondering if a bounding wall might help keep cooling consistent.

As mentioned in the previous post the printer had to be sent off for an upgrade. When it came back there were some differences.


Firstly the packaging had been updated. The printer comes in a smart Robox bag which I'm sure will nicely double as a dust cover.

2015-02-06 20.58.08-2.jpg

The full set of accessories is also present with the oil and tweezers missing from the original pack now being present.

The bed, head and z-axis are now all clipped in place with some 3D printed clips to keep them from moving in transit. The bed is also different with metal clips holding the bed surface in place.

The paper documentation still has some of the issues mentioned back in my early reports.

2015-02-06 21.00.56.jpg

There are some other differences such as decals showing the path for the filament and a screw keeping the electronics panel shut.

2015-02-06 21.04.24.jpg

Otherwise the appearance is the same, however I checked the serial numbers for the printer and head and confirmed that those are both different.


Before printing I did a purge which was a good thing as it would appear the previous colour was a kind of lilac. At the end of the purge I did spot a small problem. The bed travelled forward so quickly it bounced off the front and the locking catch did not properly disengage.


For my first test I picked the Robox Robot model that was supplied on the supplied USB business card. It reported a 4hr print so I abandoned that for the time being and picked something smaller.


Instead I picked "Support free strong Bolt  by Jack Imakr" and selected to print that at Normal resolution. That had a more reasonable 1hr 18minutes print time.


The bed adhesion was good and it set off printing. I notice that the initial bed levelling checks for more points on the bed than we previous version of the software. Another thing I spotted was that it was only using the fine nozzle not the 0.8mm one. I'll check with CEL to see if that is normal behaviour.

2015-02-06 21.47.23.jpg2015-02-06 21.46.52.jpgNuts.jpg

Once the nut had finished printing it was interesting to note a distinct slow down in the rate of printing. This is the smarts in the software kicking in as they know that the smaller bolt thread would over heat if printed at a faster speed.

The print went smoothly and there was no warping of the parts in contact with the bed.

2015-02-06 22.55.35.jpg

When the bolt was nearly finished the printing stopped and the light started flashing red. The printer status in Automaker when blank. I think that might have been because my laptop was running on batteries. I unplugged the USB and replugged it and the status re-appeared. I could then press play to continue and it printed the rest of the bolt. That happened again after a couple of layers so I plugged in my power and tried again. This time the print continued to the end with no issues. I suspect the best strategy if your laptop is running on batteries is to disconnect it once the printing has started.


I waited till the fan had stopped and the nut and bolt could be simply lifted off the bed with no effort.

2015-02-06 23.18.41.jpg

The nut screwed onto the bolt easily and securely.

2015-02-06 23.22.08.jpg

All in all a very good experience and a top quality print.

2015-02-06 23.25.39.jpg

My next challenge is to print some parts that I can accurately measure and to print some gear wheels as originally planned. I'll also give those challenge prints another go, I've got high hopes for success on those.


I've managed to find yet another source of gears and that's the McMaster-Carr catalogue.


This Roadtest has been done on an "Engineering Release" version of the product so there may be differences to the released product


Fan issue

CEL told me that the strange fan issue from the last post was a known issue with these engineering release printers and was caused by a MOSFET failing due to high voltages when the fan is turned off.


New Bed

CEL also sent me a new bed with a different texture. Installing a new bed was very straight forward, simply unclip the old one and slot the new one it. The bed heater is separate from the bed so you don't need to do any rewiring etc.


The old bed is on the right of the picture, the new one to the left, you can just tell that it's smoother in texture.


Only a simple alcohol wipe was needed to prepare the bed.


Latest AutoMaker Software

Whilst I was chatting to CEL on Twitter they also commented that there was an updated version of AutoMaker I could take a look at.


  • Cura is now the default slicer - you can change between Cura and Slic3r on a system-wide basis or per-profile
  • The GUI has had the first of a series of major changes, paving the way for more functions and easier use
  • Calibration has had a major update and now includes XY calibration
  • Language preference can be chosen in the preferences screen (press the cog at the top-right). It takes effect the next time you run AutoMaker.
  • All of the profiles and macros have been updated for improved results.
  • Improvements in homing, levelling


It's still a release candidate so there are some issue but CEL thought it might be better for screenshots than the previous release. I tried out the calibration, there are easy to follow step by step instructions and my nozzle alignment produce a very high quality print.

