Custom Digital Dashboard Project

[brentp]

Shared here https://www.facebook.com/permalink.php? ... ream_ref=1 as well as on twitter.

Also, polycase.com may be a good source of an enclosure if you want to CNC route the LEDs and LCD window

[emdash]

Inexpensive hobbyist 3D printers may eventually lead to a revolution in manufacturing one day, but at this monent, they're just like any other manufacturing method: they have advantages and disadvantages. Good results require skill and experience on the part of the operator.

The obvious advantage to 3D printing is that the process, once initialized, is entirely automatic. The other advantage is that there are fewer constraints on the types of shapes that can be made, and those constraints are easy to understand and work around once you understand how the machines work. Once you have your part, you can instantly tell if you're on the right track before you invest in more costly manufacturing.

The obvious disadvantages are speed and quality. Complex or fine detailed prints can take many hours. These machines work by extruding thin streams of plastic along 2D paths, building up the object one slice at a time. And it shows in the results. "Wow, a 3D printer made that?" is not something I expect to hear any time soon. That's when things go well. There are plenty of gotchas and pitfalls that can cause prints to fail: warping, stringing, blobs forming, parts falling over. Combine that with slow print speeds and limited available print time, and there's a huge potential for frustration.

Having said all of that, my first print turned out pretty decent:

[emdash]

Printing both top and bottom pieces took about 4 hours. About an hour each, plus some down time fiddling with settings. I designed the rear cover while the main body was printing. Once both parts were done, I did a trial fit and was shocked to discover that the parts actually mated together correctly, and the screw holes were properly aligned (I can hear you snickering at the last sentence. Grow up, internet!).

I couldn't wait to get come home and try installing the parts. But I was immediately disappointed. Neither the LCD nor LED fit correctly.

[emdash]

I went back to techshop to see where I went wrong. I decide to print some test objects, to try and determine how accurate the printer is. What I discovered is that holes below about 7mm print slightly under-sized with the settings I used. 10mm and larger holes print pretty true to their nominal dimensions.

The LED dimensions in the CAD file were correct. 5mm LEDs are actually about 4.95mm in diameter on the body, and 5.5mm in diameter at the flange. So a hole diameter of 5mm should give you a nice, snug fit.

For the LCD, on the other hand, it seems I just plain got the dimensions wrong. That's what I get for hand-measuring the part instead of looking up the datasheet and working from the engineering specifications.

I've updated the CAD drawings based on this experience. I've moved the LED holes further apart, because it seems their proximity caused distortions in the print. So they're now spaced .400" instead of .300". I've added an "LED Clearance" parameter to the model, so I can fudge the LED diameter and print a slightly larger hole. I will make another test object so I can see how print settings affect tolerances. Or, I could just plan on chasing all the holes with a tap drill. Finally, I've added features to the model to secure the LCD inside the housing, and secure the other electronics to the rear cover, and a hole for the cable to pass through (going with a pig-tail for now, since I have no idea what the circuit layout will look light).

One final thing I was relieved to discover is that I was able to secure the rear cover with plastic screws, and the case didn't crack. What this means is that if I can get a good print, I can actually use it for testing. It won't be pretty, but it will allow me to refine my design before sending it out to be made by someone who knows what the hell they are doing.

[emdash]

I ran another print tonight. Looks like the changes to the CAD file were spot on, and the components all fit correctly now. I'm not entirely happy with the results. I mean, let's face it: it's ugly. I've been having the trouble with lots of "string" forming, and nobody quite seems to know why. I've also been having trouble with part warpage.

For now, I'm satisfied that the interior of the enclosure is reasonably close to correct. And what I've got is good enough that I can continue hacking on the electronics, the goal of which is to miniaturize my prototype to the point that it fits into the enclosure. Once that's accomplished, I'll revisit the enclosure.

[emdash]

Took a few days off from the project. We had some good weather over the weekend, so I had to go riding.

Went on a bit of an on-line shopping spree today. Ordered some parts for "Rev2".

