Jump to content

This is *NOT* a 3D printer (not only). My new prefered desktop computer !


Recommended Posts

It is a Voron PC, powered by a RasPi 4 / 4GB.

For now, the trailer :

Front :


2x USB 3, audio output, and µSD card reader, for easy OS swap :

Rear :


1x USB 2, 1x USB C, 1x RJ45 Ethernet Giga port, 1x full size HDMI port ; and of course the mains inlet :

More to come...

This project is near to completion, but still unfinished : a real project is never completed. Started october 2022, with the design and build of the actively heated bed slinger that will print Vorons. Took 6+ months. Next post will be about this printer, dedicated to Voron builds. What you can see on the pictures is moiré. The "Voron buiilder", AKA the "Super Tornado", produces zero VFAs. Will explain how and why, and maybe debunk the BS about these artefacts (most likely invented because of the design flaws of a dated, well known and overpriced printer)

Of course, many mods I'll share later ; isn't teamfdm forum dedicated to mods ? So, let's mod !
This desktop computer is based on a LDO Voron 0.2 r1 kit...

Once it's completed, will be a 2.4/350mm :


Edited by YaaJ
Link to comment
Share on other sites

The 0.2, still unfinished, but functional. Protective sheets are still there on the acrylic panels. I don't like the hinges on the right side, and about to mirror everything...



This printer wouldn't exist without the bed slinger that was built for this purpose, and for this purpose only. The "Super Tornado". Kept the frame, the bed, and the 3 wheel X carriage. Everything else was properly redesigned and made. It will be scraped or repurposed after the two Vorons (0.2 and 2.4/350) are operational.

An early design :


The Y carriage, reinforced. 4 wheels are on eccentrics, because assuming that 3 wheels can be inline is pure nonsense.


Adjustments require a special tool :


The Z carriages. Machined from 8 mm thick aluminium (reduced to 6 mm). Double eccentrics for wheel adjustments.


Machining the carriages...


First, had to design and machine a jig...


All corners are reinforced. No cast aluminium rubish. Machined ! So I am sure the corners are all real 90° (as accurate as the machine that made them).


Also made some multithreaded M5 T nuts", from mild steel. Then, they were zinc electroplated.


Frame/Z axis junctions. Machined for a tight fit (AKA machinist fit), at true right angle.



Z motion. One motor, 2:1 reduction. The two leadscrews are suspended on a pair of preloaded deep groove bearings.


Y axis ; the tensioner :


The motor :


More to come, before the Voron 0.2 mods...

  • Like 7
Link to comment
Share on other sites

I'm sure there are many. 3D printing subs have been very active for the last 15-20 years on machinists forums.

Will post a few more pics, showing why I am retiring this crap ASAP. Yes, it prints wonderfully (prints fast with no VFA and no ringing, at 60°C), but it is a pain to use. Maybe some components will be reused, and it will  become a laser or PCB engraver...

Soon, I'll post a couple of mods I made on the zero. Mostly about cable management and cooling. After many prototypes and testing. And lots of fails of course; The most important is showing what failed, and why !

Just a picture for now ; have to take more. No adhesive tape, no DIN rail (really bad for cooling because of the added height).


I have a question : not sure about the Nevermore to make for the 0.2. LDO recommends the Nevermore Micro V5, but I see no V5 on Github. Only V5 Duo.

Is it this one : https://github.com/nevermore3d/Nevermore_Micro/tree/master/V5_Duo/V0 ? It is huge !

Or is the LDO website outdated ?

Should I make the V6 for V0 ? https://github.com/nevermore3d/Nevermore_Micro/tree/master/V6/v0




Link to comment
Share on other sites

Finally figured what parts to print, and have it working, with charcoal pellets.

The V6. Maneuvring the jalousie is extremely satisfying !!!

A few more pictures of the printer that printed the Zero, then the Zero mods.

The Z transmission, on top :


Homebrew device combining a mini-clicky probe dock, a Z endstop and a auto-Z device :


The controller (SKR 1.4 Turbo + Pi 3 B+ + a few custom PCBs for various features) :


The ghetto enclosure ; it's a pain to work with this thing (it's now falling aparrt...).


Heated chamber ! A food dehydrator recycling air ; advantages : a mechanical thermostat and a thermal fuse.



Reaches 60+ °C :


It produced this :


And is also making that :


The 0.2 will really be a (secondary) desktop computer, on a desk, with a 1024p monitor, a keyboard and a mouse.

