Kerbal Instrument Panel: In-Desk Apollo Themed Hardware Controller

[cubinator]

Great work! That's an awesome looking meter!

[richfiles]

Thanks! It was a lot of work, and I took extra time to do things exactly a certain way, but in the end, that was because I wanted them to look this good! 

Once I finish the meters, I will have almost the entire left half of the instruments ready for panel assembly. I really need to figure out how I'm gonna do my rate meters/velocity readout bezel that will fit around the Navball. That'd be the only thing left up in the air. Man, the Russian LEDs that I was looking at look GREAT, but dang, they are expensive! It'd cost me nearly $100 to have enough shipped for the project. I feel like it'd be WAY cheaper to figure out a small SMD PC board with SMD LEDs soldered to the edge (basically, anode on top of the board, cathode on bottom of the board, with pads right up to the edge. I can get SMD LEDs at bulk prices... I feel like that would likely be cheaper, and probably look fine, IF I can find the right shape SMD LED.

After that, I just need to create the scale for my vertical velocity meter, create a bezel for the DSKY, create a bezel for the Auxiliary Digital Readout, and build the ∆V Carriage Meter, and the instruments will be complete! Once instruments are complete, I can move onto the control panel and start that process.

[richfiles]

Youtube Video
Direct Video Link: https://i.imgur.com/fmlSXS0.mp4

And with that, my first meter is 100% complete!

I installed the high and low movement stops, painted the edges of the meter face white, to help reduce light spill even further, and obviously, reinstalled the meter movement mechanisms.

Put the covers back on, reinstalled the face and the top and bottom bezels, and with that, it's done! I probably could have done this the other day, but the stops require very fine tweaking, and the pointer is incredibly fragile... I didn't even wanna look at it tired.

Im pretty happy with the amount of light spill that's been minimized. Most important, is there are no glaring pinpoints of light, and the gaps where the pointers pass through are nice and dark!

The finished shot!

Edited January 6, 2021 by richfiles

[richfiles]

Not much to update today. Worked on the remaining meters. I'm doing some steps in batches, to streamline. I will be doing the single edgewise meter separately from the remaining three dual edgewise meters, as both it's internal construction, backlight mod configuration, and even the scale itself has been done entirely differently. I also, unfortunately, discovered I may need to reprint that scale. I made 2 minor errors. The movement configuration text that I copied off off of the original meter faces is placed too low to be visible, and the color gradient for Atmospheric Density covers the entirety of the face width... The only things that would be illuminated in the aqua color would be the text and scale tick marks, and two small vertical strips at either end of the color gradient. I'm not sure where I went wrong with the measurements, but I also realize I never test fitted the early paper printout, which is sad, as I could have caught this early on. Now I've wasted a print opportunity with the Nekoosa Synapse XM synthetic paper, and I only have a few sheets of the stuff. I don't wanna re-print until I can fill an entire sheet. I'll probably assemble it with just a paper scale, and get the mod built up, and then re-label it with the Nekoosa synthetic paper at some point in the future, probably when I start printing nomenclature tags for the control surface. It's a small bummer, but it is what it is...

As for progress today, I have the proto boards installed on all three dual meters, and started the tedious preparation of the front face of the second of the factory backlit meters. The one I finished had that step done ages ago, and the two non backlit models faces were just straight up cut off and replaced with smooth polycarbonate, which has no face features to even be removed. The one meter has a center ridge that is higher than the rest of the face. I have to slowly snip that off, then file the ever so slightly tapered front face into a flat (side to side), even curve. If I don't do that, the scale will have unsightly lumps and warping in the surface, as it's stretched tight over the curve. No pics, cause it's the same old stuff that's already been covered in past posts, and I'm now focused on getting all the steps done as fast as possible, as opposed to re-documenting every step. Today, I will be busy with other things all day, so I might not get back to this till Friday. I'll probably finish that front face modification, drill and tap holes for securing the scales and the two meter's replacement face plates, and if I have time, solder all the remaining LED modules. I likely won't do more than that, as my Friday afternoon is also occupied. I'll have the weekend free, so I'll see how far I can get then.

Edited January 27, 2021 by richfiles
I'm alive... Just had work I had to do first.

[mountaineerman]

Top-notch, professional job. Way to go!

[richfiles]

Decided to Work on the Atmospheric Density Gauge, along with the other three gauges. The other three are partially completed, still have some work left to do, but they're moving along nicely.

Construction of this one was markedly different than the other four. This one had a gradient scale and two entirely separate illumination zones. I painted the edge of the polycarbonate first, and fed the printed scale into place and secured that. I built up another LED board, similar to the other meters. Again, each LED has it's own resistor (on the other side of the board) for even current limiting and thus even illumination. I tweaked the angle of the LEDs to center and evenly space them as best as possible. You can also see the metal shield I cut to separate the two illumination zones.

I cut a diffuser to spread the light evenly over the back of the scale, and made sure to cut it's shape to match the contour of the bent metal shield.