* This Display: http://www.buydisplay.com/default/240x6 ... odule.html
* Teensy 3.1: http://www.adafruit.com/products/1625
* Waterproof 4-wire cable set http://www.adafruit.com/products/744

These are all speculative purchases. I'm going to keep hacking on my Rev 1 design because I need something working when the season starts in the middle of next month.

If I can get all this working, rev 2 will be a huge improvement over rev1. The 240x64 pixel display is plenty big enough to be seen and has more pixels than necessary. I did a test visualizatoin with GIMP and a laser printer). I opted to get the thinner FFC variant. I've never worked with FFC style connections. It will require fabricating a PCB, and I've never done that either. But I know people who have and can help with that.

Teensy 3.1 is an arduino pro mini on steroids. It seems like the way to go. With 34 IO pins, I can feel free to do whatever I want. I don't have to fret about using 7 whole IO pins for the display, or worry that a JSON parser won't fit in 2k of ram. Plus there's 3 on-board UARTS, so that leaves open some interesting possibilities. For example, the display could have serial pass-through. Meaning you could daisy chain serial peripherals, and have your display cake and with some telemetry icing. And I still have another UART left over for debugging. Plus it's got USB on-board, so I can be indecisive and reprogram it easily. At $20, it's really a bargain. I'm not sure why you'd bother with anything else.

Inexpensive printers are finicky. I think the filament I bought has slightly different properties than the other filament techshop sells. It melts at a lower temp, and it's less viscous when melted, so it oozes out of the extruder when it's not supposed to. I have some ideas to get better prints, but it'll be a while before I can try them. Some of them involve changes to the printer settings, and others, sadly, involve changes to the design to make it more printer-friendly.

Actually in a lot of ways it would be easier if I could just use a mill or one of the shop-bot CNC routers. Techshop's class calendar hasn't been very compatible with my calendar, so I still don't have access to those.

Anyways, enough complaining. I never thought this was going to be easy. Tonight my plan is to cut down some of the perf-board I have to a size that will fit in the enclosure and try to assemble the majority of the components. It would be encouraging to see everything working together.

[emdash]

I "finished" the enclosure, but I'm really not happy with the results. Don't feel good about putting it on my car -- it's too funky and too big.

I've realized the importance of doing mock-ups. I would have saved myself some trouble if I'd focused on the exterior design first. I've also realized the importance of making PCBs. Proto board is hard to work with. It's error-prone. It complicates exterior design, because the component placement is constrained to a .1" grid. It limits the choices of connectors (many common connectors are .05" pitch, for exaple). It also wastes a lot of space. A better design sequence seems to be:

Choose components (screen, connectors, hardware) -> Exterior design incorporating components -> evaluate -> if happy, design PCB else goto 10;

Of course, in reality there's a push-pull tension between exterior design and hardware considerations on the interior. All I'm trying to point out is that it's a waste of time to focus on interior design only to discover that you can't stand the resulting exterior.

My order from Adafruit arrived. Still waiting for the LCD. I have enough information and whatnot to make a detailed mock-up of rev 2. So I'll spend some time on that while I wait for the LCD to arrive. The Type A machine printers are fine for that sort of thing. I'm just not sure I feel comfortable using them for the final product. The prints tend to be porous, and there's no easy way to fix that.

In the mean time -- to give myself something I can finish in time for the next regional (too late for the upcoming one) -- I've decided to scale back my ambitions a bit -- 6 shift lights and a master alarm. Since I won't be using 7 whole pins for the LCD, I don't have to pair the LEDs. The whole package should be pretty small. I think I can handle that, even on perfboard.

[emdash]

More pics.

[gizmodo]

Instead of printing the housing why not get some 3/32 or 1/8 inch ABS sheet to work with? It is pretty cheap, you can heat it and bend it and attaching can be done with ABS glue or a plastic welder.

[emdash]

Instead of printing the housing why not get some 3/32 or 1/8 inch ABS sheet to work with? It is pretty cheap, you can heat it and bend it and attaching can be done with ABS glue or a plastic welder.