The 2.4 is in standby (thinking of mods : balljoints, Canbus, reinforced Z tensioners, etc.)

Currently sorting the pictures showing the Zero mods...

  • Like 2
Link to comment
Share on other sites

Last mod project.

Do you see what I mean ?

A Voron Zero also can be a lamp !

(most expensive LED panel is 5.99€, least expensive is 3.99€. From the dollar shop, waiting for another project 'got a bunch of those).

To be compared to LED strips and other stick... Yes I know, it is laughtable when comparing with a LDO kit. Only real tinkerers will understand.



Edited by YaaJ
  • Like 1
Link to comment
Share on other sites

I'm a f***ng genius !

Spent this saturday tuning the 0.2. The Ellis' way, except for the squish. Don't like this concept... (if we want to push material to the bed for adhesion, it has to be done with the extruder, not with the Z motor and coords... 0.2mm has to be 0.2mm, period. Calipers against the brim or skirt, and that's it. Just play with the flow settings. The Voron parts bevels are there for elephant foot compensation (IIUC), and Cura even has a compensation for the first layer overextrusion. Cura has plenty of flow adjustments, and flow adjustments all over the place is the reason why I gave up with all other slicers. Tested Orca yesterday, searched for flow for 1st layer, top layer, walls and infill : no success, gave up instantly)

Yes, many desintoxication attempts, always came back to Cura within an hour. Orca has a much better GUI, but Cura has much more smarter settings.

Then, printed a Voron Cube. First two mm : fine. Then more and more crappy. Accel 10000mm/s², speed 100 to 150mm/s (after IS tuning), the maximum allowed by the stupid standard Revo nozzle that came with the kit (12mm3/s). 250€ for a full loaded, brass, high flow kit, not sure the hotend will not be replaced with some Phaetus, that will work with the V6 nozzles that are in stock, high flow and hardened ! Same for the 2.4.


But I digress...

As I was looking at the plastic being extruded, investigating for the issue, all the screws were running away from the two Y rails.

Just forgot to tighten and thread lock them ! As well as some other ones !

When I work on an engine or a gearbox or whatever critical, I use some paint or marker to mark what was torqued...

A genius !

(hopefully the printer will produce a better cube after re-tightening, thread locking and re-tensioning ; because for now, the homebrew bed slinger tank produces better parts. A "bit" slower, but much better. And it has to go to the cellar : can't stand it anymore, and worse : The Wife doesn't too ! Attempted to sell her 'the-printing-lamp-that-can-show-YT-videos-with-ad-blocker on the TV' : din't work...).

Tonight : another speed cube, and Ubuntu for ARM. For now, will two SD cards (thanks to the card reader extension). Later : dual boot.

Also, will start to upload to Github ; nothing revolutionar, but could be usefull to those who hate CADing.

  • Like 2
Link to comment
Share on other sites


I'm desperated... Took the gantry apart, retightened everything, etc. and still the same issue.

The cube on the right, printed at 45° in order to see if there's an issue with the A or B axis, is sh*tty (but much better than the other ones...). Had problems with the eSun roll : far too small flanges, loosy winding, I experienced a couple of tangles (the filament diving under the top ones) ; this should be solved with a different spool holdeer design (never happens on the other printer).



No idea what the problem could be. It's definitely related with the extruder, but what ??? Took it apart, nothing special, and same results.

The filament is not very consistant. But no such problem with the full size Stealthburner. Currently printing a cube with it, with the same filament spool : so far so good (except I forgot activating the part cooler). Same speeds, only the accelerations are lower (of course) [EDIT : no, the settings are more agressive on the Supe Tornado]. There are some horizontal artefacts (it's a stress test), but not to this point.

On the V 0.2, I even calibrated LA for the two speeds/accel I use (outer walls and other extrusions). No effect. Seriously thinking of the Bowden version. On such a tiny machine, it is probably worth it. It will never print stringy materials, it will be printing ABS only. It has to be a reference machine, extremely fast and accurate. Got this kit with this in mind.

I'd prefer be designing, showing and sharing mods, than attempting to get decent prints... Definitely the zero is not an easy build.






Most beautifull Voron cube ever. See below. Left, the previous crappy cube. Center, the cube produced by the "Super Tornado" (48mn) ; we can see where the part fan was turned on, and it was insufficient. Right the cube after the fix, with the same gcode than the left one (51mn). Of course, artefacts are exagerated by the lightings.




