-
Amazing project! Would love to get my hands on one of these eventually if you plan to scale up and sell them! -
-
Just wanted to drop a quick update, in case anyone is wondering where things are at. I haven't dropped the project, but it is on hold for now - the core processor I am using (and a number of other parts, but the biggest deal is the main CPU) are totally out of stock due to the chip shortage (32+ week lead times at the moment). I can't afford the time to radically change the design to use a new main processor, so for now.... she is paused. When parts become available again to mere mortals like myself (big companies always get priority) I will try and do an actual prototype production run.Leave a comment:
-
Quick little update today - nothing really exciting to show off. Smaller stepper board is done and (mostly) programmed, wasn't too bad. I do think if I want to go with ultra smooth motion I will need to make some additional changes and spend a lot more time screwing with firmware, but really as it is it is very good and certainly where I am going to leave it for now.
Still need to finish off the updated electrical for one more board, which I have been dragging my feet on. Been very busy with unrelated stuff lately, as is usually the case, but I think in the next couple weeks or so my schedule should be back to normal and I can put in the final push to get the prototypes running.Leave a comment:
-
Figured I would put together a brief update today. Nothing really too exciting - firmware is moving along, updated stepper board/small gauge is in hand and parts should arrive shortly so I can get that all dialed in. New drive strategy should give me really fine grained control over the motor, and basically zero noise (I found a few whitepapers on the exact motors I am using, basically someone did a bunch of math to get things perfect and I just need to implement it in software). Almost ready to test the back plane in the actual car - but naturally I ran into (what I think is) a fuel pump issue over the weekend so I have been kinda tied up with that. Realistically it isn't stopping me from doing some basic testing but I do need to get it sorted.
In terms of mechanicals, things have moved along a bit. Final 'draft' of the enclosure and pretty much all the pieces is together (and printed). These are just quick and dirty 'fast' prints to check things like mounting hole location and whatnot, but you can start to see the overall design taking shape. Size wise it is just a touch smaller than the factory cluster, so a little easier to jam in behind the steering wheel :). The smaller size will be made up for by a updated ring piece that essentially matches factory dimensions, but this will be designed last as I need to figure out a window/cover setup.
With the mechanical layout I am quite happy. Currently I think I need a very slight angle adjustment (the upper arms need to be a touch shorter) but otherwise it is basically where I want it. I will likely leave the slightly different curve shape as it is, but I am also toying with the idea of basically copying the factory shape almost exactly. Either way, any gaps/differences will be made up with the ring piece, which will replace the factory item and also cover the window component. The biggest difference to the stock cluster (IMO) in terms of looks at the moment is that the gauges are not angled 'inwards', i.e. tilted slightly around the vertical center axis of the cluster, so they are a little more flat. I don't know if this is going to be an issue or a feature - I have yet to make my mind up. I can certainly tilt the gauges but this will require adding cost because the daughter boards will need to be connected via cables, which is another set of 2 connectors for every dial. For now I will likely leave it as is, and maybe at some point in the future (if there is enough interest) offer a slightly different front enclosure and cable set so that the tilt can be emulated.
In terms of electrical work, I need to make a mildly modified version of the small stepper board to allow for an econometer. This really isn't a big deal; in terms of electrical/schematic design it will be identical to the small 52mm board (which already supports dual motors), and just have a slightly different layout, but I may also leave this out of the picture for the prototypes as it is another item to bring in, assemble and test. Additionally, it is easy enough to add it to a prototype cluster - simply replace the translucent dial backing piece (and faceplate) with one designed to match the new board dimensions. One other goody is that the new 52mm board has provisions for a back lighted pointer (full RGB, like everything else).
At some point in the near future I will also experiment with acetone smoothing on the cluster pieces - this should remove the 3d printed layer effect and also improve strength slightly (not that strength is an issue - even these draft pieces are solid enough that the only thing I am worried about is if somebody drops it on concrete while installing it, or decides to use it as a blunt weapon). Although the plan is still evolving, I am leaning towards offering a smoothed, lightly sanded and ready to paint (or not, up to the end user) set of cluster pieces in the kit option. I really don't want to do a ton of manual labour on the kits though, so this will depend on how easy the smoothing and a quick sanding process actually is.
