​It has been a while since I worked on the Pi Panel Skirt for my Voron printer build, but I'm closing in on when I may actually need to use it.  The design has not changed much on the outside, since I originally posted about it back in October, but there have been quite a few changes under the hood recently.  I printed a first version and decided that if I changed the SD holder to a design that removes the stock plastic shell similar to this, I could get some more area to add a shutdown button for the Pi (safe shutdown button).  The pics below are before I re-worked it (yet again).  I know, the prints are not good, since I was using PETG which I did not dial in well, and a 0.6mm nozzle, where I should be using PLA and a 0.4mm nozzle due to the fine details.  It's just a prototype though, and it did prove the fit was good on some things and not so good on others.  In any case, it's moving along and I have some PLA on the way which will hopefully match the rest of the red on my printer, which is Sunlu ABS in red (which is a muted red, and I guess has a matte finish on the stuff they sell in China).    

This is the newest version, which is still needing to be printed.  In this version, there are two general purpose LEDs on the left, a disk activity LED for the Pi (shaped like an SD card) and the raspberry also will have a LED backlight for the Pi power status.  The power button is the tic-tac shaped thing in black which I sort'a designed so it would look like the power/reset symbol.  There is also a LED ring light for the USB ports.  I am not sure yet if the design for the SD port will change, since I may go with something simper if I need to add a black masking layer in the print.  I have a feeling there will be some bleed through unless I switch to black after a few layers.  There is a wire management cover on the back which will hopefully keep the chaos under some degree of control for all the LEDs and switch wires.  I plan to route then back along the USB cables to the Pi.  I also plan a version for folks that want to add a 0.6mm hexagonal grill to the skirt.

Update 4/29/2023: After printing the above model, I decided that some changes were needed.  Most of the work required (of which there was quite a bit), is below the surface, however some cosmetic changes were made as well.  I dropped the inset power/reset symbol, and went with a simpler and cleaner looking printed inset layer for that, and the same for the two function LED's on the left side.  The SD holder is also updated to print easier, and I think it looks cleaner as well.  Overall the design is much less cluttered and simplified.  Below the surface, I made a few light diffusers for the Pi logo, using either Lilly pad LED or a WS2812B LED chip (which of course I now have to wait for).  There were also a bunch of changes to the SD card holder, and the cable management cover. 

The BOM so far is as follows:

• (qty 8) M3 4.2mm (dia) x 3mm Brass inserts.
• (qty 1) USB 3.0 dual panel mount extension (this is the side by side blue one which is available in 30 or 50cm lengths)
• (qty 1) 25cm MicroSD female to male extension
• (qty 4-5) LilyPad LED modules in multiple colors
• Wires and dupont female connectors and crimps
• (qty 1) Optional WS2812B LED chip 
• Superglue Gel (I use Gorilla SuperGlue Gel which works really well)
• I may use an old floppy disk ribbon cable to keep the LED wires simple.
• 6x3.5x4mm Momentary Push Button Switches (I used the switches from this assortment).
• Black Sharpie (optional, but may help to mask out stray photons)

I'm still working on this and again waiting on the WS2812B, but that won't stop me from testing this with the LilyPad LED's.  What will stop me from working on it, is... work.  I will probably get back to this next week.  So far though, the design looks to be going together as planned, and everything at least fits.  Printing it will require a properly calibrated printer (esteps, flow and possibly horizontal expansion compensation, with dry filament).  Although I designed in a good margin for fitting the small parts where possible, there is a lot crammed into a small space, so things will be tight.  There is also a mild bit of hacking to extract the captive M3 nuts in the USB extension, and I leaned the hard way to drill them out from the back side and not the front, but they should come out without too much trouble.  I plan to glue the USB extension in for some extra security, once everything else is in place.

Some of the pics below are earlier prototypes, the later pics are close to the final version.  This is not yet posted (I need to test it first).

4/30/2023:  This is probably all the progress I can show for a while, waiting on some WS2812B LED boards from AliExpress.  The testing today did show a few issues which I think have been resolved.  The first is that the Pi logo will not be adequately illuminated by one Lilypad LED, but the WS2712B should have no problem with that.  I went ahead though and made an insert to allow using two Lilypad LED's to light up the Pi Logo, which should be enough.  I'm pretty happy with the way the rest of the lights will illuminate though.  The pics below were done using an orange LED which was the only one I had handy (also waiting on some more Lilypad LED's in green).  I'm not sure which colors I will use, and if maybe I should use one color (red or blue), or if I should try different colors.  I'm also not sure which two functions will be mapped to the LEDs on the left side (which may determine their color).  