Calibration.png CalibrationResults.jpg

Automaker now has some nice tool tips on the buttons and hints when your filament spool is not correctly loaded etc. It makes it easier to understand what's going on. There's also nice big buttons to turn the ambient light off so I can get better photos although I still need a bit of practice at that

The new bed and software made printing a lot simpler and my latest print, a phone stand has printed very well with no curling or warping. To form the pivot the printer has to bridge between the two sides effectively printing in thin air. It did that very well, much better than it had done on the double helix in my last print session. There was a tiny bit of clean up to do on the pivot so that it clipped together and swivelled properly but otherwise an excellent result first time.

Printer to be returned

It was the intention of CEL and Element14 for us to RoadTest the production printers so this particular printer will be boxed up and sent back to be fixed and refurbished. This means there will be a pause in this series of blog posts but there should be even better results when it returns.


This Roadtest has been done on an "Engineering Release" version of the product so there may be differences to the released product


After my success printing a sheep I thought I'd have a go at something else. Later in January I'll be meeting with the Alumni Association of the Student Television of Imperial College (or PSTOIC for short) and every year we have a mini award ceremony. For this reason I thought I'd make a medal for the winner based on an old version of the TV Station logo.



I used OpenSCAD to build the model as that gave me an opportunity to try out the software in advance of printing some gears. OpenSCAD was straight forward to use with the exception of the text which was not supported in the version I used although there does seem to be some work in progress with regards getting that to work natively. I downloaded a simple Text to DXF tool (designed for engravers) and imported my two words into OpenSCAD that way to produce the results above.

PSTOIC_Medal_Print.jpg PrintedMedal.jpg

Again I gave the bed a quick polish with some kitchen roll before running and had good adhesion to the bed. The print went well but there were a few holes in the top surface at the top, I'm not sure if this is an issue with the slicing software or with the printer but it would appear that it did not really put a top on the letters either so something is not quite right with the print, however I might be able to fix this with a bit of post processing in the form of primer paint.


Following this print I had another case where the printer refused to recognise that the element was fed into the head despite the fact I could actually get it to extrude by manually opening the valve and advancing the filament. My work around is to heat the head up and then eject the filament followed by heating the head again before re-loading as dougw had suggest I try.


Another strange issue was that at one point AutoMaker suddenly told me that there were no printers attached, I removed and replugged the USB connection and it correctly re-recognised the printer.


My next challenge print was the double helix, this is a seemingly impossible print to do without support structures. However the strange thing is that it the cross pieces do print, I can see that carefully taylored GCode would make them print better though such as making the first crossing down pass all the way across the supports rather than trying to reverse immediately it got to the other side, perhaps that's something the software could pick up on?

DoubleHelix.png DoubleHelix.jpgDoubleHelixCleanedUp.jpg

My printer made it upto 3.5 rungs before it became unstuck from the bed, cleaned up it does not look too bad, perhaps I'll just tell people it's Flea DNA hence why it's smaller.


My final bit of strange behaviour was that the PSU / extractor fan stayed on for all of this print, I thought this was something to do with the thin thickness but it stayed on afterward so there might be a problem there? I'll let it cool down and see it the problem re-occurs next time I use the printer.


This Roadtest has been done on an "Engineering Release" version of the product so there may be differences to the released product


At the end of my last post on test prints the printer was out of commission with the bowden tube and extruder parting company. CEL responded promptly to my support issue and despite the New Year managed to get me a new tube very rapidly.


Bowden Tube Change


Old tube on the left, new tube with metal ferrule on the right.


The Bowden tube upgrade instructions are straight forward and have an overview and photos. They refer to a couple of other sets of instructions such as removing the head, replacing the belts and removing the side panels. One of my biggest issues was finding a 2mm hex key which was needed to remove the old fitting, luckily I had one with my lathe tools. The head removal instructions could do with clearer reference to unscrewing the finger wheel at the back of the head, you could potentially do some damage if you don't undo that first. Otherwise I did not have any issues with the instructions.


I ran a manual purge before continuing, as was not sure if my 0.8mm nozzle was the cause of the problem, it purged quite easily so perhaps not. I managed to burn finger doing this so won't be doing it again unless absolutely necessary. The manual purge instructions mention to modify the "B" settings but don't specifically mention setting them back. I made a note of the values and set them back to the original values following the purge.


It took several attempts to put the head back on and get it recognised by software, it is a bit fiddly to get the correct alignment so that the pins make connection before you screw the head back in place.


Challenge Models


As mentioned in my plan of action I'd got some "challenging" models from Richard Gain who's been working with 3D printers for some time now.


