y2kiah's A-10C cockpit build

Can I ask why you're only cutting the templates on the CNC and not the actual ribs?

For me the reason is time. On the machine each rib would take 4 or 5 times longer to cut out and stress the machine much harder. It would probably be machining for over an hour per rib. In that length of time, it is very prone to losing steps and getting out of position, requiring frequent re-zero, and that turns the 1 hour job into 2 hours. Plus if it screws up the cut, I'm not ruining a $20 piece of ply instead just a $6 piece of hardboard. If I get a faster machine that I don't have to babysit in the future, I'd probably fire the job off and go do something else, but for now a jig saw will make short work of those ribs.

Makes perfect sense. Just curious if you had discovered a special technique. I have seen woodworking plans where they use a template with several pieces stacked together and use a bandsaw to cut the stack all at once. I could see where that would save a lot of time.

what font are you using on you panels?

Some that have been recommended on other forms

You will notice that some have rounded ends and some don't

 

F-16 panel font

MS 33558 font

Trade Gothic 18

So far I've been using Franklin Gothic Medium on the panels, would be interesting to see how they look in those other fonts though

I found a website that has MS 33558 as a True Type Font File, and a whole bunch of other stuff.

http://www.simpits.org/fileproc/showfiles.php

Hi guys, it's been a while since I've posted in my thread. Thanks Avilator for saving me from having to necro my own thread.

Anyway, I thought I would post pictures of some of the panels I am shipping out. I wish I had a laser engraver like some others, but I made do with what I have. Sorry the pictures are kind of yellow.

These are 1/4" thick panels of layered acrylic, with 1/16" aluminum back plates. They can be back lit, except for the ground override panel. All of the lettering was filled with bright white after engraving, and they are sealed with a satin clear coat. They are shown sitting on rails made from 1"x.5" aluminum angle.

I said in an earlier post that I would show some pictures of my painting station. It's nothing fancy, just very cheap and easy to make. It can be folded flat for storage.

The parts you need are:

1) 5 pieces of foam board

2) duck tape

3) air filter

4) 2 safety pins - used to secure top to the sides when in use

The cheap box fan sits behind the air filter and pulls air through the filter, and none of the paint spray makes it through.

Here's a before and after painting the small screws that hold the panels to their back plates.

Those look awesome! And your painting station is very clever!

Looks like you found a good process for engraving your panels. I never had much luck as some areas of the panels would not get engraved as deeply as others. I suspect my bed was just not level, but I was never able to get it perfect.

Anyway, great work!! :thumbup::thumbup:

By the way, how's your project with the arduino card coming along?

You did a great job for using a Cnc! I never had much luck either...

Nice update! And timely, too - I think I killed some brain cells by spray painting over the weekend. A fan and air filter just made it onto my shopping list. Oh, and I'm stealing the screw painting idea, too ... just thought you should know. ;)

Those panels look fantastic.

How exactly can the panels be backlit? Do they have space to embed leds inside the acrylic?

edit: Do you have a website to sell your panels?

Thanks guys!

By the way, how's your project with the arduino card coming along?

It's coming along, I haven't looked at it in a little while honestly. The DCS lua stuff is done. The arduino firmware is pretty much done, I need to get around to release that so people can start using it. That's only one piece of the puzzle though. I'm working on the saving/loading of project files. When you save a project in the editor, an XML file is created that is in read by the server. It's so boring to code that serialization stuff though, ugh...

Oh, and I'm stealing the screw painting idea, too

Cool but if you can find a flat piece of foam you'll be better off. The cup got a little hard to manage towards the end.

How exactly can the panels be backlit? Do they have space to embed leds inside the acrylic?

edit: Do you have a website to sell your panels?

one easy way is EL sheet, and the other way is to drill holes for LEDs. I didn't pre-drill anything because everyone seems to have their own solution for back lighting.

I haven't set up any website for selling my panels, I'm kind of shying away from taking orders and think I'll just produce one-offs and sell them at my own pace instead. At least until I can get a quicker production line going. Right now there is just too much manual labor involved with creating these things.

BTW one of the things I'm kind of waiting on is the release of a new board that the Arduino guys are working on with a built-in ethernet connection instead of a USB connection. Read about it here.

Right now I use a Wiznet ethernet shield for connectivity, so I would save around $30 per board when the new one is released. The TFTP feature is very interesting because I'll be able to upload new firmware via ethernet, so making updates to the firmware will be much easier, and could even be automated.

I've been mucking around with electronics the last few days. I've been debating between Arduino and building my own PIC based boards.

I'd like to have each panel independent, but the cheap Arduino's only have 14 I/O pins which is not enough to run a panel. If I end up doing a PCB it ends up cheaper to just design my own PIC based PCB. How are you dealing with this?

I'm using the Seeeduino Mega which basically has more of everything for cheaper, and it's really small. Price-per-pin it's the best Arduino deal going, but shipping from them took a month and one never arrived at all (they did send another though).

Of course, if you're going to build your own PCBs anyway, Arduino being open source you could always use their design and maybe produce it cheaper, not sure on the economics of that though.

To increase my effective number of pins, I use an "off" position for each digital switch so the absence of an active signal on all other switch pins indicates that the "off" position must be active, and I send that event. So I only need 1 pin for a two position toggle, 2 for a 3-position toggle... and so on, I'm sure you get the point.

You can use shift registers or a matrix setup to increase effective pins as well. For the CDU keyboard I plan on using one or the other.

For LEDs I'm multiplexing them, so I only need 3 pins (SPI interface) to work with the MAX7219/21 chip for 8 x 7-segs or up to 64 individual LEDs in a matrix.

The biggest limitation I see so far is for rotary encoders. There are a lot of them in the pit and you need real-time interrupt pins to read encoders (again unless you build a separate circuit for those). Those are the most limited pin types on the Arduino board, only a few, maybe 4 on the mega, can't remember. I have looked for a good encoder shield that could for example read up to 8 encoders on 3 SPI pins. I haven't found one yet, that would be a great thing to design. Without a shield like that, you basically need one board per panel with encoders anyway.

I was thinking about doing self contained panels vs. a central box for the boards, honestly if money were no object I'd give every panel its own micro and that would simplify things a lot. With the mega being overkill for most panels, I thought it would be better to let panels share a single board so I plan to have a few per side and a few for the front, however many are required - kind of a semi-distributed setup.

Those panels looking great :thumbup: and same goes for the paint box. Normaly I wait for a less windy day and make the paintjob outside but results is less then good.

Bought an Arduino from eBay (together with some tutorials) some weeks back but haven't yet started it up. Target is to use it to interface with the engine instrument air-core (when and if I get there). Future will tell. Looking forward to see your progress. More pics please :-)

/Gus

Another idea for lighting besides discrete LEDs or EL sheet would be LED light strips

http://store.makerbot.com/1-meter-led-stripe-green.html

These have a self adhesive back so I'm picturing just routing out a channel in the back of the light plate where you would route your strip, and sticking the strip directly onto the back plate.

Is there any good and cheap 3D modeler around for a nood. I want to plan my cockpit before getting into it.

Google Sketchup is free! :)

Is there any good and cheap 3D modeler around for a nood. I want to plan my cockpit before getting into it.

Sketchup, like pitbldr said ... and cardboard.

Good i already downloaded it before asking. God to know i did the right choice on an whim.

Thank you guys :thumbup:

y2kiah, your panels looks really good! So is your paint station!

To increase my effective number of pins, I use an "off" position for each digital switch so the absence of an active signal on all other switch pins indicates that the "off" position must be active, and I send that event. So I only need 1 pin for a two position toggle, 2 for a 3-position toggle... and so on, I'm sure you get the point.

Hi y2kiah,

Of course you only ever need 1 pin for a 2-position switch. And 2 for a 3-position switch. But for more positions like an 8 position rotary switch you should wire them all. Reason is it's hard to tell the difference between all off and when you're just between pins. 3-way is a special case because the off position is between the 2 wired pins anyway.

I was thinking about doing self contained panels vs. a central box for the boards, honestly if money were no object I'd give every panel its own micro and that would simplify things a lot.

This is definitly the way to IMHO. With PICs at about $2 per, the micro is the cheapest part. The first time I did this I custom coded for the lua commands I wanted from that panel. The proved to be non optimal time-wise. Now I have a generalised PIC board that can be very quickly configured for any switch type without programming. I just timed it and it took all of 3 minutes and 45 seconds to set up a blank board for 1 each of:

momentary switch

toggle switch

3-position toggle switch

5-position rotary switch

2-bit incremental encoder

4-bit binary encoder

potentiometer

That's 16 inputs total. Not so bad really.

Cheers,

Colin

PanelBuilder how are you managing the PCBs when doing that? Are you manually wiring everything or making a PCB per panel?

If per panel are you etching them yourself or where are you getting them done?

Thanks for your input Colin,

Hi y2kiah,

But for more positions like an 8 position rotary switch you should wire them all. Reason is it's hard to tell the difference between all off and when you're just between pins

I haven't had any problems doing this with rotary switches, I debounce input not only for each pin, but also across all switch positions, so it resolves the issue with break-before-make switches. Granted it's only 1 pin saved but it's an easy fix.

I think 1 micro per panel is the right solution for the more complicated panels, but a lot of panels are nothing more than a few toggles and rotaries, and a dedicated micro just seems overkill for those. It's really a give and take. One one side, you complicate the wiring to each panel, on the other side you complicate your communication network back to the PC.

PanelBuilder how are you managing the PCBs when doing that? Are you manually wiring everything or making a PCB per panel?

 

If per panel are you etching them yourself or where are you getting them done?

PCB per panel.

I've etched my own in the past. Life's too short. Now I use these guys:

http://www.olimex.com/pcb/index.html

I use this for layout:

http://www.freepcb.com/

Cheers,

Colin

I think 1 micro per panel is the right solution for the more complicated panels, but a lot of panels are nothing more than a few toggles and rotaries, and a dedicated micro just seems overkill for those.

Well I use a 16-input board, a 10-input board, and I'm working on a 4-input board.

For panels with just a lot of momentaries, I think there's still a place for keyboard emulator boards -- that's what they're good at. For example, for that full keyboard panel I can see a certain elegance in sending the matching keystroke for each button.

I have an old Hagstrom KE-USB36 that I would probably use for that.

Cheers,

Colin