Here, you can see the diffuser generally placed. This style meter has no provision for factory backlighting, unlike the dual meters I was working with. As such, there is no integral mounting for a diffuser, so I had to come up with my own mount. The solution I came up with was actually really simple.

Drilled out and tapped a hole in the housing at each end of the diffuser, punched a hole in the end of the diffuser, and screwed it down. It was really simple, but works.

With the metal shield in place, the top illumination zone is now separated.

The gradient scale LED board has three colors of LED, Ice Blue, Blue, and Violet, to really accentuate the gradient.

Some 1/4 inch nylon spacers were used to line up the second LED board so it sits at just the right height.

Just like before, a diffuser was cut and lined up in front of the LEDs, and screwed onto the sides of the housing to secure it.
I also cut and bent a spill shield to keep light from spilling through the gap for the meter's pointer.

Almost fully assembled. I forgot to take a picture of the pointer movement mechanism. I misplaced the screws, so had to get replacements from the hardware store.

The separation between the lighting zones work well! Really liking how it turned out!

It looks great in the dark! Here is the Atmospheric Density Gauge completed! I really love the color transition of the gradient scale, and the camera simply does not do it justice.

What remains is mounting two of the final three scales on the last three meters, cutting the grooves in the polycarbonate to accommodate the color bar LED assemblies, and epoxy those in, and solder them to the LED boards. I've cut the light spill shields, and already drilled and tapped most of the holes needed. This was the one unique meter, and I think the more complex meter to build, so I suspect the remainder should go quickly.

Edited February 13, 2021 by richfiles

[richfiles]

Just got the correct screws and finished mounting the meter movement. Figured I could snap that shot I forgot to take.

Just a small touch that I personally prefer... A lot of my equipment is salvage, surplus, or pre-owned. In the case of this meter, it used to be in use with a utility. I had to remove that label to make way for the backlight screw terminals, but rather than discard it, I trimmed it in such a way as to allow it to fit back on the back. It's a detail only I would know about, but I kinda like it. 

Edited February 14, 2021 by richfiles

[richfiles]

Progress! ™

I made one minor error securing the scales to the housings, and I'm sure any keen eye will spot it's glaring obviousness. I will live... It's taken me far to long to get to this stage, and though it is a clearly visible error, I shall live with it, in the name of moving forward with the project and extreme laziness in not wanting to rework it. The misalignment is only about a millimeter-ish... For those not interested in not playing a round of Where's Waldo Kerman, and *somehow* can't spot the error...

Spoiler

I misaligned the top of the Solid Propellant/Xenon fuel gauge (second meter).

Apparently, I was also delusional in my belief that these would be finished today. My decision to individually current limit every individual LED with it's own individual resistor added a lot of individual moments of extra soldering time to this individual Kerbal fanatic's schedule! It took a long time to make all the color bar LED assemblies. Took a good while to mount the scales to the fronts of the meters too. That pile of resistors is the pile that goes onto the three circuit boards inside the meter, where I'll mount the Blue-Green LEDs. Both resistors and LEDs will need to be mounted, and the color bars epoxied into grooves cut into the back of the front polycarbonate. I can imagine all that taking only about a day, so see y'all next week! 

Not gonna lie... I'm really looking forward to seeing this cluster of meters completed, with all trim installed, including the side bezels. This will also mark the start of attempting to mount components into place in the aluminum panel. I will be jamming those meters as far to the left as I can make them fit. I can work out how I want to mount the tape meter. I'll have to focus on figuring out how to construct the bezel that will wrap around three sides of the navball. That bezel will contain the velocity digital readout, and three LED bar graphs to represent Rate Gauges. Once I'm at that point, I'll have constructed half the instrument panel! 

[Misguided Kerbal]

8 hours ago, richfiles said:

The misalignment is only about a millimeter-ish

Man, I hate when that happens. Is it possible to remove and reattach it?

[richfiles]

14 hours ago, Misguided Kerbal said:

Is it possible to remove and reattach it?

It would have been if I had noticed it before trimming the sides down. Notice how the black portion extends to the sides of the meter, but the white portion is narrow? That's because it gets trimmed down to the width of the polycarbonate section I'm reducing the length of the gap for the pointer to travel along, to get those wide top and bottom black spaces, like the Apollo meters had. I just didn't see it when it was wider, but once the gap got trimmed down, it became more obvious that there was a misalignment... Eh, it's not a vertical alignment, and I expect a small amount of bloom from the light shining through it, so I'll live with it. It'll simply be indicative of fine Kerbal quality construction. That'll be the meter sourced from... checks notes... "Found lying by the side of the road"... Okay...

Small update, not worth a new post, I think...

I got ALL the dremel work on ALL the remaining meters done. Visually, it don't look like much, but this means I can proceed to finish with nothing more than epoxy, soldering, and basic assembly. Literally nothing is holding back the finishing of these meters now, especially no "I can't use the Dremel this late, cause I don't wanna disturb the neighbors" spiel! 