Richard's first mode the Lightbulb model seemed to get stuck on 70%, "Slicing skirt". So I abandoned that but will try to give it another go at some point.


The second model the temple did not cause any problems slicing and auto generated some "support structures" for the peaks above the windows.



The temple model correctly reported that it was too tall for the chamber so I scaled that to 95% and continued. With the default settings this did not stick so I swapped for some custom settings I'd been playing with. That worked better with some filament sticking but it still messed up at the back corner. As an experiment I move the model over to one side to see if it would work better there, that did not help either. I tried the fine and normal default settings and that did not help either.


So I thought I'd try cleaning the bed and bit more thoroughly. I took a tiny amount of "swarfega power" this contains cornmeal and citrus oil so is great at degreasing things but is still gentle. I followed this with the usual alcohol wipe. I tried the default settings with again no luck so moved back to my custom settings. If anything, things seemed worse. I followed this with a quick "cook", 5 minutes with the bed heater at 99°C. Once it was cool again I felt the edge of the bed and it felt quite greasy so I gave it a good scrubbing with some kitchen roll. I followed this with a default "fine" print and stickage was much better but not perfect so I stopped and repeated with the kitchen roll.



2015-01-01 23.45.27.jpg

I posted a few pictures of work in progress on Twitter. It was at this point that CEL told me I'd got an old heated bed and they were going to send me a new one. So it looks like I'm going to be testing out another set of instructions.


I left the print running for some time and just as I was about to leave to go to bed one of the support structures fell over so I decided to abandon things at this point. Once I'd cracked off the support structures the main print was very good with just a tiny amount of warping in one corner.

2015-01-02 00.27.29.jpg

I ran it again using the "normal" setting (with no support structures) and it progressed quite well but whilst printing the roof one of the nozzles snagged on the structure and it was pulled off the platform which is a shame as it was printing quite well until that point.


With the help of my daughter we picked a less challenging item to print and despite a rather major failure of the support structures, it actually printed ok. I'll make a note of that for future prints and see if I can adjust the support structure settings.

2015-01-02 12.21.33.jpgSheep.jpg

Fan test


Fellow road test thundersqueak mentioned that she'd had some problems with the fan. I did not think this applied to my printer but even so I decided to try out the G Code direct entry and ran the following commands in turn to see if the head fan was working correctly. These run the fan at max speed down to a lower speed. Note that you can't run these whilst the head is hot as the printer won't let you go below 50% in that circumstance. The G code syntax does support loops and variables but I've not done that for all long time so don't remember the syntax.


M106 S255

M106 S200

M106 S160

M106 S60

M106 S10



My fan was fine with all of these speeds so I can confirm that there's nothing wrong there.


This Roadtest has been done on an "Engineering Release" version of the product so there may be differences to the released product


For my first few prints I thought I'd make some toys for the family, the results would not be critical and they should give the printer a bit of a work out.


Before each print the printer does a short purge of the print material and cleans the tip using the little wiper at the front, it's important to clean this off periodically otherwise the nozzle will grab take this blob of plastic away with it on the next print and that will make a right mess.


Train Track


As this print started up, I knew that I did not have good adhesion to the bed but I thought I continue anyway as so far I'd not managed to print anything at all.


Although this was a bit touch and go to start with but the nozzles did not snag on the flapping ends.


The bridging for the trackbed was successful and the part infill worked well, however as the part continued to print the corners started to lift off the bed.


However, given that this was just a toy the end results were not too important, the end results are below.

Top, side and pre-tidied bottom. Once the "brim" had been trimmed off with the supplied chisels the bottom of the part looked pretty respectable.

FinshedTop.jpg FinishedSide.jpgFinishedBottom.jpg


Balanced Die


This part was selected to check dimensional accuracy, I'll be writing a further post on that once I've printed some other parts.

I scaled this model so that it would be 30mm so that I could then measure that it was accurately reproduced. There was good adhesion to the bed and it seemed to start ok, however as the print progressed it was clear that the corners were curling up again as the printer was rubbing against these as it printed the perimeter. Apart from this the print went well.


The One was the Top of the print with the Six being the layer lying on the bed.



Motorbike Jigsaw


The model told me that it was going to take 7hrs to print so I thought I'd scale it to 50% that also allowed me to put all of the parts on the bed rather than having it split over three prints, as previous mentioned the software did not managed to auto layout the parts.