Currently the cluster is setup to accept a cut piece of flat whatever for the window/transparent cover. I will likely end up using acrylic cut on a CNC, but may also end up going with a different material or simply leaving this as on option on the kits.
I plan to spend at least another two solid weeks hammering out firmware, and maybe a month total getting everything ready for a final set of prototypes that I can actually distribute to a few people for testing. After that there will likely be a bit of a lull, however, as I need to sort out some more work so I can fund a low volume production run of the project. This is actually a good thing though - I can work through the issues that will inevitably crop up in the test cases so that there are no major bugs in the final build. And even if there are, I do have a firmware update provision - so I can add functionality as needed and obviously fix issues that arise (this same interface is what will be used to customize the cluster, if so desired. Initially this will be a pretty full featured text interface (i.e. complicated), but I can likely throw together a python program with a UI that exposes most of the interesting bits to make customization more accessible).
Leave a comment:
-
Thank you, this is great info. I haven't worked much with plastics before getting this printer, and it did not occur to me that the wall thicknesses would be a factor in terms of flammability.
For overall mass of the enclosure I think we will be ending up in the ~400g range, depending on the window material, and yes, relatively speaking it is likely the least of your concerns if there is a fire, especially in a stock E30.
Maybe for now I will shelve the fire retardent plans. I do have a kg of some ABS rated for it here though, so maybe once the design is finalized I can do some crude experimention... definitely not just looking for an excuse to light shit on fire.Leave a comment:
-
Where any of the materials land in the UL94 flame retardation performance test is going to be hugely impacted by the flame retardants added to the resin. ABS is all over the place with the minimum wall thickness needed to meet V0 (0.8 - 3mm generally), but it is pretty common stuff and presumably some of the filament vendors will actually provide some sort of traceability for their material (although I am a bit skeptical of this unless it's coming straight from Sabic, LG, etc).
ASA looks like it generally has lower flame test performance, at least based on the few materials I pulled up on Matweb, but again I'd assume it depends on the flame retardants added. There were a couple on there rated V0 at 1.6mm walls.
Based on experience injection molding with high performance PC/ABS blends with high flame retardant content to meet V0 at thin wall sections on some consumer devices, it's sort of a mess. The retardants have been, in a number of cases, known to plate-out into the tooling and make a hell of a mess. The trouble that stuff would create in a 3D printer seems to be to be significant, so I'd wonder how common those materials would be in printing filaments. Then again, I have never looked into those filaments, so maybe the options are good.
In all cases, I am not sure how much any of it applies to 3D printed parts since they generally will have very thin & porous walls, at least on the top & bottom faces where you (assuming) have 3-5 0.2mm layers, plus all the infill and air pockets which all sort of seem like the perfect setup for fire lol. The wall porosity probably does a number on V0 wall thickness requirements just on its own. Besides that, how flame retardant is anything else in the E30? The good thing about the 3D printed cluster housing is that, while it is a fairly large volume, it is low mass, so there's not that much to burn.Leave a comment:
-
-
Are you aware of any flame retardant ASA's? Overall I would definitely prefer to use it, better UV stability as well as all the other bits you mentioned. However, none of the ASA filaments I can readily find seem to be flame retardant/self extinguishing, which is something I would really prefer to have (especially for track usage).Leave a comment:
-
I've recently switched to ASA, and far prefer it to the other materials I've printed in (PLA, PETG, ABS). It doesn't shrink like ABS, is easy to sand/finish, and is strong, real strong. I use it as the bobbins for custom guitar pickups, which wind up under a lot of tension after they're wound with thousands of winds of extremely thin copper wire. Stands up to heat great as well. I was printing on my MK3S's stock PEI steel plate, but recently switched to the Prusa powdercoated satin sheet.
Yeah, ABS has it's challenges for sure. I actually do have an enclosure, and the bed heater is as high as it will go (PEI on aluminium @~110-115C). I think my challenges are more so related to bed leveling - the large bed definitely isn't perfect, and what is less perfect is my application of the PEI sheet. I may have to drop on a piece of bs glass, but for now for prototyping it is not a big issue.
edit: On bubbles, I've had to deal with them a bit, pierce them with a needle and press the air out. The hole isn't large enough to cause any issues, but is large enough you can get the air bubble out.Last edited by Elysian; 03-22-2021, 12:49 PM.Leave a comment:
-
Yeah, a small oven would be ideal. I have used little IR ovens in the past and they work great. I need to make up my mind about whether or not it is worth it to get one - given my likely very low production volumes it may well be cheaper to farm out the build, but it would be really nice for prototyping or even just doing partial kits where the end user has to assemble all the not impossible to hand solder bits.
In regards to the bubbles... yeah they are a pain. But luckily the bed is so massive that even with some bad spots, I can just move a ~350x200mm part around a bunch and avoid them. The second draft of the front just finished, zero warping this time (finally think I have that all dialed in) and the two second draft enclosure pieces fit together great.
I definitely need to slightly widen the factory connector holes, a bit too snug of a fit for my liking right now. I am also working towards a modified rev of the backplane where one LED is in a better position, I was originally thinking of an angled light pipe but I don't think it is going to work nearly as well as just adjusting the LED position.
Overall I am extremely happy with the dimensional accuracy across the part. Everything I read about 3d printed ABS seemed to indicate shrinkage made for smaller parts - mine seems bang on to where I want it, certainly within the tolerances I used when modelling the draft (at least now that it isn't peeling off the bed). The only minor nitpick is hole sizes are definitely a little small, but this is to be expected, especially in the small flat pieces where the extruder hot end spend a lot of time remelting around the holes. And realistically I just need a guide hole of roughly the right size anyway for a screw to get some decent bite into the plastic. Since this is not something that will be coming apart often I don't think it is worthwhile to integrate nuts or anything like that. Just holding the assembled front and back piece with only half the screws installed, the unit feels very solid.
Here are a couple more snapshots, I still haven't removed the glue from the bed and I did zero finishing work, but there aren't any big bubble spots :). And yes... the ABS is switched to transparent partway through. Finally burned through my first kg of plastic!
It really is too bad I buggered up the wiring on the main connector, if it wasn't for that I could actually power the board as shown here. As it is, the dual 60 pin headers are backwards and when plugged in short power and ground (oops), so i can only power the board when the 104mm gauge is disconnected and tied in with jumper wires instead. There are a few other minor things that need to be fixed, like the LED position and some series resistors here and there, so a board re-spin is needed regardless. I also still have to check the factory connectors, which I will be getting to shortly.Leave a comment:
-
Damn, yeah those are some sizable bubbles! No possibility of peeling the film off and replacing it? I'm surprised they could not be popped, unless there is some crap trapped under them.
Good stuff with the electrical proto work. It's always satisfying to see stuff working for the first time! Also, QFNs in general are the devil for hand soldering, especially fine pitch ones, and doubly so for ones with big thermal pads! A future project is going to be to use a T-type thermocouple, a PID box and a toaster oven to build myself a reflow oven. A guy I work with uses an un-hacked toaster oven for reflow and it seems to work fairly well, so I bet one that can follow a closed-loop profile would be great.Leave a comment:
-
Definitely helped in a few spots, but I certainly wouldn't call it perfect. Some of them just do not want to deflate completely it seems.
