totm june 2019 Custom hardware / simpit repository. For people who take KSP a little too far.

[richfiles]

Ah! I see. Maybe design in KiCAD, print out 1:1, and check, check, check those measurements in balsa. Once you feel confident about the prototype, have it made. Those FR4 protoboards would also work. Despite using FR4 protoboard, the first thing that comes to mind when I hear "protoboard" are the phenolic style "paper protoboards", which are quite thin and brittle, and don't even tend to sit entirely flat. Honestly, if you prototype anything on your higher quality FR4 boards, they'd probably just be usable as is then, though having a design that can be manufactured, and made available to download would be great for the community in general.

[catonthekbd]

If you were to get rid of those square internal corners, and made sure the cutouts still worked with 1mm radii, your pcb manufacturer would thank you

 

[richfiles]

Ah yes... @catonthekbd is absolutely right. It's easy to forget basics like that when it's been over a year since I last ordered a PCB... My first PCB. 
Definitely do a "DRC" (Design Rule Check) to make sure the design falls within the manufacturer's basic requirements. In this case, for cutout portions, consider what is easiest to machine. If they are milling out a cutout, consider the radius of their tool, and make sure there are no hard interior corners sharper than that radius (exterior corners are fine).

Another thing to consider, is if you have multiple holes, try to make as many of them the same size as possible. This will reduce the number of tool changes, making it simpler to manufacture, and potentially cheaper for you. Some manufacturers might charge a fee for exceeding a certain number of tool changes.

Another thing to consider, is that many board houses will send you three identical boards when you order their usual low cost prototyping option. If you have a design that requires multiple boards of similar dimension, consider seeing if you can create design overlap. Maybe a V-groove section that can be snapped off to provide a piece you need elsewhere, to allow the part to be used one way for one part, and a completely different way when used as another. I don't know if this will apply to your project, as I've not had a chance to look it over in detail myself, but it's something to consider. I'm going to be using this "feature" of many board houses in my controller for my navball. My navball project is using a real FDAI salvaged from an Israeli P-4 flight simulator. I need three identical circuits to control each of the three axes, but I need a single reference circuit and reference amplifier to provide a signal to all three circuits. By building the board to include the reference, a reference tap point, and a single axis control circuit, and a reference power driver, I can order the ONE board, but get three identical copies of it. I'll populate all three axis drivers, but on one board, I'll populate the reference generator and the reference power driver. Boards two and three will just have those parts left out, but will be connected to the reference signal from board one. Tricks like that can drastically reduce the board cost by just taking advantage of the way manufacturers supply these boards, usually in triplicate. Another way to benefit from triplicate boards, is durability. You can stack two or even three thinner boards to make a stronger assembly, if it's needed.

[Francois]

On 5/4/2019 at 12:59 PM, catonthekbd said:

If you were to get rid of those square internal corners, and made sure the cutouts still worked with 1mm radii, your pcb manufacturer would thank you 

Ha thank you for this advice, I didn't realize but it makes sense. For the same reason, I'm happy with cutting in balsa wood with a hobby knife, but when I have to do a shape in plywood with a dremel and a drill bit, it's another story...

On 5/5/2019 at 12:03 AM, richfiles said:

Another thing to consider, is if you have multiple holes, try to make as many of them the same size as possible. This will reduce the number of tool changes, making it simpler to manufacture, and potentially cheaper for you. Some manufacturers might charge a fee for exceeding a certain number of tool changes.

Good to know as well, so far, my CAD design can be improved on this side.

As I'm making progress, I'm understanding better what doesn't work between the theory of a CAD program and the reality of my fat fingers and poor tooling. I think I'll keep prototyping and maybe replace parts as I go. I could see a working prototype with wood and protoboard evolving to manufactured PCB to replace the main board + 3d printed parts like my new bearing mount for easily reproducibility by the community.

Here's the progress I made yesterday:

It spins, sort of... 

(Also let me know if I should give fewer updates or start my own thread)

 

[klond]

 Wasn't sure where to post, thought here was most appropriate.

 

Pretty amateur compared to what's going on in here, but it's my first project, it's nothing special, but it's complete!

 

Amazon 10-key pad for about $10

Terrible hand-cut avery labels run through a laser printer

A 4-pack of those clamps was a couple dollars.

Drilled some pilot holes for the clamp.  Collected a few matching from work, threaded for plastic.

Autohotkey to overwrite numpad buttons.  Took some reading and experimenting to figure out their script programming.

Electrical tape over the annoying blue LED.  Wonder if I can utilize it somehow.

 

[richfiles]

@klond That is by far, simply THE MOST Kerbal thing I have seen all year! Bravo! 

For the Blue LED, you could replace it with something a bit more Kerbal, like a green LED. You might need to add a small resistor in series with it, since a blue LED almost certainly will have a higher forward voltage than most green LEDs. As a result, blue LEDs will often have a smaller resistor than what you'd use for a green LED. Adding a resistor in series will increase the total resistance, further limiting the current through the LED. The larger the resistor, the dimmer the LED. A subtle green glow might be more preferred than a blazing blue annoyance.

Edited May 13, 2019 by richfiles

[Francois]

A bit of progress tonight: both motors are screwed on the plate and it fits in the 100mm ball.

Next, I will make the mount to hold the axle properly, something like this:

[James Kerman]

Congratulations on being awarded Thread of the Month @Mulbin.

A special note about this particular thread: All the contributors here are deserving of this award, not only because of your amazing hardware creations but also because of your helpful and polite demeanor. Keep up the good work!

[c4ooo]

Does anyone know a good method of painting wooden boards (MDF) so that they look nice? Years ago I tried matte black spray paint, but the finish was rough and the paint rubbed off onto skin. 

[qzgy]

On 7/4/2019 at 6:49 PM, c4ooo said:

Does anyone know a good method of painting wooden boards (MDF) so that they look nice? Years ago I tried matte black spray paint, but the finish was rough and the paint rubbed off onto skin. 

I wonder if just normal wood paint/spray paint plus some form of clear coat would work.

Maybe a primer coat might also help but I'm not sure.

[hugopeeters]

MDF sucks up paint like crazy. You need to use primer. I used a high gloss spray paint which worked fairly well.

[Tabb]

Hi all,

I'm currently working on my own KSP controller based on the one from hugopeeters.

Here is the drawing I plan to use to lasercut the box. The texts will be cut as well so that led in the box will backlight everything.

Do you have any comment before I order it ?

[Francois]

Quite a while since my last post. I've been down a great rabbit hole  I was wondering how to achieve precision with my terrible tools, my Dremel and my Swiss army knife. I don't really want to get into 3D printing even though I'm thinking more and more about it, but after discovering that guy https://www.youtube.com/channel/UCe5ZPXFmhf00k4sB8cGMVOA I realized that a CNC would help for the navball parts and for panels later. So I bought a cheap Chinese CNC machine (3018 Pro on Amazon) and went deeper into the rabbit hole to learning about it and upgrading it with an emergency button, some limit switches and a z-probe.

Weeks later and a bit wiser, after a lot of design and redesign, I made progress on the outer case of the navball to the point that it's taking shape. Here's the model:

The process and the current state:

I'm taking a week of staycation next week so I should make a lot more progress hopefully.

Oh, and I visited the Air and Space museum in DC and Kennedy Space Center in Florida (and saw STP-2 launch and land \o/)

[tsaven]

10 minutes ago, Francois said:

I was wondering how to achieve precision with my terrible tools, [...] So I bought a cheap Chinese CNC machine

I feel like this comment summarizes this whole thread pretty well.  We want a controller for our lego spaceships game for no functional reason, and then end up spending hundreds (thousands?) of dollars on tools and parts for these kind of screwball passion projects. It's really quite Kerbal, when you think about it.

Your plan looks seriously badass man, I can't wait to see a functioning mockup.

Also, a minor life-pro-tip:  I've found that clamping a sacrificial sheet of MDF or plywood to the bed and then using 2" wide double-sided tape to stick your work piece to it will be plenty strong enough to hold the work piece in place while cutting, while allowing you to cut all the way down through the work piece (and into the MDF) to separate it without damaging your bed.

[tsaven]

My project isn't nearly as pretty or as finished as many of the other setups here, but I've got the basics working so far off of my Arduino and SerialIO.

I've also started to bolt together a temporary stand to hold most of my equipment while I continue working on the code.  The various timing functions (Such as displaying time to AP/PE/Target) have been giving me a world of headaches as I try and keep the formatting correct, so when I feel like taking a break from lines of C++ I get back out the drill and sword of Exact-Zero and start wiring stuff together.

[Francois]

On 9/9/2019 at 1:21 PM, tsaven said:

I feel like this comment summarizes this whole thread pretty well.  We want a controller for our lego spaceships game for no functional reason, and then end up spending hundreds (thousands?) of dollars on tools and parts for these kind of screwball passion projects. It's really quite Kerbal, when you think about it.

Your plan looks seriously badass man, I can't wait to see a functioning mockup.

Also, a minor life-pro-tip:  I've found that clamping a sacrificial sheet of MDF or plywood to the bed and then using 2" wide double-sided tape to stick your work piece to it will be plenty strong enough to hold the work piece in place while cutting, while allowing you to cut all the way down through the work piece (and into the MDF) to separate it without damaging your bed.

I think I've started making real progress when I gave up on making it happen cheaply. But I agree, my Kerbal thinking for this: if it doesn't burst into flames, that's a good first step :p

I already put a thing piece of balsa under the piece, but I should switch to tape indeed. I've run a few times the machine into a clamp. Thanks for the advice.

Good luck on your project! looks like a good start! What's the final plan?

Edited September 13, 2019 by Francois
typo

[tsaven]

45 minutes ago, Francois said:

I think I've started making real progress when I gave up on making it happen cheaply. But I agree, my Kerbal thinking for this: if it doesn't burst into flames, that's a good first step :p

I already put a thing piece of balsa under the piece, but I should switch to tape indeed. I've run a few times the machine into a clamp. Thanks for the advice.

Good luck on your project! looks like a good start! What's the final plan?

For cutting soft things like wood and even aluminum, as long as you're cutting into a sheet with a lot of surface area I've found that firmly pressing it onto a continuous bed of strong double-sided is plenty strong to keep the work in place.  And it keeps your clamps out of the way so you have full access to the piece, plus the aformentioned sacrificial nature of it.

I'm still working on my plan.  I think for style I'm going for a Space Shuttle/Apollo look as much as reasonably possible, but I'm having a hell of a time finding the exact type of buttons that I want.  But so far, this is a kind of rough plan.  The cutout at the bottom is for my keyboard.

[tsaven]

Quick update for those who are interested.

Thanks to staying up until 2:30am on a work night, I've made a lot of progress on my test controllor. 

 

Only about half of the buttons work so far; The SAS mode select/LEDs are still running on the breadboard, as are the Altitude & velocity selection buttons, and the throttle isn't hooked up yet.  But the 7-segment display readouts are working great, as is staging and RCS/SAS toggles!  The left displays are Apoapsis/Periapsis, the right two are Altitude and Velocity (switchable via buttons on the breadboard between Surface/Velocity/Target and AGL/ASL)

Bit of a video of some basic functions in action!

 

Edited September 16, 2019 by tsaven

[Francois]

Yesterday progress: added that spoilboard to my tiny CNC, definitely makes life easier, thanks @tsaven. I just have to be careful where I put the tape because the machine chokes bad when it hits the double-sided tape glue.

And so far, it fits surprisingly well together. The slip ring is perfectly aligned with bearing and the motor

More updates to come, I'm off the whole week

[tsaven]

54 minutes ago, Francois said:

Yesterday progress: added that spoilboard to my tiny CNC, definitely makes life easier, thanks @tsaven. I just have to be careful where I put the tape because the machine chokes bad when it hits the double-sided tape glue.

 

And so far, it fits surprisingly well together. The slip ring is perfectly aligned with bearing and the motor

 

More updates to come, I'm off the whole week

Looking forward to seeing progress!  I hadn't considered the problems that the double-sided tape glue might present for smaller systems, the machines I've got access to are a little beefier.

Any reason you're going with plywood as opposed to a plastic of some kind?

[Francois]

6 minutes ago, tsaven said:

the machines I've got access to are a little beefier

haha, that doesn't fit in my apartment

7 minutes ago, tsaven said:

Any reason you're going with plywood

you mean for the mounts I'm building or the spoilboard? For the mounts, it's not even plywood, it's a pine board. The spoilboard is a 3mm plywood sheet. I guess I'm staying away from plastic to limit plastic waste and dust (I mill in my bedroom :-/). I'll use acrylic for the front panel though.

[tsaven]

A while back someone posted a link to some 3D files for some toggle switch caps that look like they'd do a decent job of emulating the switches on the Space Shuttle.  Just for fun I had like 50 of them printed, I got them in today:

 

Overall they're a bit smaller than I expected.  I think if they were upsized by 30% it would feel a bit better and will surely fit on the toggle switches easier.  As it stands I think I'll have to drill them out and trim the toggly part of the SPST switches from Sparkfun to get these to fit.

Not really sure what I'm going to do with 50 of them, but they were like $1/ea in this quantity so whatever.

[KerballingSmasher]

Those are... kinda small @tsaven. Glad you can use them though, not wasting $50!

[c4ooo]

Do you guys think covering the wood panels with aluminum tape would make for a good aesthetic?

[mountaineerman]

I think the edge between the rows of tape might not look super great. My plan is to use some kind of dark-grey paint over aluminum to give a cockpit-type feel.