In a word: tech shop.

I haven't taken the safety courses (SBUs) I would need to work with ABS (shopbot CNC router, upright mill, etc). You can't put ABS in the laser cutter because apparently it releases hydrogen cyanide or something equally nasty. It was easier to take the 3D printing SBU due to silly scheduling constraints. Also $$$.

So right now, what I need to do is come up with a design that I'm happy with. Then find someone who can help me make it using whatever method.

[gizmodo]

I was thinking more along the lines of cutting it with a razor or hot knife and heating it with a propane torch or heat gun. Or if you want to get a bit more involved you could build a simple vacuum table and do it that way. I made these two parts with a home made vacuum table and some .090 ABS. I just used MDF and Bondo for the molds.

[emdash]

I was thinking more along the lines of cutting it with a razor or hot knife and heating it with a propane torch or heat gun. Or if you want to get a bit more involved you could build a simple vacuum table and do it that way. I made these two parts with a home made vacuum table and some .090 ABS. I just used MDF and Bondo for the molds.

Interesting. Is that texture deliberate, or a result of the molding process? I'd be curious to learn more about how you did this. Also, tech-shop does have a vacuum table, though I don't know if we're allowed to use it with ABS.

BTW, it turns out that a friend of mine has access to a much better 3D printer. Much more accurate, and can print with a combination of ABS and water-soluble support material. This means it can handle any shape I care to come up with, and with much less guess work. I'll probably go that route for the next iteration.

It will be some time before I get that far, though, because the racing season starts this weekend!!!. At some point you just have to stop hacking and drive.[/i]

[gizmodo]

The ABS that I buy is smooth on one side and textured on the other. The rest of the dash in my car has a slight texture to it so I decided to go with the textured side out.

I looked on craigslist for a free oven that someone was getting rid of. I used the element from that and build a tin-lined wood box that I use to heat the material. I built the vacuum table out of 2x4s that I found in the dumpster of one of the houses they were building by my house. For the top of the vacuum table I just used 1/8 inch MDF and drilled holes in a 1x1 inch pattern. I used some of the left over 1/8 MDF to space the top sheet away from the bottom sheet/ In the bottom sheet I drilled a hole and glued a piece of PCV pipe that I hook up to a shop vac. I have been pretty happy with it.

I have had luck forming ABS, HPDE, and Lexan with a draw depth of 5-6 inches. Any more than that and I will need a bugger vacuum.

Here is what I used to base my design off of: http://www.tk560.com/vactable4.html

[emdash]

I had a fantastic weekend (apart from accidentally flooding the engine of my street car, and having to spend Monday in Willows waiting for help. Not winning anything yet, but getting faster and more confident little-by-little. Saturday's race went well. No mechanical problems. lap times consistently in the 2:10s (not competitive, but not last place either). Car's brakes need work.

As for the project, It turns out that the dimensions of the "Rev 1" enclosure are actually just about right. It. And my guesses about the mounting bracket size and location turned out to be spot on. To top it all off, some of the other racers actually seemed impressed by it (or something similar) will go on my car after all. One of them offered to run some prints on his company's machine (a commerical SLA unit). He says it would be good for him to learn how to use it.

Design wise, that was made abundantly clear during Sunday's rain race, is that everything needs to be water proof -- including the RCP. Th existing enclosure is really inadequate in this respect, so I'll have to come up with a solution pretty quick -- even if it's just to put the entire unit a Tupperware container. Meanwhile, looking for water-proof connectors. Adafruit has waterproof 4-wire and RJ45 pigtiails. Summit racing has some stuff as well. Connectors add up.

I took some pictures, but forgot to grab my charger cable, so I guess I'll have to remember to update them when I get home.

No word on my rev-2 display. I see the charge on my bank statement, but haven't received any communication from buydisplay.com.

[brentp]

Thanks for the update! Yes, for water proof it would be good to get it inside of a compact sealed enclosures with grommets.

We're planning a fully waterproof version of RCP as a premium offering.

-Brent