I also spent some time cleaning up and optimizing a wire management method for this part.  I really hate to see wires everywhere, and so there will be a three part wire management module.  The module will secure the LED's (which will also need to be glued in), give some space to manage the wires, and finally cover the whole mess up, while also adding some extra security to the USB ports.

Update 5/18/2023: I recieved the WS2812B LEDs and also some LilyPad momentary switches, which I think would work better than the tiny switch I am using in the current design.  The LilyPad switches are more, but I think for reliability and being easier to assemble, they are probably worth it.

Update 5/20/2023: The updated design with the LilyPad momentary switch is ready to have a prototype printed to test.  My printer is tied up at the moment though, with yet another part for the Voron.  I was able to assemble some earlier prototype parts I previously printed (to test the fit only), and made some changes to the current model, which will help with assembly.  The good news is that there were no show stoppers, and the assembly went pretty smoothly.  I won't be using the test model though, since it is an earlier revision printed prior to changing to the LilyPad button.

Update 6/2/2023: I feel the design is probably done, however my printers have been tied up on other projects and there have been other life related issues going on so the Pi Panel has not yet been test printed.  It is not forgotten though, I am gonna get it printed as soon as possible (but after the Voron Cable Management Duct Remix is completed since I need that to move my Voron build along).

Update 8/3/2023: I have not tested the design, but it has been way too long and so I posted the latest version (which I think will work).  You can read about that here.

I finally got to the point where I could almost see the end of the tunnel with my Voron build, and then started down the rabbit hole of how to get my Mellow Fly SB-2040 (version 1) hooked up.  I thought I was starting to figure it out, and then stumbled on this thread on their Github.  In addition to the issue in that thread and possible fix (hook up an external 5V buck converter), it was also interesting to realize there is a version 2 of this board.  I could not bring up the schematic but the CAD file shows some changes in the area where the buck components are located (so it would make sense if they updated the buck converter design).

The buck converter is required to provide a 5V source on the board, and it's output is what the processors (after being stepped down again to 3.3V) and the LED's use, since they cannot take 24V directly.  Since the board takes 24V as input, it has to first drop that down to 5V using a buck converter, and then though a less efficient linear regulator to drop 5V to 3.3V for the Pi.  When a voltage is dropped using a buck converter, some of the energy is wasted (heat), but buck converters are still the efficient way to lower a higher voltage to a lower one, when there is a significant difference between the two.

I am not an electronics expert, and barely a hobbyist, but ​several things I've read make me question if the buck converter is implemented well on the V1 board.  I decided to see if I could compare a reference design to what is used on the V1.  I found something even better than a reference design though, and used TI's (really cool and free) webench tool to spec out a BOM and reference design using the same TI TPS5430DDAR (datasheet here).  Using the webench, I set the input voltage range between 23-25V (and I also tried 20-30V and even 10.8-13.2V), and generated several BOMs and schematics.  Despite the input voltage range being different on several runs, the component values I stayed mostly the similar, except for the input capacitor value, which changed depending on the input voltage. 

When I compared the values from the webench tool to the Schematic for the SB-2040 Version 1, I saw that the Fly SB-2040 V1 designer used different values for several of the key components (compared to the webench).  There may be a reasonable explanation for this, which I just don't understand, so I'm not calling this out as a problem, just something interesting.  The input capacitor looks to be 4.7uF on the SB-2040 V1, which is the same as what I found when running the webench tool with a input range of 22-26V (which is good).  However the diode, inductor and output capacitor value or type were different than what was specified in the webench tool.  

For example, when I ran webench using 22v-26v input (that is probably too tight a range but I was looking for the best case), and compared them to the SB-2040 V1 schematic, I noticed these differences:

D1
Webench shows a Schottky diode, while the SB-2040 V1 uses a standard diode (IN4007)

L1
Webench shows a 15uH inductor, while the SB-2040 V1 uses as 10uH inductor

Cout / E1
Webench shows a 330uF 6.3V capacitor, while the SB-2040 V1 uses a 220uF 6.3V capacitor

Like I said, I am not expert, but this does make me wonder, and so I plan to check the board with a thermal camera to see if the buck converter looks to be heating up (I read that may happen).  If it does appear to be heating excessively, I will either re-consider using this board, or possibly try swapping some of the components with values which better fit the values from the webench tool.  I'm also wondering (in case of an issue) if swapping the slower standard diode with a schottky diode may be worth trying first. Aside from that, if there is a problem I may just switch to a different CANBUS board, since for the price it may not be worth the effort of swapping several components (since I suck at soldering small stuff).

I did not trace the parts in the pics out below, but the highlighted parts look to be the Diode (red), Inductor (purple) and 220uF Capacitor (green).  The TI TPS5430 is in the orange color box.

Update - I went looking to see if I could find some components to maybe have on hand if I want to try some swaps, and noticed that if I increased the temp in the webench, still using 22-26V as input, I did get a BOM using a 220uF output capacitor.  The diode and inductor were still possible candidates to swap.  I cannot seem to find a Schottky diode in a SOD123 package with similar specs though, the highest current version I can find is 3A, which is also the max of the 5430.  Finding a 15uH inductor that would fit looks like it will be a pain as well.

It looks like the version 2 uses a different buck converter so that is a possibility to swap the board, or maybe the BTT CANBUS board which looks similar (designed for the Stealthburner).  Given the well known issue of heat and my new understanding of the root of the problem thanks to rbrtwtrs at Github, I don't really think it's worthwhile to use the V1 board, and will look for another option, the heat gremlin is one thing I'd rather not have to deal with.

Update 2 -  I read of other issues with the SB-2040 implementation such as possible driver issues with SK6812 LEDs (I guess some LED strings being sold for the Stealthburner use SK6812 LEDs and not the standard WS2812 LEDs, and some folks have had issues).  The driver issue could be fixed though, as it's software.  Due to the heating issue, which seems well understood in the community to be related to the buck converter on the V1, I've decided not to continue with the Mellow Fly SB-2040 Version 1 board.  I ordered a BTT EBB SB2209, which will be here in two weeks (around the second week of May).  I could not find anyone selling the SB-2040 Version 2, which looks like it could have saved me some time, but I feel a bit like BTT may be a safer bet anyway, since I have not read of many serious issues with it, and it seems more polished. 

I bought an all metal Z-axis tensioner for the Voron 2.4 from AliExpress, not knowing at the time it was a knock off of the Chaoticlabs design (I should have bought it directly from ChaoticLabs which may also have saved me a bit of trouble setting it up). 

So this is NOT a post about the official ChaoticLabs Z-axis Tensioner, but a review of the AliExpress "Funssor" version.  When I received the parts, I thought they looked OK and were well made.  However when I went to swap the printed Z-axis tensioners, I ran into the first problem.  The shoulder screws which were included did not fit the stock GT2 20 tooth (9mm) idlers.  I measured them and found they were about 4.97mm in diameter so they should have fit, however when I checked the point where the threads stop and the shoulder begins, I noticed a very slight bump in the profile of the screw (under a microscope).  

​I ran the screws in my drill and used a small hand file to bring down this area and then used some fine grit sandpaper to clean it up, which fixed that issue.

The second issue which is gonna require parts, is that the idler will grind to a halt if I snug down the shoulder screw that holds it in place, even just a bit.  I was able to reduce this problem if I added another M5x8x1 flat washer to the stack, however they only include four of these washers in the kit (so I would need four more).  I think some blue threadlocker is a must on this as well, since even with the extra flat washer, the screw cannot be tightened down due to the design of the tensioner.  One trick to use when assembling the idler into the part, is to stick the edge of the washer to the idler using some tape, and then when the screw is started into the threads, pull the tape out.  Otherwise getting the inner flat washer installed is a bit of a pain.  

So I am again waiting on parts and re-installed my printed Z-axis tensioner, which will work just fine.  The advantage to the Chaoticlab design is the knob, which will allow much simpler adjustment of the tension.  They have made the design files available for printing as well.  I'd suggest if you plan to buy these, try the official version from Chaoticlab instead of saving a couple bucks on the AliExpress knock off like I did, since it can't be worse than the knock off, and is probably higher quality.  I'd suggest picking up some extra M5x8x1 flat washers in any case, since they could be handy and are easy to lose during assembly.

It's gonna be a while before I get the washers, so will update when these are finally installed.

Update 5/4/2023: I received the washers and installed the tensioners and have mixed feelings about them.

Pros:

• Aluminum construction should be stronger than printed
• The knob is much easier to use than tensioning using an allen wrench

Cons:

• The construction is not great, the shoulder bolts required filing, and the knob feels crunchy.
• Required ordering washers, which only helped somewhat.
• I'm not a fan of the way the pulley is held in.  The shoulder bolt which holds the pulley cannot be fully tightened without seizing the pulley.  The retaining bolt is also somewhat cantilevered, since one side is only supported by the bolt head.  They could have just used the same design to retain the pulley as stock, and it would have been better. 