BTW, about the (in)famous VFAs... There's a bloatload of BS about VFAs all over the internets. VFAs have to see with the machine rigidity. The fact than an east european printer manufacturer is gaslighting everyone changes nothing. If their machines were rigid, they wouldn't make VFAs. Does a Tevo Tornado / CR10 produce VFAs ? Yes ! A lot ! Does a massively rigidified one does ? Not at all. Does a Voron Zero produces VFAs ? Not even the tiniest one. VFAs have a structural origin, and nothing to do with 400 or 200 steps morors. All motors on the Super Tornado come from Amazon, unbranded or from some obscure brand, all are 200 steps, the extruder is a Stealthburner made from a old, used Trianglelab BMG clone.

Now testing this. Because why not.


Works fine :


Let's test the same, but upside down...


Edited by YaaJ
new pictures
  • Like 2
Link to comment
Share on other sites

Forgot to say.

The fix.

Just retightened the Revo nozzle before printing. It prints much better. This nozzle *very* easily moves when cleaning it with a brush (I use a nylon toothbrush). Move the brush so it doesn't unscrew the nozzle ! I wouldn't consider this system as reliable. Never had a loose nozzle until now ! Or maybe is it a lemon ?

And don't print upside down. The bed + the Kirigami are stronger than the anti backlash spring. It will print... into the PEI coating. And later, good luck for removing the print. Ask me how I know.

(please don't make fun of me ; I do stupid things so you don't have to)

  • Like 2
Link to comment
Share on other sites

While you don't have to crank it down tight, you do have to firmly snug down a Revo nozzle. I have been running it on my Trident for over a year-and-a-half, and run a nozzle scrub every time. I've never had a nozzle loosen on me.

That cube on its corner... wow! 😮

Oh, and we've all printed into the PEI. It's a rite of passage.

Link to comment
Share on other sites

Whenever possible, I've been using adhesive spray on sanded bare stainless. because of PETG, and now ABS. First time I used PEI, it came with the part. It was advertised : for PETG !!! Smooth PEI. As a result, I didn't reuse PEI until it was mandatory for some Voron parts. And I slowly killed another one ; even had to create a custom STL bed for Cura, showing where not to put the models.

Definitely wasn't tightening enough : it didn't move again (or maybe never tightend it).

Anyway, the Revo's will be replaced with a Dragon or a Dragonfly on the Zero, so I can reuse the Nozzle X's and the BMG CHT's, and for the 2.4 it will be a Rapido. Will stick to the old nozzle standard ; plenty of V6 and Volcano.

The cube on the corner : so cool not having to worry about the bed going back and forth ! Had to try. 100% success. We can read everywhere that it is a huge CoreXY and deltas advantage for tall parts : had to try. Will be good for the pin support challenge. Will also try the Maker's Muse lattice cube.

Printing at max flow allowed by the standard Revo (11 mm3), 140 mm/s, 10000 mm/s². Way below what this printer can do. Just watched an old 247printing video. Have to experiment with realy fast inner walls and infills. The goal being a sub 15mn quality benchy. (I remember my first benchy took 2 hours and 15 mn with the stock Tornado and Marlin firmware that came with, 5 or 6 years ago).

I'll try to share one or two designs tonight on Github. But there's always one more project on the drawing board !

Edited by YaaJ
Link to comment
Share on other sites

Made a strain relief for the Picobilical. It also solves mitigates the problem with the reverse Bowden rubbing against the tophat. But adds a bit more bend to the harness... (Teflon insulated wires with the LDO kit). The radius looks similar to a cable chain, I'm not too worried. Here's an early itteration, made in a hurry, because there was a lot of stress on the connector. It attaches on the Z extrusions. Would be nicer on the motors extrusions, but didn't want to mess with the micro M3 T nuts. Will have to make something similar for the hotend because it is a bit too short and a zip tie does not fit xell.


Printed at 45°, for better mechanical properties.


The Zero became a desktop computer today. Still no OS nor dual boot, but it will come one day or the other. For now it prints.


The Super Tornado enclosure is falling apart. And the Zero produces better prints :  currently reprinting the 2.4 parts that were already printed


The third iteration of the stain relief. The second was not printed.


Another mod : the door ; I don't like the door opening to the right ; yes, it can be removed, but the hinges being screws, they will act as little files on the long term. And I don't want the door on the way while tinkering with the spool. But we have a problem : the LDO door is special. It has two 6mm diam recesses, 2mm deep, for the magnets (and how is supposed the handle to be glued ?). As a result, it is asymetric and chiral. The solution : gently drilling at 3mm, and replacing the magnets with M3 contersunks. Then redesigning the hadle for inserts, and the latches so the countersunks get in contact with the magnets (not strong enough if inside the plastic).

Works wonderfully !



The Nevermore was printed on the bed slinger. Works like a charm. Love the jalousie. Thinking it could be better I also printed the jalousie with PETG. It broke instantly. Not the ABS one (in fact I broke one because I stress tested it ; reprinted two, with a spare, just in case).

Had to re-tram the bed and re-tune the Z offset a couple of times. Either the Kirigami is not that rigid, or the Z endstop tends to move, or parts are just settling... Some thoughts ?

This printer desktop computer is amazing !


Edited by YaaJ
  • Like 4
Link to comment
Share on other sites

Remixed the LDO toolhead strain relief so it also holds the wires : didn't like the constraints on the crimps. Not 100% happy, it looks a bit fragile, and more material could allow for a support free model.


Will not be able to design for some time : will be abroad.

  • Like 1
Link to comment
Share on other sites

  • 2 weeks later...

Not much mods on the Zero these days. It's printing parts for the 2.4. The bed slinger hasn't been used for a while, but still by hand just in case a large part is needed...

Made squash ball feet. I find the design cool (original design)... One advantage is that these feet can be mounted on a stock Zero : they just replace the "foot accents"



They make the printer quiet, but two caveats :

- input shaping doesn't work perfectly anymore, as the printer is shaking, and the machine accelerations pile up with those from the head... Thinking of measuring the bed accelerations, and substracting the numbers from the accelerations measures at the head (playing with csv files). Does it make sense ? No idea.

- the center of gravity is not centered, (mostly because of the spool holder, and it looks like the Pisa tower !

Not sure : maybe some VFA (more testing and comparisons needed).

A PiCam is on the way (next mod).

  • Like 2
Link to comment
Share on other sites

Just created a secondary Git on Github for sharing Voron related designs.

Called it CachecticLab.

BTW, discovered one interesting thing. A thing that probably everybody knows. I made a rear plate with two fans, on top, near to the A/B steppers. Spinning the fans makes, even as low as 20%, and the chamber tepemrature drastically drops. As a result, they are now unused. The Pi 4 has one of these large heatsinks, mounted passive. Didn't aven mount the 30mm fan. Will remove the fans, and just redesign the rear plate with vents and no fans. The Super Crappy Tornado will make it.

chamber : 65°C
Picobilical MCU : 64°C
Raspberry : 56°C
SKR Pico : 63°C

Far from throttling. But this says nothing about the stepper drivers temps (pushed a bit the currents, as I was getting false endstop triggerings, and also layer shifts with the LDO default parameters). It seems everything is working great. Except one of the belts is spitting black particles. Will reprint the A and B drive frames, something was wrong with the Tornado prints (warped ?), and is out of alignment.

Have some ideas for replacing the M3 bearings/pulleys axes with G6 rods : have some, salvaged from a paper printer. Just a bit of machining (just threading with a M3 die) to turn them into studs. (hate screws as axes : they don't and can't comply with the tolerances usually required for bearings with fixed axe and spinning load : G6+H7, the bearings being H7 ID, and axe G6. IIRC).


Link to comment
Share on other sites

Today, an idea... A bit tired with the Zero shaking and dancing. Made adapters for the feet that came with the LDO kit. We get an extended support polygon, more stability. And swapping between squash balls and stock feet takes less than one minute...




  • Like 1
Link to comment
Share on other sites

This camera bracket was an epic fail...


Made another one :


Centering is not too bad ; of course, it will need better lightings, and the part being printed often is hidden by the X beam... But this placement offers a better depth of field for the same focus, compared with the usual one (horizontal). Also been thinking of another placement and a mirror. Overcomplicated...


This tiny machine did a great work for the 2.4, printing cardridges for all my old V6 toolheads, high flow and hardened !


Also added a FPC connector for the accelerometer.


Currently designing an enclosure for a 4"3 LCD. BTT and Waveshare will be supported. Real enclosure, not one of these "PCB holders". For the Zero, but also for the 2.4 (I don't like the things I found on Github and Printables)




Also working on the lightings. For this project, unlike the LCD, found great ideas on Printables and Github : mpbiv421 and Mapple Leaf. Will be a remix combining the two designs. Also thinking of CANBUS, with a dedicated custom Raspberry hat. Will do it if Klipper supports SPI -> CANBUS. I found no reliable information (or no information I clearly understand). SPI being much more capable than the serial interface (already played with the Pi SPI, up to 24 MHz ; C, bare metal, Pi <-> STM32, not Python ; no idea what the limitations could be with Python/Klipper, and is it supported ? serial is desapointing, never got better than 250 kbps, OctoPrint and Klipper)


The BTT U2C is too large. And I'd like not to use a USB port.

Edited by YaaJ
  • Like 2
Link to comment
Share on other sites

The capacitive touchscreen is now in use. The BTT one (that came with the 2.4 kit) is unusable because the flat cable would make the enclosure too large. The screen is mounted on a pair of hinges, in order to accomodate various feet heights.




Two inserts for an adjustable foot (todo...) :





This Waveshare 4"3 LCD is really crisp, and high quality, and the capacitive touch excellent. BTT and Waveshare only differ by the electronics, the BTT having a brighness potentiometer. I'm sure the displays are the same.

Now working on the lightings, based on four V3D Daylight Neo Mini. These LED sticks have two problems :

- they are a bit too long (a few mm) for the bed lighting ; but if those are the last ones in the chain, the PCB can be trimmed (I don't want to take the printer apart - for now, cut the back pannel, and make them protrude in the electronics compartment)

- their pinout is the opposite of the Picobilical. Will have to unsolder the JST connector, and solder it on the other side (on components side). Luckily, the Kicad project is open sourced, and I just had to modify the schematics, the PCB layout, and finally export the STEP model. Then a diffuser was easy to create in CAD, massively inspired by mpbiv421 work. Work in progress, no images for now. The JST connectors have to be flipped anyway : see here (no explanations, but it's trivial) : https://www.printables.com/fr/model/389682-voron-v02-v3d-daybreak-neo-mini-led-mount

Daylight Neo Mini, modded in Kicad :


Will be attached using clips ; these ones are too short by 2 mm and the diffuser would interfere with the XY joints. Also, the diffusers are too long for the 0.2 😞

Why didn't V3D made these sticks just a bit shorter ??? 7 LEDs instead of 8 for example... Or just make their spacing a bit smaller ( 7 x 2mm would be just fine)


Edited by YaaJ
Link to comment
Share on other sites

  • 1 month later...
On 9/19/2023 at 6:04 PM, YaaJ said:

Made a strain relief for the Picobilical. It also solves mitigates the problem with the reverse Bowden rubbing against the tophat. But adds a bit more bend to the harness... (Teflon insulated wires with the LDO kit). The radius looks similar to a cable chain, I'm not too worried. Here's an early itteration, made in a hurry, because there was a lot of stress on the connector. It attaches on the Z extrusions. Would be nicer on the motors extrusions, but didn't want to mess with the micro M3 T nuts. Will have to make something similar for the hotend because it is a bit too short and a zip tie does not fit xell.


Printed at 45°, for better mechanical properties.



Have you made the .stl for this and your hotend strain relief available anywhere? I'd love to print a set for my own use.

On 9/19/2023 at 6:04 PM, YaaJ said:


Link to comment
Share on other sites

I didn't share this design because I don't use it anymore. It looks like it adds strain elsewhere : the umbilical has a shorter radius with it when the head is at Y = 120, as the cables cannot go to the back. (no sure my explanation is clear).

Moreover, I will soon CANBUS the machine : would require a slightly different design.

Here's the STL. It's a quick print (1 hour or less), it's easy to install. You can try, but I think it is worthless. It was a bad idea. I sincerely think that no strain relief is better, any strain relief will add strain somewhere else. The other end of the cable, or the cable itself. It was used maybe 50 hours, the printer being 211 hours "old".

To be printed at 45°

Same with the squash balls. But will be reworked for an extended base and more stability, without the balls. Currently on the 2.4, out of time for the Zero ! Two months and still designing mods, with kilograms of failed or flawed designs !!!

Picobilical Strain Relief v2.STL


Empty boxes of filament :


and the v 2.4 mods coffin :



Edited by YaaJ
  • Like 3
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...