Before I printed this one I had a horrific noise from the printer, it sounded like the extruder was slipping, I managed to pause before any serious problem occurred. The filament would not feed after this so I ejected it chopped off everything that felt rough and re-fed the filament, the printer seemed to load correctly after that. However, it turned out that all was not fine and my bowden tube had popped out of the printer and the filament rather than going into the head was spooling out into the build chamber.


CEL responded quickly to my twitter message and are sending me an upgraded tube which looks a bit like the following.



This Roadtest has been done on an "Engineering Release" version of the product so there may be differences to the released product


The Robox case and mechanism feels very robust, with a mix of metal and plastic, it's held together with lots of hex socket screws. The case is split into two parts with the reel and electronics on the left and the printing space on the right. The top, front and right side provide great viewing area to see the printing in process.


The printer has an RGB light at the top and I was initially a bit sceptical about the use of this. However given that you could run multiple printers from the same computer it could be useful to colour code each one as the colour is displayed both by the light and on the software so you could use them for identification.



The door lifts up and out of the way and has a good thickness to it so also feels robust.

RoboxDoorMechanism.jpg RoboxDoorLock.jpg

The door locking mechanism is quite elegant in that it is opened by moving the bed all the way to the front. The software won't let you unlock the door if the bed or heat are too hot and a warning is displayed on the screen.


The reel holder has a large spindle with locking pins to hold the reel in place. There are two lots of contacts on the reel so it's not possible to mount it incorrectly, the smart reels seem to have a small chip (eeprom) that stores the settings for that material. The middle of the spindle acts as a status light and control button, one press to pause and a long press to eject the filament. The holes for the filament are large and funnel shaped so it's easy to load the filament into place.

RoboxSpoolHolder.jpg Spool.jpgSmartSpoolElectronics.jpg

The bed is moved with a belt as is the X axis, all three axis are supported by metal rods and linear bearings. The Z axis is driven by a screw thread.


The bed itself is heated with a mains powered heater under the bed, on top of this sits a simple sheet which can be easily unclipped to be replaced or repaired.


The front edge of the bed is a wiper which is used to clean the nozzles before each print. The front of the printer case also unclips so you can tip out any plastic pieces that have fallen off the bed and down one of the slots.


The head has dual nozzles the fine size is used for the outside and the larger head is used to fill the shapes. This means that you get faster printing speeds without sacrificing the detail. These are on a rocker so that the heads move up when not in use. I had to take the head off to upgrade the Bowden tube so here's some photos of it up close.


The head has a fan on the top to control the cooling of the nozzles and filament, there is also a larger fan on the top of the printer by the power supply to rapidly evacuate the hot air once the print is complete. By controlling these two fan the build temperature can be controlled. This is all done automatically for you but you can over-ride the default temperatures if you like.


As you might tell this particular print has an issue with the back edge curling up from the bed so it turned out a bit misshapen. I'll write up another article on the different things I print and any issues I have.


A side effect of a problem I've had with the bowden tube (the tube that feeds the filament) is that I've got some photos of inside the electronics compartment. The brains of this model is an Atmel micro-controller. You can see in the last picture the bowden tube that's no longer attached to the extruder.



This Roadtest has been done on an "Engineering Release" version of the product so there may be differences to the released product

As per the software the documentation looks smart and professional but needs a little polishing to get it perfect. The following might sound a little pedantic but please don't think that I'm being negative, my comments are there so things can be improved.


The first bit of documentation you'll see is the wrapper on the box. That's followed by a couple of quick notes on the flaps to help you get the printer out of the box.

The "quick start" guide repeats these instructions but mentioned a "plastic handle" on the accessories box, this box actually has two finger holes not a handle. There's also a mention of the warranty card on the card insert but a different reference to that being on the printer bed, although in my case a warranty card was not provided.


On the package content there is mention of a axis lubricant and tweezers, neither of these were in the package, I have yet to ascertain if this is an issue but I do have both a variety of lubricants ans tweezers if I need them. Again being slightly pedantic, it's not actually possible to check the contents of the packages until after unpacking which is the next step.


My next issue is slightly more significant, on step 4/5 it was not possible to open the door as the bed had not been homed to the front. This meant that the door latch had locked the door shut. Luckily I spotted this and released the catch (with a selected kitchen implement) rather than forcing the door open. I don't know if this was an issue with the long journey or if the printer had simply not been put away properly as the bed was right at the back.


On step 5 there is a mention of a blue clip for the print head this was white in my version and the print head was not secured to the right as per the diagram but was in the middle. Again I don't know if this is an issue with the transatlantic journey that the printer had undertaken.


A couple of minor differences on the diagram in step 5 is that the microSD Card has a cover on the printer but not in the illustration, the power switch and power socket are the other way around to the diagram.


The explanation of the components is clear and everything exists as expected.


As I'd already installed the software from the web I thought I'd uninstall and re-install from the provided USB. The USB has software, models and manuals as well as linked which I checked worked ok.


The screenshot for the final step is also slightly out as there is an option to run the software from this menu. The USB stick had a version 1.1 of the manual on it with several pages reporting that they were not complete, the 1.2 version could be downloaded from the web and had these omissions fixed although they still said 1.1 on the front cover.


The software provided appeared to be a beta version "AutoMaker-beta-installer-windows.exe" but I thought I'd try it to see if the autoupdate worked. The instruction suggest that there will be a language option but on both the provided and downloaded version this does not exist.


The autoupdater worked successfully and downloaded and installed the version 1.00.17 from the web. As mentioned previously the account registration page was not displayed, later when looking at the USB stick there does seem to be a PDF warranty card. The USB stick also contains a manual but there is a newer version on the website.


I continued to the next step and attached the USB and power connectors as mentioned in the instructions the drivers load at this point. I turned on the power and the AutoMaker software prompted me to update the Roxobox Firmware.


The firmware rapidly installed and the printer status was reported.


There was a reel of white PLA and a reel of green ABS provided in sealed but resealable ziplock bags.


I thought I'd try the PLA so I cut the filament as described and fed it into the top slot, the instructions suggested these are numbered but that was not the case for my printer. I fed the filament in and there was a whirring noise but the filament did not get pulled it. I pulled it out and tried again, this time it was pulled so I clipped the reel into place and was told that the reel had been updated and no action was required.


Warning: Don't be too hasty to clip the reel in place, wait till it's finished pulling in the filament before clipping in the reel otherwise it won't let you continue. The printer is not very good at telling you that it's in this state and I did manage to get it into this situation a second time (without trying very hard). I turned it on and off then ejected and re-loaded the spool.


My first print failed to stick to the bed but luckily I spotted that and pressed the pause button in the middle of the reel. I tried purging the head and by the time that had completed both nozzles were sticking correctly. So I then tried to print a little version of the vase model that had been supplied. This also failed with one corner sticking up but otherwise a good print. So my next print I tried another test piece that was provided "test_150sq.stl", this also failed badly with the job not sticking at all.

2014-12-22 21.14.15.jpg

So a poor start, back to the manual to see if / how the bed needs cleaning and also to review the forums to see recommendations for getting PLA to stick to the bed.


The forums are relatively quiet but there is a good support site with detailed walk throughs for various processes such as changing a head or reattaching the Bowden tube. So in summary the documentation is good but some of it just needs a few updates.


This Roadtest has been done on an "Engineering Release" version of the product so there may be differences to the released product


Fellow roadtester dougw mentioned that he'd had to upgrade his video card to run the automaker software so I thought I'd also check my machine which is a Dell OptiPlex 755 from a few years back. It's the ultra small form factor so there's no option to do any upgrades.


I downloaded the Automaker software and it reported a problem.


I had to register online to get access to the support site and then found the following issue:


"I am running Windows and AutoMaker tells me that my computer does not satisfy the 3D requirements"
"AutoMaker runs on JavaFX, which requires a certain minimum level of 3D support on your PC. Please make sure that the appropriate video drivers for your PC are installed and fully up-to-date. We have tested this successfully on Windows XP, Vista, Windows 7 and 8.1. We expect there not to be any problems with a modern PC but if you continue to experience difficulties after updating your video driver then please contact us."


My machine appears to meet the requirements on paper so it's a bit annoying that this error occurs. I tried Dell, Intel and Windows to ensure I had the latest drivers but no joy. As requested I let them know but it would appear my only option is to use a different computer for this.

Luckily my work laptop did not have the same issue.

For my first experiment with the software I loaded up some motorbike parts.

You'll notice that some of the parts are blue and some orange. This is because the check to determine if parts are overlapping is using a bounding box rather than the actual component. It seems fairly straightforward to move the parts about but when scaling and rotating the parts you need to press Enter or Tab (rather than clicking on the next field) or the changes don't take. You can resize either by typing in new dimensions or by using the transforms, it might be nice to do this using the mouse too. I tried auto positioning these parts but it did not manage to successfully position the seven components on the bed although there was plenty of space when they had been scaled to 50%.


I failed to manage to save the layout as the menu that was supposed to allow the saving did not seem to exist, there is supposed to be a registration screen when the software is first started but that did not appear, the "preferences" button also does not seem to be available.


The "toolbar" buttons along the bottom don't seem to match the manual, with the "group" button missing, some tooltips here would have been nice too.


The right arrow moves you forward in the workflow but disappears if the printer is not ready e.g. the filament is not fed.


The printer status screen allows you to see the temperatures of the bed, head and ambient temperature as well as jog controls for the head position, and switches for the fan and lights. There are some advanced controls on the right that allow you to open and close the needle valves on the head but you should not need to use those.


The other issue I have is that the advanced screen is a bit too well hidden. If you look in the screenshot above you will see 3 small circles between the status pane and printer picture on the right. This button opens up the diagnostic tab with the serial number etc of the printer.


The time estimates for printing seem a little optimistic, the time estimate was 2hrs 30 but the print time was 2hrs 45 although I did not do a particularly accurate timing, a follow up test seemed to be fairly accurate being just a few minutes longer than estimate.


So my first impressions are mixed, the software looks quite nicely designed but perhaps not quite as polished as it needs to be. Hopefully my missing menus will appear once the printer is attached.

Cel and Element14 have been at Electronica 2014 this month showing people Robox printer.


Photo credit doctorcdf


I'm looking forward to it arriving but have been thinking about the Roadtest.


One of the tasks we are asked to do for a Road Test is to mention what comparable products you've looked at.

In the Kickstarter project Robox suggested the following list for comparison.




Given that even the most basic printer starts at more than £500 I'd not actually considered buying a printer before as I don't have the budget. However, now that I have a printer on the way I've been looking at what it's capable of doing and what other similar printers exist. The problem is that there are so many other printers around that it's difficult to know how to compare. I decided to use the following criteria to pick some candidates for comparison.


  • Ready assembled
  • Enclosed build space
  • Can use PLA or ABS
  • Price < £1500


Based on this and the help of some 3D printing stores I came up with the following list.


  • Zortrax M200
  • Flashforge Dreamer Dual Extruders
  • Ultimaker 2
  • Cube 3


The technical specs of these printers are similar too although you do get more bang for your buck with the Robox. Specifically for the price it has a finer Z resolution than the competitors, supports a wider variety of materials. There are more expensive printers such as the Leapfrog Creatr HS that matches the resolution but that was out side of my hypothetical budget.


I don't plan to do a detailed chart comparison to these other printers but they might get mentioned again in my later blogs. If you have one of these printers let me know and we might be able to do a comparison test once the Robox arrives.

A big thanks to Element14 and CEL for letting me road test this printer, it's not arrived just yet so here's a video from the IET about the printer.


Gears, shafts and clock parts


I plan to use the CEL Robox 3D Printer to print gear wheels and components for electric clock mechanisms. The fine resolution print head should allow the detail needed to print gears, screw threads, shafts and hands and other parts. The higher print speed should allow different versions to be made quickly and for larger components to be made. I'm interested in dimentional accuracy of the print so plan to use a digital camera and scanner to test that along with some digital calipers.


Specific features of this model to test are:


  • Speed of printing
  • Quality of finish
  • Ease of deployment and packing up
  • Thin items such as the clock hands to check for warping
  • Moving parts such as bearings and concentric shafts to check for accuracy and surface finish


I've also been warned by Nate about issues with the Bowden tube coming out of the housing for which there is a known fix. I'll check that is in place and retrofit it myself if necessary.

There are also some users who have had blocked nozzles which is something I'd like to avoid.



Using the provided software Automaker is a key part of the roadtest as it controls the setup of the printer. I will likely need other software to create and manipulate models before they are printed.


OpenJsCad is a scriptable STL generator a bit like OpenSCAD but written in Javascript. I've found a gear generator written for that which looks useful.

I've had a quick look at STLView and to review files before moving them over to Automaker.



Richard Gain a 3D printer enthusiast has pointed me at some some really challenging prints. Richard printed me some flywheels the other year and I met him briefly at the IMakr London store opening.


Light bulb is a good test of printing overhangs

Temple needs good cooling when it gets to the point at the top

Double helix tests bridging capability


I've also found a nice selection of gears on Thingverse so I'll likely start with those for my test prints before moving onto creating some of my own.


There are also some 3D site which may also provide some models to prints




3D Shapes



One less obvious source is a site which is more focused on the 3D artist than the 3D printer enthusiast. However they have many files in OBJ format which should be easy to convert to STL. I also have some of my own models from the time when I was making a Flea Circus film which I may try out if I have time.



Filter Blog

By date: By tag: