Jump to content

Kerbal Instrument Panel: In-Desk Apollo Themed Hardware Controller


richfiles

Recommended Posts

  • 2 months later...
On 8/17/2018 at 11:59 PM, SpaceN00b said:

Where did you find your 8 ball?

Yeah, the repository has a lot of info that never made it here.
Ebay is a fine resource, if you can afford the bill. Mine specifically was imported from Israel, formerly used in a P-4 Phantom flight simulator.

I'm still alive, but work is infuriating... I work two jobs, both part time...
This past week (not pay period... just ONE week) I got a nice 23 hours on my second job...Not bad... 51 hours at my main job though.. "Part time" :mad:
That's part time, with no benefits, I might add. I've stayed part time cause I value my time more than a few extra bucks, and my health insurance is currently affordable, with both reasonably priced premiums and co-pays, but is also strictly available only to those who don't have access to employer provided health insurance plans... Something that going full time would disqualify me from. The "idea" of more free time and better health insurance is worth staying part time... Except my employer has obviously ignored this. Today they asked me to start coming in an hour earlier from now on. I've been pushing for another employee for my area since November.


The ONLY reason I'm still there, is it's the most stable job I've ever had. I can't deny that. Government regulated food testing is a big business, and it's not realistically possible to export it... Customers require fast testing and results. Shipping items every single day, expecting it to start testing same day isn't feasible without local services. That's why we never have downturns. I've been there 5 and a half years, and there's never been a downturn beyond seasonal fluctuations in food production. It only goes between normal and heavy. I don't thing the job is going anywhere unless either the government goes belly up, or people stop eating. Either one would be a far worse scenario than a job search! :/

That said... If i could get enough hours at my secondary job, I'd ditch the lab in a heartbeat. Problem is, that job isn't longterm stable... I can have tons of work... Or I can go months with no orders. That's why I need the stable job. Doesn't change the frustration with balancing the most stable job I've ever had, vs a company that's CLEARLY EXPLOITING me.

Edited by richfiles
Link to comment
Share on other sites

On 8/23/2018 at 2:25 AM, richfiles said:

This past week (not pay period... just ONE week) I got a nice 23 hours on my second job...Not bad... 51 hours at my main job though.. "Part time" :mad:

Dude do you atleast get paid extra for more than 40 hours a week? 

Link to comment
Share on other sites

Yeah, I get OT pay, which is nice... At least there is room for flexibility in hours... Truth is, the hours tend to be quite flexible. If I say I need to do something before work, typically I am able to do so...

Originally, I used to be scheduled a strict 8 am to 12 pm shift. Due to changes in when we received customer samples to test, they decided they wanted me to come in from 10 am to 2 pm. Problem was, there was no adjustment to pay, no increase in hours... They just wanted to shift my time by two hours. Over the course of 5 days a week, that would eliminate 10 hours of my afternoon, with no compensation for me. I argued against it, and suggested I come in at 9 am, but be permitted to come earlier, or stay later, as needed, when additional labor was needed, and when I could actually do so without interfering with my other jobs... It was great... At first. Then they asked that we "check out" before clocking out. This was often cumbersome for me, as both my supervisors took lunch during the hour I normally would clock out at. Then things shifted from, "Okay, see you tomorrow", to "Okay, is everyone in your area fine with you taking off?", and then to ""Okay, is everything done in your area?" (That's what the full timers are for, aren't they!?! To stay a full shift and finish up!?). Now it's become not replacing employees who are no longer working there, and asking for even more hours from only two people. We don't even have enough people to run alternating Sunday shifts anymore... So they eliminated working Sundays, and now stack Sunday's 6 hours of work for 2 people on top of what we already gotta do on Monday...

I'm just so tired of it. At some point, I need to skip my department supervisors and just go into the office and I dunno... file a complaint or something.

Edited by richfiles
Link to comment
Share on other sites

  • 10 months later...

Man, another year gone... Yikes! Anyway, I've been getting the itch to build again, and some stuff showed up recently that I wanted to share.

I recently received the Godspeed Minuteman key cap set from the company formerly known as Massdrop (why would they change a recognizable name to something as meh as "Drop"?). Honestly, I had forgotten I'd ordered them. Anyway, they are a space themed low profile "DSA" key cap set made from glow in the dark plastic. I ordered some 405nm UV LEDs recently, and experimented with illuminating them from underneath. The results speak for themselves...

yBsmQDg.jpgIZt7mF0.jpg

The UV LEDs have an interesting pattern in the camera. The green glow is the plastic glowing, not simply light shining through. The UV light fluoresces the plastic.

Anyway, while I have definitely been out of the loop for a while, I'm not gone. I have had even more practice working with KiCAD, and have no doubt that the FDAI controller boards will be more than possible to produce. I need to finish up a few things that are already started, such as modding my analog meters and finishing the DSKY diode ROMs (I still have one left to make).

Anyway, the reason I'm talking about these glowy green mystery goo keys, is I've decided to return to the original keyboard concept for the controller. The original plan was to mount an Apple //c keyboard into the controller. Then I got the Danger Zone key caps and built my custom keyboard using those. My revised plan was to simply have a notch in the controller that the keyboard would slide into when the desk's keyboard tray was retracted. Now that 6 years have passed since I built my computer, I realize it's getting to that time when I want to build a new one, and my workshop definitely needs a computer. My plan is to basically migrate the old computer to the workshop, and build a new one as my main computer. This brings up some interesting questions... Could my new computer be squeezed inside the kerbal controller itself? My existing computer is bare right now, and the Philco Radio case is in the workshop getting work done... It looks GOOD sitting on the desk in there. I'm half tempted to leave it in there, but I DO like the thing to be more visible too... Thing is, it tends to be better lit and at eye level in the workshop, while it has to sit on a second level upper desk shelf to be in here. Anyway, regardless of what I end up doing, I'll be needing a Keyboard for each computer, so I may as well go back to building the keyboard into the Kerbal instrument panel. I'll still keep a retractable keyboard tray, and still have my Danger Zone keyboard, but it'll be stored away when not in use. I've actually got a set of typewriter styled keys that I'm seriously considering building up for the Philco Radio Computer. All that's up to whatever happens in the future.

Next big thing to do is to machine the front panel. It has to get done, no question. One of my roadblocks is the fact that all my parts are well... Parts. Once things start fitting together, I can actually begin wiring, and once I begin wiring, I can actually begin to figure out how to make every part of the system function. The front panel is too big for my milling machine, so I need to deal with it in one of two ways... I need to either split it and machine each half separately, or finally try to get it done at the machine shop. My preference is to keep it in one piece. There's gonna be a lot of measuring to make it all fit. I am also definitely going to need to machine out a notch on the edge of one analog meter to fit it around the internal support frame for the front frame. I need EVERY MILLIMETER of space to fit everything... That's not a joke. I'm even probably going to have to mill a notch out of the outer chassis mount on the other side to make room for yet another meter to skirt right up to the edge.

Another roadblock ended up being the tape meter. The screen printing shop that I went to completely flaked out on me. I guess they decided a one off job wasn't worth the effort. I can give it another shot, but it looks like I'll be forced to use those stamp kits and build some manner of jig and do the whole thing manually. That's not gonna be very fun... But, it's either that, or find another screen printing shop that's willing to take the job. I REALLY want the tape meter. It's such rare piece, and I wanna make it work, I wanna make it fit... I just really wanna make it a reality.

So yeah... Popping back on the radar here. Plan for now is to finish steps that I started on but haven't completed. Top priority is finishing the last Diode ROMs for the DSKY, finishing and backlighting ALL the analog meters, and making progress on the main front panel. Once that's all done, I'll move forward from there.

 

Edited by richfiles
Link to comment
Share on other sites

Dude, this whole thread is mind-blowing.  I thought I was a big nerd when I was just soldering some headers onto pre-made arduino shields!  I'm light-years away from anything you're doing!

Link to comment
Share on other sites

Curious as to where you got those little 16-segment displays?  And were you able to get a driver board for them, or do you have to address each individual LED segment individually?

I'm planning on using a bunch of these https://www.sparkfun.com/products/11441

for driving my displays, as it's much easier to just send a string to an Ic2 address than it is to keep track of which LEDs need to be lit for what.  But I haven't been able to find a pre-made driver, or any 16-segment displays that are in stock.

Link to comment
Share on other sites

On 7/26/2019 at 1:43 PM, tsaven said:

Dude, this whole thread is mind-blowing.

Thanks!

On 7/26/2019 at 5:51 PM, tsaven said:

Curious as to where you got those little 16-segment displays?

I got 'em surplus, but look at part suppliers like https://www.digikey.com (my favorite one), https://www.mouser.com, and https://www.newark.com for individual components. Those are all US based suppliers, but they ship internationally. You can also look up regional suppliers. For surplus, I mostly use https://www.allelectronics.com and https://www.goldmine-elec-products.com

So, for me, I built the diode ROM to control the 14 segment displays just because I enjoy hardware, and retro stuff like that. I could totally have driven them from an Arduino or something, and it absolutely would have worked, but I simply chose a different route. the issue is most of the serial to row/column driver chips aren't meant for driving over 8 segments. The popular (and cheap) MAX7219 and 7221 chips are not really able to drive 14 or 16 segment displays. You could get away with it using shift registers, or even port expander chips, or even just using the digital I/O directly. You may need a transistor to drive the displays, especially if you scan them. Fortunately, you only need to cover the number of rows and digits with a transistor driving each line, but you'll always need to control it by switching segments. There ARE controllers for 16 segment displays, like the Maxim MAX6954, but they tend to be expensive... Mouser has them starting at $18.76 per chip, and they can drive up to 16 digits of 7-segment displays, 8 digits of 14 or 16-segment displays, or 128 individual LEDs. If you do get one of these chips, consider doing like I did, and ONLY place the 14 segment displays where you NEED them, and let the rest be 7 segment, That lets you use the cheaper, and much more common Max7219 on those.

Since microcontrollers are dirt cheap these days, I honestly recommend using one controller as a display driver if you have any concerns over whether or not you can support the timing. Have it handle the segment switching and scanning, and it's input would be the plaintext data transmitted by whatever is handling that. I absolutely plan on spamming controllers in my own project. Maybe it's possible to efficiently code one controller to multitask between several things... Or I can keep the code simple and have a controller dedicated to one simple task.

As an example, that's my plan for the FDAI controllers. Each axis will have it's own separate Arduino micro handling the packet input, and calculating the amplitude and polarity output that corresponds to that particular angle, and send that data tot eh DAC and analog switch IC. Since each axis is identical, with the only difference being the piece of data it reads out of the packet for updates, I can literally create code for one, and just copy it onto three microcontrollers, changing only the location in the data packet that it reads from. Depending on the memory available, I might be able to get away with an amplitude value lookup table rather than calculate it. Absolute value the number into a variable, add them together, if the result is 0, set the polarity output to represent negative, else set it to represent positive. So basically,

I'd start with whatever command(s) lets me read specific data out of an incoming serial packet and store that data to a variable (we'll use PacketDataA).

PacketDataB = PacketDataA + 120
PacketDataC = PacketDataA + 240
If PacketDataB >= 360
 Then PacketDataB = PacketDataB - 360
If PacketDataC >= 360
 Then PacketDataC = PacketDataC - 360

After generating these three angles based on the angle sent by the packet, we want to then convert that to a sine value that corresponds to the attenuation at that angle of a syncro control transformer. As I said, a lookup table probably makes a lot of sense here. I'd need to take the angle values in the three PacketData(A, B, & C) variables, and read out attenuation values out of the lookup table into three additional variables. We'll call them AttenA, AttenB, and AttenC.
 

B = Abs(AttenA)
C = AttenA + B
If C = 0
Then PolarityPin = 1
Else PolarityPin = 0

... And basically repeat that three times total for each of the three attenuator values and their polarity.
Repeat all of the above three times, to get the outputs required for all three axes.

Pardon my lack of coding skills, my foundations are in BASIC, so it's not gonna be syntax correct for an Arduino, but I think the math is solid. I imagine it might be possible to easily code this all on a single chip, but why make myself suffer if I don't need too. I imagine I could maybe just run it all on one, and it'll be fast enough to keep the FDAI updated, but honestly, either way is fine by me.

So, back to the 14/16 segment LED displays. You can use a dedicated controller like the MAX6954, or you can drive segments with row and column drive. If you're better with programming, and want to keep things simple, get either 14 or 16 digit displays and control up to eight at a time, per MAX6954 chip. It's nearly $20 per chip, but at least they have a built in character font for the displays, and you can send the chip plain ASCII characters to display over SPI. I think the MAX7219 chips are simpler, and about half the price (cheaper ones are available on ebay, fully assembled on a PCB, with LEDs, but beware the quality and test them first).

Edited by richfiles
Link to comment
Share on other sites

  • 3 weeks later...
On 7/27/2019 at 2:37 AM, richfiles said:

Nerd stuff

Jesus dude you are like so far beyond the stuff I'm trying to do :confused:

As a warning, because you seem . . . fairly . . . well versed in hardware stuff, I might end up pestering you a fair bit in the near future because jeeze do I need help.  I've got a very tenuous understanding of basic electronics and have played around with some Arduino's a few years ago to make very basic projects, but I'm embarking on a far too advanced KSP Controller build myself as a learning experience and to keep me out of trouble.  That layout isn't close to finished yet and I'm worried I'm going to have to expand the board because I keep adding features that I want...

To be on topic, regarding driving 16-segment displays; I was able to find this useful little $6 item to drive 14-segment displays, which I think will do just fine: https://www.adafruit.com/product/1910

Your suggestion of only using the 14-segment displays for indication is exactly my plan, specifically to tell me if my AP/PE readouts are in meters, kilometers or more.  And while I'll have to connect one of the two-digit segments to the controller board with some jumper wire, I can easily use that board to drive two separate 2-digit 14-segment displays for each of my indicators.  And it works over IC2, so I can probably wire them up right in series with all of my other read outs for data as long as I'm careful about the addressing

I have another question for you; where the heck does one get those relegendable backlit push buttons that you said you salvaged from an old video controller board?  Those sort of things are exactly what I'm trying to find for my project, or at least a reasonable emulation of them. I've spent most of the weekend combing through Mouser & Digikey and haven't had much luck; the few options I've found when looking through switch maker's catalogs are not in stock anywhere and are listed simply as "Call us for a quote" which usually means "We're not interested unless you're buying 10,000 of them".  Or they're $25/ea and that's a lot to swallow when I'm looking at needing *checks* at least 35 of them.

I mean, really I'm hoping to find a way to emulate the look and feel of the Tellite switches that were used on the Apollo consoles, but those are designed and built for aircraft and I've been around enough airplanes to know how much anything that flies costs.

I had hoped I could emulate them pretty well using Cherry MX switches and these square flat keys, but sadly they're not nearly translucent enough to shine sufficient light from an LED through.  Plus I'd have to come up with a way of labeling them.

Right now I've been checking out rectangular buttons from arcade machines, which are readily available for less than $5/ea but have other problems to them in terms of the look at feel.

 

Link to comment
Share on other sites

That Adafruit board looks perfect for the job. Nice controller plans too! :D

Regarding the backlit keys, I think they were about $6 each when new, but all mine are salvaged. I have a bunch of extras. I think 107 of the smaller ones and I think 62 of the big ones. The small ones are about 0.45 inches square, and the large are about 0.7 inches square. They all support a 2 lead 3mm LED. Mine have red/green opposing polarity LEDs installed, which allows the creation of red, green, and any yellow or orange if I PWM it. You can fit any color 3mm 2 leaded LED in the switches. Given the difficulty in acquiring them, I'd really prefer to hold on to some of the larger ones. The small ones, I'm willing to let go of more readily, since I have so many. I know they are smaller than the Apollo ones, but you're also not wearing a space suit, so they would probably work. :wink: PM me if you want.

Man, $25 a switch is nothing... I once price checked the Honeywell tab lever toggles... Oh boy... Three digit price tag... Each! :0.0:

It drives me mad that such an iconic toggle design hasnt been mimicked by the Chinese switch makers. Sure, they got plenty of Bat lever and paddle lever toggles, but nothing like the textured tab levers. You can get 3D printed caps from places like Shapeways though, to epoxy to a standard toggle. You just gotta find a toggle with a lever that doesnt spin. Some of the cheaper ones will spin.

Link to comment
Share on other sites

7 hours ago, richfiles said:

That Adafruit board looks perfect for the job. Nice controller plans too! :D

Regarding the backlit keys, I think they were about $6 each when new, but all mine are salvaged. I have a bunch of extras. I think 107 of the smaller ones and I think 62 of the big ones. The small ones are about 0.45 inches square, and the large are about 0.7 inches square. They all support a 2 lead 3mm LED. Mine have red/green opposing polarity LEDs installed, which allows the creation of red, green, and any yellow or orange if I PWM it. You can fit any color 3mm 2 leaded LED in the switches. Given the difficulty in acquiring them, I'd really prefer to hold on to some of the larger ones. The small ones, I'm willing to let go of more readily, since I have so many. I know they are smaller than the Apollo ones, but you're also not wearing a space suit, so they would probably work. :wink: PM me if you want.

Man, $25 a switch is nothing... I once price checked the Honeywell tab lever toggles... Oh boy... Three digit price tag... Each! :0.0:

It drives me mad that such an iconic toggle design hasnt been mimicked by the Chinese switch makers. Sure, they got plenty of Bat lever and paddle lever toggles, but nothing like the textured tab levers. You can get 3D printed caps from places like Shapeways though, to epoxy to a standard toggle. You just gotta find a toggle with a lever that doesnt spin. Some of the cheaper ones will spin.

So . . . I've got a kinda funny idea.  Hear me out.

I read a couple pages back that you were having machining issues related to the size of the part you're trying to cut.  Is this something you're still stuck on?  My place of work has access to machines that have some pretty gosh-darn big work envelopes, because we're that rare California tech company that actually makes a physical product.  And while I'm not particularly chummy with the guys on the shop floor, I've been meaning to take a wander down and say hi.  Our production rates keep them pretty busy and I know these machines are run a lot, but . . . perhaps there's some space in the schedule for a personal project?  How big and what is it that you're trying to cut?

I hear you on professional switches being stupid expensive!  Those rectangular Tellite switches from the Apollo mission control consoles are flight-rated and even on e-bay used you're lucky to get them for less than $80/ea.  I can't even fathom what they are new.  But as for emulating them, arcade-style buttons seem to be pretty close: https://www.amazon.com/EG-STARTS-Rectangle-Buttons-Machine/dp/B07R8CNW47

They're not a flush pop-in mount, I doubt they'll feel nearly as nice to press.  I'm also not sure if they're relegendable, but some of the colored ones do look like you should be able to pop the clear cover off and stick a label in there?  I guess I'm about to find out, they should be delivered tomorrow.

If they don't work out, do you have a brand or part number, or perhaps a source where those video console buttons you got out of the scrap board would have been sold new?  $6/switch is a perfectly acceptable price, and I had the same idea that you mentioned with a bi-directional LED.  I'm still having trouble finding a good LED for it though; the ones from the local Frys were hilariously dim.  Something like 12mcd.  Do you have a source for some bi-directional Red/Green and Yellow/Green LEDs that have some actual oomph to them when run at 5v?

I've also been giving serious thought to your suggestion about running multiple controllers.  Initially I had planned to use at least two controllers, but now I'm toying with doing it on a single device by using an Arduino Due. It's powerful enough that there should be no problems with computation and speed, and it can do the native USB keyboard emulation which means that I can add even MORE buttons and controls to my console! :D

The real downside is that the Due is a 3.3v controller, but almost all of the other bits on the console are going to be running on 5v.  So now I've got to figure out how to get 5v devices to talk to a 3v board.  Is there a trick to that?

Link to comment
Share on other sites

I actually work at a machine shop. I have access to a large mill, and even the option to use it, but I also don't have the final dimensions for the tape meter, DSKY, and ∆v carriage meter designs finalized, so I've not cut anything out yet. One thing I may do is cut the space for the edgewise meters, the navball, and the vertical velocity meter. Not too worried about that.

As for the arcade switches, I don't know if those ones you linked specifically are relegendable or not, but I have a pair of similar switches that definitely are. The tactile feel is "clicky". Most arcade style switches typically use industry standard Cherry style momentary switch hardware for the actual switch. Honestly, I absolutely hate Amazon for anything parts... They have very poor representation and search features... To the point that I avoid it. For basic components, I tend to favor Digikey first, then Mouser, then ebay, then aliexpress. I trust ebay to provide a better selection of arcade style switches, and aliexpress to provide the lowest cost (but greatest ship times), since it's a Chinese vendor site. Here is an ebay link for 10 of those switches, with clear covers, for $20. This video shows round buttons, but the rectangular and square versions tend to let you change the insert in exactly the same way.

If you want flush or semi flush buttons, and the switch has significant height over the surface of the panel, then you can use two panels. The top panel would have cutouts sized just big enough for the actual button itself, and the second panel would be mounted below the top panel with spacers. The second panel is where the pushbuttons are actually mounted. The video effects panel switches I have are PC board mount, so you'd need to solder them to a board, then mount that board to your panel with spacers. I just used generic proto board PCB material and drilled the holes for the alignment posts with a small drill. You can see the pictures of how I mounted them on this thread.

For the 5v vs 3.3v issue, look for level shifters. I definitely recommend digikey for finding specific LEDs, due to the parametric search, but if you know what you want, check ebay. I recently picked up 100 3mm UV, 100 5mm White, and 100 5mm slow changing RGB LEDs, Spent only $8-12 per 100 piece lot of them. I got the red/green bi-color LEDs from Digikey years back, and sourced them there, cause I knew exactly what I'd be getting. For indicators, I highly recommend diffusing LEDs. They will have a lower brightness rating than clear lens LEDs, but the diffused lens will more evenly spread the light out over the button. You don't need that much for indication, else it could be a distraction. 10-14 mcd is common for indication. I know the bidirectional LEDs in my DSKY keypad work well. If you need brighter, you might need to see if you can fit a pair of individually brighter LEDs in a back to back configuration, and fit it where the single LED would normally go inside the switch.

Also, if you haven't already done so, I'd definitely suggest starting a build thread for your own controller project. I'd be happy to help out where ever I can.

Edited by richfiles
Link to comment
Share on other sites

Oh boy... I hope all that fresh mod-ability they talked about in the KSP 2 announcement translates well to the KSP serial IO mod needed for this instrument panel project to work!

Very importantly, I hope that the data packets can be maintained between versions. It'd be awfully inconvenient to have to change things to make this work on KSP 1 vs 2.

Link to comment
Share on other sites

  • 1 month later...

I am working on a controller as well and noticed you were looking for a way to print the letters on the buttons.  I am going to use this product to print out the labels and put them on my controller (Which will be a face plate of 1/8" aluminum I am going to mill out).  After printing they can be sealed with a UV sealer for added protection.  They are cold transfers: Print them, put them in lukewarm water, then slide them on.  Basically just like model decals from the old days, just custom.

 

https://www.transferpaperexperts.com/translide-clear-laser-transfer-paper

Hope this helps.

Keep up the good work!

 

Ben

Link to comment
Share on other sites

I actually got some of that stuff, for my model train hobby. It's far simpler though, for the buttons. I have Nekoosa Synthetic Paper. I can just print directly onto it, cut the square out, and insert it into the switch housing. It works great!

Link to comment
Share on other sites

  • 3 weeks later...

qdJeIcb.jpg
These silicone grommets are super cheap on AliExpress. They're used with tattoo guns. I've decided I want to 3D print Apollo style tab levers for the toggle switches, and I've ordered these to put on before epoxying the tab levers to the switch levers. I wanna see if using these will visually mimic the silicone hermetic seal that was on the MIL spec switches used on the Apollo instrument panels. I hope it looks good!

I'll need to find an alternative to Shapeways. I have no idea what is going on over there, but my shopping cart that had $60 worth of Apollo tab lever switch caps was showing $300 yesterday, and today $150! I have no idea what madness is going on over there! Yikes! :0.0:

Anyone know a good alternative to Shapeways?

Edited by richfiles
Link to comment
Share on other sites

  • 1 month later...

Once again, China can't even. Colors, how do they work? Is this a red grommet? :huh:
I'm annoyed, but I can't help but laugh, cause it's such an absurd color swap! :rolleyes:

aAkFlaU.jpg

Well, I knew that I'd eventually find the right type of silicone grommet, and sure enough, I got these in today.

7EkneLh.jpgrfu7Rgt.jpg

This might just work! Slip these grommets on, after cutting them to the desired shape, and I think we might have a winner!
All it should need then is the Apollo style Tab Levers for the ends. Some 3D printing and some epoxy and paint should do the trick!
I had noted how stubby the Apollo switches were, so I decided to get some mini switches to experiment with.
I may still go with full size toggles, particularly if I can find one with a flat lever. I know those won't twist, and might be easier to 3D print Tab Levers to fit onto.

Compare to these actual Apollo switches.

6lgIhYm.jpg

Link to comment
Share on other sites

  • 5 weeks later...
  • 2 weeks later...

I've been wracking my brain on how to sense the position of the tape meter and the carriage meter. I need them to be able to perform fairly rapid motions, to accommodate rapid altitude modification (planned and unplanned) or rapid small burns, while not losing track of their position. One idea was to print a pattern on the backside of the tape, and have photodiodes detect the coding. This would double my work printing the tape, and also is impacted by the fact that there's already numbers on the back side, and any printing that doesn't sit in the center could wear on the rollers.

Ultimately, I saw my boss had left his digital caliper on at work, so I turned it off to preserve the battery. It occurred to me, from my experience with using Chinese manufactured caliper hardware as DROs for low end milling machines, they they encode their position using an absolute position encoding method. Now, there's too much friction in those to get good rapid movement, and I worry that they would wear out prematurely if there were enough slop left in the movement to reduce the friction, but it did inspire a search.

I looked up Absolute Rotary Encoder, and found exactly what I need... At avionics prices... WOAH, YIKES!!! :0.0:
Okay, the navball was expensive enough, and I don't currently have the cash for big purchases like that at the moment... And I need two of them...

I saw a few crusty ones without easy to connect shafts for the $60-100 range but what I really need is a something with a standard shaft, or that can be attached easily to a standard shaft, that I can easily connect by belt to a motor or between the motor and drive wheel for the meters. I did actually find a few that might work out okay, including a good brand name unit that only has 256 steps, a short but usable shaft, and is about $20, but now I'm wondering if another device might work out just as well...

I saw an angular hall effect sensor: the AS5600 Rotary Position Sensor with 12 bits of resolution. It's basically a chip that senses the angle of a magnetic field. It'll output an exact value over I2C. I don't even need 12 bits. 8 bits is double the numeric position (I can code for tape positions equal to half steps, so point at one number, point at the space between two numbers, point at the next number, etc. With 10 bits of resolution, I can quadruple that, and again, it's more than enough for the tape meter to display it's values. Regardless of the version I buy, All I need to do is have the motor drive a toothed pulley attached to the tape meter's input cog, and have a magnet attached to the system. The drive wheel of the tape meter has 20 sprocket pins in one revolution, with the tape being 120 sprocket holes in length, giving me a 6:1 ratio. I can handle that in two possible ways. I can have a 1:6 pulley driving the tape meter. It means the motor shaft would be pretty large, and it reduces my motor's overall torque being transferred to the drive sprocket. The low torque means it can't come up to speed that fast, and might lag. The benefit, Is I can mount the magnet to the motor, and one rotation of the motor would then equal one rotation of the tape, and I'd have a true accurate position feedback with minimal hardware. The second option is to have two pulleys belted onto the motor. It'd be more complicated, but I could even have the motor have an even smaller shaft than the drive sprocket, giving me more torque, but I'd have another pulley, 6 times larger than the sprocket pulley, and that would only exist for the magnetic absolute position measurement. It requires I might need to do extra things, like maybe have an idler wheel, and a serpentine belt configuration. Might be able to get away with a triangular belt configuration too. The meter is still a low torque thing. This configuration might end up being the route I actually take. The last option is to combine the absolute position sensor with some other sensor to determine which rotation it's on.

Regardless, whether I splurge for an optical encoder, or go with a magnetic angular position sensor instead, I finally think I have the tape meter figured out.

The carriage meter will likely be very similar, but I'll steal the head assembly from an electronic typewriter/word processor and drive that mechanism. I can drive the steel cable that moves it easily.I think I'll want 10-12 bits of resolution on that one.

Anyway, mostly brainstorming, and recording thoughts before I go to bed. A lot of issues got solved today, in a very short period of time!

Link to comment
Share on other sites

TMz9jx2.jpg?1

So, did some mockups and measuring and researching parts online. Thanks to cheap CNC and 3D printer parts, toothed belts are dirt cheap online. I've also got meter placement to worry about. Sadly, Space is very constrained in my housing, and adding the Delta V readout takes things from a comfortable fit, to shaving millimeters off bezels to find a working fit. The above image shows the fit around the DSKY.

5zNhiqV.jpg?2dazhhE4.jpg?1

I wanted the keypad to be centered between the annunciator LEDs and the displays, but if I go this route, I'll have to settle on centering the keyboard on the displays, due to the space taken by the pulley. You can see above, that the space is so tight, I'll even need to trim the edge of the PC board off, just to make room for the sprocket. I may or may not be able to shave a small amount of distance, depending on if I can shave out the inner diameter of the pull wheel to it fits over the bearing housing (the bit that sticks out behind the gear... That gear will be removed, and replaced with a toothed pulley wheel)

I measured and cut out some mockups for the pulley wheels. The large wheel has 60 teeth, and the two small ones are 10 teeth, giving me a 6:1 ratio between the motor+tape sprocket, and the large wheel. The large wheel will have one of those Hall Effect rotational position sensors reading a small magnet mounted into the wheel. Every six turns of the motor generates 6 turns of the tape sprocket, and one full loop of the tape, and one full turn of the large wheel. All I'll need to do is have a basic closed loop position based motor driver program. That's something I might even be able to just find an example online for, and throw an an Arduino Pro Mini or something. You just send a desired position to the microcontroller, and the program looks at the position sensed by the position sensor, and determines how much error there is, and runs the motor at a proportional speed to to eliminate the error. I can put in a max speed limit and max acceleration/deceleration rate to protect the tape from being torn by agressive motor movements. The belt will be in a simple triangle configuration. The motor and the shaft+bearing that mounts the large wheel will just be attached to the plate that will screw into the main tape meter housing. If I slot one of the screw holes, I can adjust the tension out of the belt by sliding that assembly forward or back. It'll function the same way the tape tension slot works (visible at the top of the tape meter housing). Adjust the tension, and tighten the screw. That's it! I'll look through my collection of bearings and see if I have something that I can mount the large wheel to easily.

This all assumes I even mount it to the left of the DSKY. Left of the DSKY offers me one major advantage. There's literally nothing behind the annunciator LEDs, leaving me plenty of room for the mechanical elements, including the large pulley wheel. There are two other positions I could mount it, but there are issues regarding space for hardware in both of those positions.

37GMyJl.jpg?1nERPuXa.jpg?2

I could optionally mount it right of the DSKY, and left of the vertical velocity meter. Mounting it here keeps the radar altimeter, altimeter, and vertical velocity all in a collected cluster of instruments, which I like, but I'll get to the issue that mounting it on either side of the DSKY results in, in a moment. On this side, the bulk of the Vertical Velocity meter packaging creates an obstruction, making the triangle belt configuration impossible. A simple solution is to add a fourth pulley, a non-toothed idler wheel that holds the belt below the edge of the VVM housing. It's a very simple solution, and won't hardly add any cost, and it should work just fine.

Honestly, it might be easier to mount the LED displays below the VVM, instead of above. It gives me a little extra clearance, but still not enough to eliminate the idler wheel. I unfortunately do belive that space limitations dictate that I may need to buy smaller LED displays for this as well. That wasn't part of my original plan, but I wanted room for some pushbuttons, and some LEDs by those displays, and I think I just have no choice but to go smaller. That's alright though. It'll keep it from looking like the DSKY puked a few extra displays onto the side, and make that bit look like it's own module in the panel.

mvPyeIe.jpg

Finally, there's the third possible mounting location, left of the Navball. I can't say I really want to mount it here, because it makes for the hardest mechanism to build, as well as pulls it's information away from the altitude cluster of instruments, but I guess, photo included for reference purposes... Mounting it left of the Navball has some disadvantages, but it does one thing that's very lacking in this panel... It helps bring the Navball closer to center. Over time, I've eliminated three of the seven edgewise analog resource meters (in favor of four dual readout meters). I've also since come up with the idea for the Delta V carriage meter and found the tape meter to make into the Radar Altimeter. The LED display failures of the original smaller DSKY digits resulted in me buying larger LED displays, which also further widened the DSKY. That took out three meters to the left, and added two to the right, plus the wider DSKY. This has had the unfortunate consequence of almost eliminating space for any large bezels, and shifted the navball rather far from the central position, and honestly, I'm not really a fan of that. The Navball was supposed to be the star of this build! Anyway, I mocked up the positioning of mounting the tape meter to the left of the Navball, main;y just to see how it looks.

The Navball is very long (it's my longest instrument, depth wise), and completely obscures the space beside the tape meter for any mechanics. If I did choose this method (I kinda don't really want to though), I'd have to use a double shafted motor with two pulleys mounted on each shaft. A short belt would sit on a pair of 10 tooth pulleys mounted on the motor and tape meter. The other side of the motor would have another 10 tooth pulley on the other shaft, and would then connect with a second belt to the 60 tooth pulley for the position sensor. All that hardware would be positioned directly behind the tape meter. It's a bit more complicated, and I don't much care for the positioning though, so I likely won't do this.

Jcz9H9I.jpg

That leaves the Delta V meter. I REALLY want it to sit on the right side of the navball, just like the ∆v meter built into the game... In fact, it's being built especially to model itself after the one in the game. I laid out all the instruments on the front plate to get an idea of the space I had available. I installed the bezels onto the two outer edgewise meters, and lined them up to the mark I made indicating the flange on the inside of the main housing. Since the Navball has an integrated bezel as well, I just left a very small gap between the meters and Navball. On the right side, I again aligned the Vertical Velocity Meter to come right up to the mark I made indicating where the main housing flange sits. Placing the Radar Altimeter tape meter and he DSKY in a reasonable position, I'm left with the small gap between the Navball and the DSKY, which very nearly ends up being centered. Ideally, I'd love to have two of the smaller display side by side. I need room for an LED illuminated scale on the side those displays in the center will have an LED masked to look like a pointer, and the entire assembly will be able to move up and down, just like how the ∆v burn meter on screen, next to the Navball works. When set, the meter will physically move the displays to the top of the scale. When you perform a burn, the meter will move toward the bottom, indicating the percentage of the burn completed. The displays will display the ∆v and the burn time. There may be some additional displays at the bottom to indicate more details, if needed. The big gimmick, of course, is the physical movement of the digital display and pointer LED. The mechanism will be different than the one for the tape meter, but reuse a lot of the same ideas. I'll snag the linear bearings out of a printer or typewriter, for the display assembly to move up and down on. A pair of pulleys will be mounted above and below the assembly, and a cable attached to the assembly will be spooled around a drum attached to a motor. This is literally just the guts from an electronic typewriter I have, reused. I'll have to figure out how much rotation of the motor translates into how much distance is moved by the display assembly. If I instal a large enough drum, then the circumference itself could simply equate to the distance traveled, and again, I'd only need to mount a magnet to the drum, and read it's angle with a Hall Effect Rotary Position Sensor, same as the tape meter. I honestly don't know yet just how wide the carriage meter will end up being. I'm hoping it is narrower, simply so I have some space for bezels.

sbAs10e.jpg

One final note, I need to rework the current dropper resistors for the diode ROMs. Pic is a bit blurry, but if you look at the second row of pins, from the edge, you can see they've all snapped off. I took a gamble and tried to use ceramic surface mount resistors, mounted on end with the sockets for the ROMs mounted on the end of those. I was very concerned that they could crack, given how brittle ceramic resistors are, and sure enough, it did. Why did I use ceramics, if I knew they had a risk of cracking? Simple... I had 'em on hand, and I wanted to see it light up when I built it. The solution is simple though. I just need to order the through hole version of the resistor, so I have sturdy leads to solder to. It'll probably take an hour or two to rework. It's more a waste of time than anything. There's about 30 resistors per display pair/diode ROM set, plus several more for the "Time To:" ROM, so just about a hundred to replace. Cost won't be an issue. 250 pieces will only cost me $3.84, and will be enough to do all the LEDs.

Edited by richfiles
Link to comment
Share on other sites

  • 4 months later...

I just had an absolutely hilarious idea... I kept a brushless motor that was part of an early prototyping/evaluation run for an aviation manufacturer, back when i worked my old job at an old employer, back in 2006ish. That motor was intended to adjust the control surfaces of a refueling boom on a fuel tanker jet. The company wanted to try a direct control system, as opposed to the traditional hydraulic controls. I kinda want to see how low a voltage I can actually operate it at... It might need too much voltage to even spin, in which case, It won't be suitable, and it's kind of a dumb idea, cause the motor is SO overkill for the application, but I kinda wanna see if I could make use of it for my tape meter. No reason other than that I already have it, and it just so happened to use a resolver for commutation, which happens to be a type of absolute position sensor, something I need anyway. That one early production motor would have sold for about $5K if it's leadwires had been just an inch longer, and would have directly driven control surfaces... LOL... And I'm thinking of using it merely to spin a tape meter! :rolleyes::confused:

Honestly, I had other motors that were much smaller, and much more suited, but i don't know if I kept any... Many of them went to a scrapper when I moved back in 2012, and none of those had resolvers. Those motors went into other aerospace applications, but definitely aren't nearly so massive... I once considered building a scooter with that refueler boom control surface motor! I never did, cause it would have been a lot of effort, and I needed high voltage to get good torque. If I can get just enough torque at low voltage to spin the tape though... Maybe...

Regardless, I just dropped a new order for parts. My last order appears to have never even shipped out at all. Very annoying, but given the ongoing craziness going on in the world, I can kinda understand that some places might not be able to fulfill orders. It's just excruciating to wait for an entire month before find out nothing's gonna show, cause it never got on the slow boat. Each failed order is another month of delay... :(

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...