New project - MSFS C++ USB interface

[richard]

i find super glue won't last on "common garden variety" pla. there are quite a few pla variants around now, silky smooth , pla+ , pla super strong
that i have not tried gluing up. glue may not be the only issue, one of the silky variants i tried wont hold its shape over time in normal summer time temps/humidity either. its a bit of a mine field, abs is more durable/predictable but the fumes give me trouble.
had similar issues with resin printer, the nice fume free water wash up resins fall to pieces if exposed to sunlight for a couple of weeks
they come off the printer looking lovely, it was marketed as dental resin like .... not!
the first abs print i did to repair my weather station wind vane bracket has been out in the sun and rain for 10 or 11 years now and is still fully intact.
can't remember gluing abs but i think that was not easy either to get a good result , got some memories of acetone being involved

[Demon]

ABS is definitely superior. Check the plumbing section at the hardware store; look for the ABS glue close to the black pooptoobs (not to be mistaken with the PVC glue - those white tubes).

I'm using ordinary amazon PLA, and I was able to separate by hand the 2 pieces just now. Took an effort, but it did separate.

You do bring up an important detail about PLA; exposure to UV destroys it quickly. I'm using LED lights in the basement, but I want my original prototypes to live long without being OCD about keeping it in light-proof wrapping (I don't want to reprint later, then refinish the surface again - I know there's a lot of detailed work involved - I've done something similar before).

I guess that means development and testing with a glued-together PLA prototype (I'm gonna try Gorilla super glue - they tend to make quality products), then printing an ABS unit when I'm ready to make the mold.

[Demon]

Speaking of longevity, I made this silicone mold way back in mid-2000.




Last time I checked it maybe 5 years ago, it looked "not bad", now it's got a slimy feel, and the main bulk has shrunk relative to the 3 small partitions. I'd say it was good quality for at least 10 years.

I don't know if Smooth-On has "watered down" their products by now, but I can guarantee they were awesome when I was playing with this stuff.

I also tested one of the liquid plastics that you can make threads. All I remember is that it was graded industrial strength, and was mixed by volume, 50-50 (I didn't have a digital scale back then and didn't want to be bothered with them).

[Demon]

So I did a full-frame test to see how cooling, warping and shrinkage would affect the end results. The frame shrank nearly 1mm at the widest part, and there's practically no warping (I left it cool on the 3D printer).




I didn't do any detail work, like front bezels and covering that unused section on the top (it's actually upside-down, that's really the bottom).

Overall, I'm happy with the results. The screen fits just right.




Now I have to see how I'm going to make the honeycomb innards, connector openings along the bottom edge, buttons on the lower front as well as supports for the main controller and button board.

[Demon]

I love FreeCad. I don't have to print the whole frame to do a test; I can just crop out the rest until I get the button section right.




I still have to make a clear LED chimney in that 3rd hole from the left.

Buttons and cable fits snuggly in the frame.




I don't think Im even going to bother making caps for those pushbuttons. They are easy to operate.

[Ioannis]

It is really amazing what you can do with a relatively cheap machine these days! Endless possibilities.

I suppose the LED is between 4th and 5th button, right? Can you 3D print transparent plastic guide?

Ioannis

[richard]

clear UV set resin is great for making small transparent windows in housings, laser cut acrylic sheet can make light pipes [best used edge on], you lose a bit of brightness though

[Ioannis]

With resin you need a mold I guess.

But with laser cut acrylic sheet, the pipe may not be cylindrical, or is it?

Ioannis

[richard]

With resin you need a mold I guess.

I just tape the back side and fill the hole [if you use the thick stuff you can get a nice dome if your quick], shine led for a few seconds and ..
volar instant clear led bezel.

easiest if light pipe hole is rectangular and edge on for best effect. i have used those miniature diamond hole saws to get circle shape on the edge of acrylic sheet but its a bit fiddley. its handy out when the leds don't line up with the holes in the case, like when you 3d printed a mirrored image of your stl for 7 hours and don't have the heart to do it again

[Ioannis]

Thanks for the info.

Ioannis

[Demon]

...I suppose the LED is between 4th and 5th button, right? Can you 3D print transparent plastic guide?

Ioannis

Yup, and yup.

It's not even worth making molds to pour plastic light pipes. Since this is so small, I can 3D print a whole batch in no time.

Clear ABS:
https://www.amazon.ca/Gizmo-Dorks-Fi...dp/B00GU2FZQQ/

Clear polycarbonate:
https://www.amazon.ca/Polymaker-Fila...9KKMLKRF/?th=1

Clear PETG:
https://www.amazon.ca/OVERTURE-Filam...dp/B07SB761QR/


It's not worth using PLA, it won't survive UV light. At the moment I'm leaning towards ABS because of cost, but polycarbonate is the clear winner according to google.

[Demon]

I've been playing with FreeCad; testing different ways to align the holes for the connections to the controller board. I finally got the splice tool working properly (I didn't want to watch videos ).




I reproduced the controller, with clearances under the board for protruding pins, areas for fasteners and added flares to the front of the 4 connectors. This is a lot easier, I can relocate the controller as I want, and generate new holes on the fly.

[Demon]

Turn out I can print the main parts in one piece. I'm pushing the limits of my Ender 3 S1 Plus, but in theory "it should work". The parts should have enough rigidity to withstand the stresses of being printed at an angle. I just need to add supports.

Case front is pretty much finished. I just have to add rounded edges and a bezel, but I'm only adding them last. FreeCad is a pain with those when you make slight alterations to the shape (probably cause I don't know all the minute intricacies of updating in FreeCad).




Nearly finished the interior brace. I just have to add attachment points for a VESA-spec mount in the center.




Just got the back panel to design, that's probably the simplest of all 3.

[Ioannis]

Never thought of that, printing to an angle so it can fit in the area!

I like this thread more and more!

Ioannis

[Demon]

You just like me for my pictures...

[Demon]

45deg was too shallow, 50deg was too tall, 47.5deg was just right under the upper limit, and fit within caution area on bed.




Print time is not realistic. This was done with a 10mm thick triangle, just to test for best print angle.

[Demon]

Sometimes when you make support panels, they might not have a lot of detail and it's not always apparent which way is which.

So I add little alignment markers to help me when I assemble it for a print (like on these supports).

[Demon]

The more I think about it, the more I'm leaning towards 3D printing to make the enclosures. There's a LOT of details in using liquid plastic, the one that scares the crap out of me is the vacuum chamber to degaz the liquid. I have the chamber and pumps, but I'm scared silly to have that thing fail. Liquid plastic makes a superior enclosure, but it's also labour intensive. With 3D printing, you start the job and go do something else; it literally doesn't take any of your time

The 300mm x 300mm x 300mm limitation of my Ender 3 S1 Plus was costing me a lot of time having to print at an angle. And then I fall on this puppy.

https://www.amazon.ca/gp/product/B0CD7S18PB/?th=1

The Anycubic Kobra 2 Max has a ridonculous 420mm x 420mm x 500mm print size that runs at 500mm/s (reviews say 300mm/s is better). Considering my printer runs at 50mm/s, that's a 6-fold increase in speed.

I'm definitely setting up an exhaust port in the basement. I could print my enclosure flat on its back with ABS, no supports, no wasted filament, and at an insane speed. I can still keep my Ender for supports jobs; like to print the smaller enclosures, brackets, light pipes, etc.

[Ioannis]

Well, OK with the self 3D printing for one or two pcs. But is it cost effective for small production?

What about using JLCPCB 3d capabilities? https://jlc3dp.com/?source=JLCPCB-top-productbar Will this be better choice regarding cost/time/quality?

Ioannis

[Demon]

I didn't even know JLCPCB offered 3D services. I have only one fear when dealing with China; it's having my idea stolen. I grew up in a time when China regularly cloned western products. I'm sure they still do today. Back then, they had to reverse engineer devices; now we're just sending them the actual prints willingly (whether it's PCBs, 3D files, etc).

I know that anyone could fabricate what I'm building; there's no rocket science involved. But still, I'm apprehensive about helping them clone my idea. I'd use multiple names for various components if I could, but that could create havoc with things like shipping, customs and such.

Out of curiosity, I got an instant quote for just the front panel for the display, $87.45.

I'm clearing more workspace for a Kobra 2 Max coming in tomorrow. I pay the printer in 6 prints, and I'm not even including shipping and brokerage fees on that, so it's probably more like 4 prints.


EDIT: One thing I'm thinking of doing, is having only the switches, encoder and pot circuits in china, and have the main USB circuit done with digikey.

[Ioannis]

Well, you do not have to tell them what you are doing. And after all, having a plastic thing printed there, what really means to them? They cannot figure what you are doing, right?

The $87.45 is for one piece?

Ioannis

[Demon]

...The $87.45 is for one piece?

Ioannis

Yup. I didn't even bother checking for quantities.

My enclosures consist of 3 parts; it's not like they will assemble it, the wrap and ship (they probably offer the service, but I'm not interested in taking them apart again once they get here.

I remember how JLCPCB likes taking humongous boxes, and that's what I expect from the 3D print service; each part wrapped separately. It's not like you can pile them up one inside the other either, I don't have a slope on the case. Shipping is gonna cost a fortune cause the final box is going to be huge.

[Demon]

Now this is a lot more reasonable than the Ender.




5 hours 35 minutes for what was gonna take nearly 2 days. This is the normal profile in Cura (from Anycubic web site), with:

• slow first layers
  0.15mm layer height
  0.4mm walls
  10% infill
  200mm/s print speed
  lightning infill pattern (90 minutes faster than default grid)

Too bad it doesn't have 2 extruders. I could print my white text directly into the black casing.

[Ioannis]

10% infill means that the compact pieces are not really compact inside?

Sorry, but I am at zero, regarding 3D printers.

Ioannis

[Demon]

Yup. You can chose from many infill patterns at varying degrees of fill.

I start at the lightest just to test fitment, then move up until it's as strong as I want.

[Ioannis]

And I guess the higher infill the slowest process. Added to the more plastic usage, right?

Ioannis

[Demon]

Exactly. It's really fast to print an empty shell, but don't sneeze on it.

The hard part is finding that right balance. I don't have stress testing equipment, so I have to use my better judgement.

I can't afford to destroy a LCD and controller card, but I can throw an empty enclosure onto the street and inspecting it for damages; looking for weak spots.

I'm the last house on a dead-end street, so I can easily stand out there and play soccer with my case.

[Demon]

Well, you do not have to tell them what you are doing. ...

Yup. I googled "jlcpcb steals design" to see if there's a history of incidents, and fell on this thread right away:

https://www.eevblog.com/forum/manufa...-risk-of-copy/


Read « Reply #19 on: July 22, 2023, 07:30:50 pm » specifically. The best thing I've learned was "don't put your company or product name on the PCB".

And I JUST finished generating all my PCBs.

I'll go through them tomorrow and take out my company name.

[Ioannis]

What I read from that thread is that you do not have to worry about stealing your idea from pcb perspective.

And I never cared about that. Maybe because I use most of the cases, MCU on board and no one knows which one or what firmware does it have.

As noted others on the thread you can just print the names and NOT the values of components on the board.

In any case who does care for my 10, 50 or even 100 pcs? And if someone manages to do what I do cheaper, then I will resale theirs and have more profit!

Ioannis

[Demon]

Just got my first order of 14 circuits, these look good:




Until you see the other end:




Oh well, live and learn. I had checked too many times, and was only seeing what I wanted to see.

[Ioannis]

Fortunately PCBs have two sides! Use the other now!

Ioannis

[Demon]

At first I went "Aaaaah crap".

Then I went "YAY! I can use the other side for my tests".

And then I found out the support pins are staggered and not reversible, "Aaaaah crap".

[Ioannis]

Well, you are not alone! Happened to all!

Ioannis

[Demon]

So I have a need to override a pushbutton (from ON to OFF) and came up with this concept; a solenoid piggyback on a pushbutton switch with a 3D printed cap held by a strand of wire.




The system works, except the switch needs about 330 grams of force (according to specs, and my digital scale). The 5V solenoids are nowhere near strong enough, and this 12V variety comes up short. It struggles, but doesn't get the job done.

I sure could use ideas. It needs about 5mm travel to complete the stroke, and must not interfere with normal switch operations.

[amgen]

Hall effect switch with small magnet or if there is magnetic field around solenoid coil that the hall effect device can sense