Overall I would not recommend these.  They look nice, but for the cost, they should operate smoothly and installation should not be such a pain.  Since I was unable to fully tighten the bolts for the pulleys, I have little doubt that I will need to re-tighten them regularly or design some other solution.  Again however, it should be noted that these were AliExpress knock offs, and are not Chaoticlabs branded tensioners, so I do not know if the genuine article would improve on the problems I experienced.

Update: After reading about the heating issues with the SB2040 V1 board, I decided to go in a different direction and will be using the BTT EBB SB2209 CANbus board.  You can read more about why, here.  I was going to abandon this project but found that it could be useful for the BTT SB2209 as well since it gives much better access to the thermistor and other wires, versus the stock clockwork 2.  If you want to see the latest iteration, just skip to the end of this post below.

I bought a Mellow Fly-SB2040 board and the FLY-UTOC-3 to go with it (I think I should have bought the UTOC-1 though).  But why buy another set of boards for the Voron, when it already comes with the Pi and the Spider mainboard?  These two boards will allow me to reduce the number of wires going from the Spider mainboard to the toolhead by using them as a  CANBUS bridge to communicate and control the components on the toolhead.  This means the UTOC-3 (or UTOC-1) board will be located in the electronics area under the printer, and will connect to the Pi via USB.  There will be only 4 wires (24v, Ground, and 2 CANBUS signal wires) between the UTOC and the SB2040 board located on the Stealthburner.  This cuts out all the stepper, fan, LED and hot end wires, and will make cable routing simpler and more resilient.  The SB2040 board will break out the individual components like fans, thermistors, hot end and LED's right on the Stealthburner tool head.  

I have read however that there are some issues with the SB2040, including some unprotected (non level shifted) inputs for the Pi controller on the Mellow board, which could result in a magic smoke release if 5V is sent to a Pi input on some pins.  What that means is that I cannot use my tappety tap tap without putting it's light sensor on the so called "HV Endstop" (aka GPIO 25), which is also helpfully called out in the listing for the SB2040 on AliExpress now.  Not a big deal, except that the Voron Stealthburner is very tight in the area where the "HV Endstop" connecter is located when the SB2040 is installed, meaning there could be issues with wiring.  There is also a potential issue with either the Pi rp2040 chip, or the TMC2209 stepper driver chip overheating in some conditions, even with heatsinks.  Fortunately for me, and anyone else using the Fly-SB2040,  Printables user Paul Trautner has posted a cool solution.  I wanted to add a fan guard to his design, and while I was at it, opened up the fan bump out just a bit to allow fewer supports and possibly some more room for wiring.  The fan grill also acts as a fan clip.  I'm printing the parts now, and will use them when I assemble my Stealthburner, which uses the Tap Probe Mod, a Mellow SB2040 board, and maybe a Rapido hot end.  If things work out, the modified files will be posted to Printables as a remix of Paul Trautner's design.  

Update 4/19/2023: I printed the parts and they appear to fit well (which is a credit to the original design).  The only part of the remixed design I am not happy with is the fan clip/grill.  due to the very small holes, the posts on the clip are is too tight to go through the holes in the fan.  I'm either going to drill the holes in the fan or will need to figure out a way to print the posts even thinner (which seems like a bad idea).  I am probably just gonna widen the mounting holes in the fan.  I'm waiting on a 30x7mm fan though, so I won't be posting anything until I receive the parts and test this out.  The parts were printed in PETG, since I don't have any ABS on hand (and don't like printing the stuff inside anyway).  I should have calibrated things better since my first layer looks under extruded and there was a good bit of stringing as well.

Update 4/20/2023:  I found a 24v 3010 fan, and it's a ball bearing fan as well, so I'm gonna try and use it.  I drilled the mounting holes out with a 1/8" drill bit and it fits the cover, but I need to make a small adjustment to the grill cover before I can post everything.

Update 5/4/2023: I decided to go with a BTT EBB SB2209, but I am gonna modify this design to fit that board as well (will fit both the SB2209 and SB2040), and am printing a model to test it now.   


Update 5/5/2023: The new remix is completed, though I cannot test it with a 3007 fan. I tried the 3010 fan and it is very tight, and may not give enough clearance for wires.  The print is not the best but it fits everything very well and I think it is ready to post when I get the time to write it up.  The changes made to fit the BTT SB2209 board were very minor, but necessary since the original design was made for the Mellow SB2040 board, which is a bit smaller.  The fan shroud is also remixed but with very minor changes to allow for a fan guard and printing without any supports (if your printer is able to handle some overhangs on the hole for the fan).  These parts were printed in Monoprice ABS+, which aside from requiring a very low flow (81.5%), it printed very easily and there were no lifting issues when using the "Tab Anti Warping" plugin for Cura.

I also switched out the gears for some helical gears which just arrived.  The method I use to align it, is to install the gear loosely on the main shaft, but with the set screw tightened just enough that it can move a bit but cannot rotate.  I added a very small amount of blue threadlocker as well.  Then, with a filament installed through the stepper, I removed the tensioning arm and used a long allen wrench to set the grub screw in the correct position.  Having the filament going through sort'a helps to see where the gear should be.  Once it was tight, I just re-installed the red tensioning arm and set the max travel set screw.  I did manage to over stress the latch though, it has a stress mark on it and I will probably need to swap it out soon.

Update 5/14/2023: The remix of the CW2 cover which will accommodate the BTT SB2209 has been posted to Printables here.  It will use the same covers as in the original model it was remixed from, those can be found here.  I did not post the fan cover since I could not figure out a good way of mounting the grill without screws or melting the posts to hold the fan.  If that changes and I can make something that I think will work, I may post it as a remix as well (I'm open to ideas). 

Update 6/3/2023: I uploaded a fanless remix of Paul Trautner's original design to Printables.  It only has some minor changes where I removed some holes to make it easier to print.

www.printables.com/model/497525-minor-remix-of-paul-trautners-mellow-sb2040-stealt

It started with a clog, and ended up with a new sapphire MK8 0.6mm nozzle, a new copper heat block, a new BMG clone extruder (because I broke the arm on the old one when I dropped it), a new 50 tooth POM gear for the BMG, a new cartridge thermistor, a new 40W heater (because the old one was stuck in the block), and a repaired part here and there.  The POM BMG gear did require some work to fit though, aside from shims which help to center the filament grove in the gear, I had to drill out the cover of the old BMG extruder, since the shaft on the new gear was too long by a couple mm.  Instead of re-running the thermistor wires, I spliced the new one in using some of these heat shrink butt connectors, which have a bead of low temp solder which is melted by the heat gun (really cool).  I should also say that, aside from the BMG and thermistor, I had been collecting the parts for a while, and just needed an excuse to install them.

Now that it's back together, I did the following:

1. upgraded the firmware
2. reset the probe offsets and reset the z-offset (M851 X0.00 Y-56.00 Z-1.80)
3. roughly set the E-steps back to what they were (M92 E421.00)
4. ran a PID auto tune on the nozzle
5. ran another e-steps calibration and got the same value as originally (421.00)
6. Re-leveled the gantry using some blocks I previously printed
7. Performed the aux leveling (using the menu and adjusting the corner wheels)
8. Ran an ABL
9. Ran the "validate mesh" program from the "Aux Leveling" screen (was almost  perfect)
10. Saved the config using M500 on the Octoprint terminal window.

I almost went all out and installed the belted Z-axis remix I was working on (and printed) back in September, but I didn't have the time, and need this printer back up.  Once the Voron is done and working, I can do that upgrade.

One of my earbuds fell apart today just as I was gonna go for a run, fortunately the other one worked, so I didn't miss out on Runicorn's words of wisdom, or my audiobook, but it still sucks that most wired earbuds I've had don't seem to last.  These are only mediocre for audio quality, but they have been waterproof mostly.  When I got home, I took a look to see how bad the damage was, and decided to fix 'em.  Most speakers use a braided thin wire with a lacquer or something, which both identifies the wire and insulates it, so the first step was to try and identify which wire was previously connected where.  Fortunately the red/gold color wire left a reddish trace on the pad it was connected to, while the blue wire did not seem to leave any traces.  My $30 USB microscope was a huge help with checking out the board and wires, but the new problem was how to hold everything in place so I could solder the wires back.  That's when I went looking for my silly putty.  Failing to find it, I recalled that I had something better, Rodico 6033 Cleaning Putty, which is sold as a watch cleaning tool, but has a ton of other uses.  It worked perfectly as a third hand for soldering, and I probably would have blown my monthly cursing budget without it.  The pics speak for themselves, I'm adding a small ball of this stuff to my soldering kit.  I am not sure how it handles heat though, but for tacking down a board or wire (well away from the heat), I think it will work.