You can see it really well in this - this is the second pass at doing the rear of the enclosure. Overall I am pretty happy with it but it is very obvious where the bubbles are.... Luckily I have plenty of print bed space so I can move this around a few inches and likely avoid them completely. The rest of the gunk on the back is just the impression left by the glue I experimented with when printing the lady some hooks, as well as some bits of the skirt I didn't remove. The big test with this piece was to see if I could avoid warping - and the answer is yes, the part did not peel from the bed at all. The final big item to address is that some of the screw mounts are simply getting overheated on the top surfaces due to their small size, so I may try a very, very low fan speed and see what kind of results it yields,
In other news, bringing up the boards is going fairly well. Only real challenge has been the stepper motor driver chip, which is precisely 4.2mm X 4.2mm with 24 'pins' that are actually just wrap around tabs. If you google "QFN24" you will get a picture, basically try to imagine soldering that when it is smaller than your pinky fingernail. And yes, there is a large pad on the bottom. 3 tries later I finally got it actually working, thankfully the one mounted on the main board will be machine built, as well as the very simple auxiliary boards, because these chips are not really hand solder-able unless you have a decent setup (hot air gun, paste, etc) and some experience. Currently my vision for the kits is having the complex control board machine assembled as well as the very simple aux. gauges, as those are cheap to do in large batches. So the kit itself would require some soldering for the large backplane board which is fairly easy (some 0603 resistors and a few ICs, but nowehere near as bad as the main control board).
Anyway, all that aside I went through 3 chips before I finally soldered one completely correctly. You can see it in action here:
Note that it is really noisy on the 'slow' movement; this is my lazy ass just hacking together some quick test code. It is actually much quieter during properly controlled movements, the movement at the start of the video where it sweeps all the way back to the home position is a better example. The 'buzz' sound is me firing up the big ol power supply.
This leaves a few other bits to get setup on the main board. Note that this does not mean 'fully programmed' - what I am doing at this stage is writing software that just performs some basic tests and lets me ensure the hardware is built correctly, so there is still some significant effort to go. Essentially the board bring up checklist looks like this at the moment:
Tested:
-Power supplies, including analog reference
-Processor
-UART/serial port for PC communication
-I2C multiplexer
-Digitally controlled LEDs
-On board LED chain selection logic (the pcb can drive LEDs on a separate attached gauge, there are some switches that let it essentially decide what LED string it wants to control)
-Stepper motor driver
-Voltage level shifting logic
Not tested:
-Motor selection multiplexer
-EEPROM
-SD card
-RTC and battery backup circuit (note, no actual battery is on the board - I mean the circuit that lets it run off the vehicle battery. This needs to be extremely low current consumption and verified to be so since you don't want it killing the battery. If you were running just the RTC off a fully charged battery it should be good for many, many moons)
-Analog input filtering
The SD card and EEPROM are extremely low risk, basically cookie cutter stuff I have worked with numerous times. The RTC is a chip I selected but have not used, and while not as bad as the motor driver was also a really fun one to hand solder, so that is next up on the hitlist. This leaves the following bits, which are actually on the backplane board and not the main control board:
-Various LED strings
-Main LED control multiplexers
-Digital character display
-Rotary switch/knob
-High voltage VR input for the diff speed sensor.
-Pulse input sensor for the tach. This is a little tricky because on some E30s it is a nice easy output from the ECU but on others it is much uglier (electrically speaking; physically it is as ugly as any other wiring mess) and comes from the ignition coil.
-All analog inputs, including filters and protection circuits
-All generic digital inputs and filters/protection circuits.
-Power down/power control circuitry.
-Careful verification that all factory signals are going where i expect. This is a pain in the ass and will be left to last.
Somewhat lower priority, but also fairly low risk since I have worked with them before (and therefore will be left until last):
-GPS. Chip shortage seems to be affecting this part too, there are some substitutes I could likely work in though.
-CAN transciever.
-USB adapter (FTDI)
-EEPROM
The small auxiliary gauges are actually just copies of the motor control setup on the main board, with a connector and an EEPROM identifier, so those should be very quick to validate design wise. I do need to fix some signals on the connector, which requires a board spin, but I can easily work around that with wires for my testing.
Anyway, things are moving along, bit by bit.Leave a comment:
Leave a comment: