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

[Francois]

End of vacation for me and not as much progress as I hoped (but a short trip into the mountains, so there is that). It's been more setback after setback last week.

I was hoping to be done with the redesign of the inner plate to mill it on a PCB protoboard and layout the components and have everything turning to then work on a rewrite of the software.

I redesigned that plate in Freecad in a better way, improved the locking mechanisms now that I have precise-ish milling capabilities and I wanted to get a good motor fitting test before milling a PCB. So I mounted the motors on a balsa prototype, wired them with my circuit that's on the breadboard and... smoke... I don't know what I've done, but I'm up for some troubleshooting and component replacement now.

I ended up powering up the two internal motors with the small drivers that were provided with them and I realize that I had some clear wobbling on the axis that will hold the plastic hemispheres. I need to test more, but I'm pretty sure that the wobble doesn't come from my design but from the bearings. This will be harder to resolve than if it was my design obviously.

One last thing, the slip ring that's in the ball is rated for 1.5A per channel. After having to get a new wire stripper for such tiny wires, I'm pretty sure it's 30 AWG wires, which means there's no way it will handle 1.5A. I'm thinking about different things here:
- I may generate a 1.5A current and pass it through a wire of the slip ring (or a piece I've cut actually) and see if it melts (my noob level in electronics keeps my hope mid-level :p)
- I may develop a more complex internal circuit to pass data over 2 wires and use 4 wires for power but, again, I'm still new to electronic and the current circuit is already fancy for me
- I may add a second slip ring on the other side of the arm and have 4-6 wires for power

The last one seems to be the easiest to implement since there's room for the second slip ring (right side on the picture).

So, I suppose the biggest problem I have right now is the wobble.

Here's some assembly I got to at some point:

[cyberKerb]

Someone posted this a few months ago and another kind soul reminded people on Reddit today. First time I'd seem it and figured it needed to be added here.
So inspiring for my own project - love the work for a controller you can actually SIT in.

(French) Build page here: https://www.tutomotique.fr/fabriquer-un-cockpit-diy-pour-ksp/

 

 

Edit:  Crikey - they keep coming!  Another one was linked in the above page:

(French) Build log: https://www.oximoron.fr/R/OxiKspController/controleur-de-jeu-pour-kerbal-space-program 
Awesome custom momentary buttons they made.

 

 

Edit2:
I updated the collection post with the 4 recent projects from this forum that I've seen.  Up to 58 controllers now!

 

Edited October 22, 2019 by cyberKerb

[richfiles]

On 9/18/2019 at 4:47 PM, tsaven said:

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.

Where did you get these printed. I'd like to get some of these. I had a shopping cart over at Shapeways with a bunch of them, but they have recently DRASTICALLY increased their prices (my $60 cart is now $300). I have a creeping feeling that they are maybe self destructing... 

[mountaineerman]

Wow, that immersive cockpit is fantastic!

[c4ooo]

Speaking of reddit, I thought someone should mention the Kerbal Controllers subreddit in this thread. Seems to be somewhat active.

[Francois]

I always assume it's the same people reading, so I'm not sure where to post. I'll probably just post my finished project there.

I should have some interesting milestone to show in a week or two if everything goes well

And regarding that immersive cockpit, it's definitely inspiring. I hope to make something like that in the end, but with 10 times more panels  this one, although not Kerbal, is also interesting https://www.youtube.com/watch?v=xRoCDYLanTA&list=PLAY3QA4kl1ESoRGRkgAYZOuiczjOa1WZ5&index=2&t=0s

[c4ooo]

On 11/21/2016 at 3:34 AM, richfiles said:

It's also possible to make a T-Bar fader fairly simply. You just need to make a lever that you can mount and move easily, that attaches to a potentiometer. One style I've not seen recreated is a linear potentiometer based T-Bar fader. One of the issues is making the pivot able to both stand mechanical stress, but also attach to a potentiometer. Most people just mount the pivot of the meter to a potentiometer, and call it good. On option you could do is to have the pivot me more mechanically robust by bolting it between two secure structural parts (like a sandwich, with the lever in the middle). You could also have friction pads at the pivot, and adjust the tightness of the lever by tightening the bolt (use a bolt with a self locking nylon insert nut to make sure this assembly doesn't loosen or fall apart). Then, to make the business part of the assembly, you have a slot in the lever that fits the slider of a linear potentiometer through it (or you can use a connecting rod style connection, either way works)

One problem with this deign is that the the potentiometer position (X) isn't 1 to 1 with the lever angle (Theta). Because of this,  when the lever is far from the center, dX/dTheta is going to be bigger than when the lever is at the center. (Same change in angle is going to result in bigger change in pot position at the edge than at the center)

To get the "true" lever position you would have to do something along the lines of Theta = atan(X/L), where L is the distance from the pivot to the potentiometer track (when the lever is perpendicular to potentiometer track). 

(sorry for 3 year late reply )

Edited November 2, 2019 by c4ooo

[richfiles]

On 11/2/2019 at 12:58 PM, c4ooo said:

One problem with this deign is that the the potentiometer position (X) isn't 1 to 1 with the lever angle (Theta).

That's technically correct (the best kind of correct!  ) Putting it into practice though, it's not necessarily that big of an issue, unless you have a REALLY long slider potentiometer. A shorter pot won't have as much of an error. Don't forget that you are also a part of the feedback loop. As you adjust the lever to adjust the throttle, your adjustment will fine tune the lever's position based on your visual observation of the throttle setting on screen. You'd be shocked when looking at consumer devices, how few knobs actually line up with their scales. Any user controlled analog knob or lever will always be accurate, not based on linearity or scales, but because of user input derived from observation. The operator forms a closed loop feedback system.

If you have an absolute desire to correct for the error, then that's easiest done in code. You could use a simple lookup table to return a calibrated throttle value based on the analog value. Wouldn't even need to worry about math, unless you want to. You could certainly incorporate the math in software instead.

Edited November 10, 2019 by richfiles

[c4ooo]

19 hours ago, richfiles said:

That's technically correct (the best kind of correct!  ) Putting it into practice though, it's not necessarily that big of an issue, unless you have a REALLY long slider potentiometer. A shorter pot won't have as much of an error. Don't forget that you are also a part of the feedback loop. As you adjust the lever to adjust the throttle, your adjustment will fine tune the lever's position based on your visual observation of the throttle setting on screen. You'd be shocked when looking at consumer devices, how few knobs actually line up with their scales. Any user controlled analog knob or lever will always be accurate, not based on linearity or scales, but because of user input derived from observation. The operator forms a closed loop feedback system.

If you have an absolute desire to correct for the error, then that's easiest done in code. You could use a simple lookup table to return a calibrated throttle value based on the analog value. Wouldn't even need to worry about math, unless you want to. You could certainly incorporate the math in software instead.

Yea, just wanted to point this out

In my panel design (yes, after 4 years I finally started soldering stuff together ), all processing is done on PC and the arduino just acts as a dumb proxy between the PC and SPI led drivers, voltmeters, and switch matrix, so implementing this correction will be easy to do. 

[Cassowary]

These are so cool, this community is amazing.

[cyberKerb]

two more consoles recently posted to reddit. If there is a local forum build thread let me know and I'll add the reference to the post I'm maintaining here.

[c4ooo]

I'm trying to write up my progress here 

 

[Kerbal007]

If anyone is interested, I pulled apart my farming sim controller from logitec to see if there was anything special about it etc and to learn a bit. 

Unfortnatley it did not reveal much and the micro controller is rare if anything. The markers are "MP01006B, DFJ115, 1810".

The axis controller, is essentially two pots and maybe a hall effect sensor ? and some wiring for the buttons. 

https://imgur.com/gallery/3ss3tEu

[richfiles]

On 2/8/2020 at 8:57 PM, Kerbal007 said:

If anyone is interested, I pulled apart my farming sim controller from logitec...  https://imgur.com/gallery/3ss3tEu

Have you used it as a controller in KSP? I had suggested it as a possible well suited controller for KSP quite some time back, but I didn't own one. I'm actually curious as to how well it works, if you've tried.

[Kerbal007]

On 3/15/2020 at 10:37 PM, richfiles said:

Have you used it as a controller in KSP? I had suggested it as a possible well suited controller for KSP quite some time back, but I didn't own one. I'm actually curious as to how well it works, if you've tried.

Its ok, just buttons extra buttons really. I think that 6dof mouse (cant recall its name) would actually be better for control. Personally id rather build my own control for buttons etc

[richfiles]

This is the controller with he small stick at the bottom, right? have you tried using it for RCS?
I've always imagined this controller paired with a flight stick for attitude control would be pretty decent for the type not interested in building a controller outright.

[sandroxil]

Hello. Anyone has used KSPSerialIO with a mod that allows to use more action groups, like Action Groups ReExtended?

Sorry, but while I'm working with my Kerbal Controller, I see that I have not enough with 10 action groups.
I need MOAARRRRR ACTION GROUPS

[Francois]

II'm surprised to not see more activity here in these confined times.

Just to let you know that my navball is not progressing though definitely not abandoned (got distracted by an apartment move, life and thinking that I should read a 1000 pages book about electronics first :p).

But to get an easier milestone first, I started a small controller.

Work in progress. The fuel/life support levels work, the LCD partially works, I still need to program the LEDs, wire and program the switches, wire and program the 7-segment display and finally, finish the design in Fusion 360 and the CAM path to cut and drill the sheet of acrylic to replace the cardboard.

[gargrag]

On 12/19/2017 at 5:10 PM, ajden said:

Well here's mine.

Hey This is an spectacular job, I'm starting to design my controller, and I'm loving the way your LCD mimics the apollo numbers, can you tell me a little bit more about it? which LCD it is, it is connected through i2c? and the visigenie soft for the digits how did yo added that?

[richfiles]

On 5/27/2020 at 12:09 PM, Francois said:

I'm surprised to not see more activity here in these confined times.

Problem is, I've been trying to get projects done, but the usual few weeks to a month wait on some parts has become months... Manufacturers and shipping have been heavily impacted too, so I'm suborbital without a chute on many of my projects right now. It's incredibly frustrating! I currently have 3 big projects that are all stalled.

Obviously, there's my Kerbal Space Program controller. I need to get plenty of small bits and pieces for the DSKY, for the Tape meter, and for the ∆V meter.

My second waiting project is the LITERAL FINAL PART needed to complete my custom mechanical keyboard is waiting on a 20-pin magnetic USB-C connector,
which will be repurposed to attach a magnetically detachable number pad.

Finally, I've got a custom monitor project for my workbench too. I'm waiting for a GoPro bike swivel mount, which I'll be using to suspend a pair of iPad LCD modules in a custom frame, from a slide rail mounted underneath the top shelf over my workbench. It means I can slide my monitor off to the side and swivel it to a comfortable viewing angle, and play Youtube videos, movies, TV, etc, if I feel like playing something in the background while I work... Or I can slide it toward the center of the bench, angle it and have a pair of iPad resolution displays, at eye level, for displaying schematics and diagrams on. When centered, it's just above my workspace, with no blocking of the bench beneath. That project is on hold, cause the swivel/tilt mount I bought for it has been tied up in shipping for the better part of a month.

It's definitely frustrating to have ideas in your head, and the time to do them, but suddenly find getting certain parts is now much more difficult.

[sandroxil]

On 5/27/2020 at 7:09 PM, Francois said:

The fuel/life support levels work

[snip]
Life support???

Really???
Now I have seen it, and I want it.
I have to include it in my controller. 

I supose you use the same LED bars to show the fuel levels and the life support levels, including some switch to change between both modes. It's a good idea.

Every time I see the posts about controllers, I have more and more work to do.

About the life support levels...
¿Are you using KSPSerialIO?
The life support levels are not included in the "Vessel Data Packet". How did you included it?

Edited June 21, 2020 by sandroxil
Portions Redacted by Moderator

[Francois]

On 6/20/2020 at 10:01 AM, sandroxil said:

About the life support levels...
¿Are you using KSPSerialIO?
The life support levels are not included in the "Vessel Data Packet". How did you included it?

I use KRPC to get the data and then I made up a not so great protocol to send the data over serial to the arduino.

I'm realizing that I haven't pushed my code to Github in a while (also I haven't touched my project in a few weeks now).

I'll push the latest next time I start my Windows PC https://github.com/fgaudin/protoconsole-python/blob/master/protoconsole.py. Oxygen and other LS are available the same way as LiquidFuel from what I remember.

[Francois]

I'm making some progress on my controller. I was stuck on something that was looking so dumb that I started reading entire books about electronics to learn more on the basics and hoping that it would become clearer. But no clue.

I'm controlling all the LEDs of the controller with a Max7219 (8x8 matrix controller). It was working flawlessly with regular LEDs, but it was totally not with those LED switches:

I realized each had actually 2 LEDs mounted anti-parallel (I didn't know what it was before) which doesn't work with a max7219. Long story of the surgery here: https://imgur.com/gallery/Wtqx2Ai

Now that it's fixed, I'm finishing to solder the board that controls the LEDs and I'll switch to wiring/coding the switches. Last will be a 7-segment display before it's functional.
I will then have to do the panel using my mini CNC and I will probably rework the code because I'm not too happy with it.

Right now, it's controlled by a single Arduino but I'm considering dedicating a Nano to the switches because it seems that the LCD Display updates are very slow which hinders detecting when a momentary switch is pressed.

[richfiles]

Don't be afraid to segment the project into any number of dedicated microcontrollers.
The code for each one will be simpler, and it will indeed make each one more responsive.
I have no idea how many arduinos my instrument panel will have when finished.

[Francois]

I probably will eventually, but I'm definitely learning under constraint  weirdly, I managed to make this solution work with a 74hc165 instead of a 74HC595. I need to post the schematic so anyone can point out the flaw.