Edited February 16, 2021 by richfiles

[richfiles]

The five edgewise analog meters are ALL completed! I discovered I don't have enough cosmetic trim pieces to cover the calibration screws, but I can probably just order a few from the manufacturer. Worst case, I make a trim piece that covers all four of the vertically oriented meters. The Atmosphere meter has molded trim, so It can't be missing! 

Total current draw of all 5 meters at max brightness is around 670mA. Internally, there are 52 individual LEDs, and 25 rectangular LEDs, with a combined 66 individual LEDs across all the 25 modules (each module has either 2 or 4 LEDs, with their own individual connections. Means I soldered 118 individual LEDs and just as many resistors, between all 5 meters. I'd rather have more evenly spread light, with each LED being a little more dimly lit, than to have only a couple LEDs, but with hotspots. Each LED only average 5.6mA.

And here's a shot with my workbench light off. The color bars really pop! Honestly, I did the temp the way I did it, just to break up the monotony of the other meters. I've said it before, but the same is true of the atmospheric density meter. I chose to use SI unit prefixes for units of one Atmosphere at Sea Level (ASL), rather than doing a percentage That was purely to break up the monotony of 0-100% ranges.

Not actually a lot to post here... You've all already seen what goes into making them, with the first build posts. This is just the conclusion to a very long and drawn out portion of the build, and I am SO GLAD to have finally finished this stage! Next step is to start cutting openings into the front panel for the different instruments. I'll finally be able to get a good idea of how much room I'll really have for the Δv meter.

Edited February 27, 2021 by richfiles

[richfiles]

Apollo did have a number of round meters, and had a relatively unified style for them. Most of my meters are edgewise meters, reporting on resources or temp, but there is one meter that stands apart. The Vertical Velocity Meter in KSP stands out as an ideal use for a round style meter, and I was able to snag a very nice meter a while back. The Vertical Velocity Meter is the round one, as follows:

Anyway, i removed the meter's scale plate and scanned it at 600dpi, the same setting that my printer prints at, so i would have a 1:1 ratio image to work with. I took measurements, made modifications, and slowly but surely drew a usable meter over the old scale, and made sure to mimic the Apollo meters as best as possible. Then I realized my multiple mistakes and started almost completely over. D'oh! One thing I had considered was cutting a small segment out of the metal piece that supports the scale plate. It would permit the backlight to shine closer to the center, and allow me to increase the size of "Vertical Velocity", but with adjusting the spacing between letters, and pushing it outward a bit, I think I've got it.

My current scale revision is shown below. I think I have everything figured out, as far as backlighting goes. The entire internal structure is mounted on a sheet of blank FR4 PC board material, so what I'll do is just drill a few holes in it, mount LEDs and resistors, and design it to shine onto the back of the scale. For the logarithmic scale, I used an old slide rule, measured out the ticks with a regular ruler, and then converted the ratio to something that looked good on screen. It's not perfect, but it's close enough to actually convey the shift in orders of magnitude. I really had to tweak with the text size and space between characters to make some of it fit. Originally, both LEDs were gonna light together as a warning light to indicate decent was too fast for landing, but adding another terminal to the rear and cutting a wire trace would permit the two to light independently. I considered duplicating the Contact Light that will be on my DSKY here as well... Honestly, I just had a blue 8-10mm LED sitting around (I forget the exact size, but it matches the RED), and on the Lunar lander, the Contact Light lit up blue. It was the only instrument that lit up in that color, and was pretty prominent. I kinda wanna mimic it.

As a side note... back to that first image... This is the first time I looked at these particular photos up close. I had seen toggle switches that seemed to have some weird reflective anomaly, but I I never realized it was because the tab levers were clear plastic. I've always seen the metal toggles, and assumed the shimmering ones were maybe wear polished... Nope! Just another thing I've learned about the Apollo program. I suppose all those toggles on the Lunar Module are plastic for the weight savings. I haven't dug through enough pics to determine if every LM used these, or only later ones.   

 

[richfiles]

I ain't dead, just got a lot of work, plus some side work.
It's all been stacked up, and that meter still is staring me in the face! 
One of these days, I'll stare back hard enough that it's own face changes! 

[richfiles]

Sometimes just thinking about something is enough reason to pick it up. Truth be told, I can't remember the last time I've been this busy with work, but sometimes you just need to take a day. I have time to finish before the deadline, and know I will meet it, so I took a few hours to work on this. First step was to drill some holes to mount the LEDs into. Very conveniently, the meter movement mounts to a fiberglass board, which also serves as a perfect base for the backlight.

I put sockets where the Contact Light and Rate Warn LEDs go. I need those to rise up off the board, so I can adjust their desired height by soldering some leads to them and trimming them to the appropriate length.

The rest of the backlight LEDs are all soldered in point to point. I marked off locations where screws go, so I would remember to keep those spots clear.

The Backlight LEDs all light up nicely!

And here's with the meter movement assembly mounted. Note that I did end up trimming the frame a bit, so that I could accommodate larger text in that location.

The Contact Light LED and Rate Warn LED are socketed. I am definitely going with this particular blue LED. It's such a soft, mellow blue, not like the harsh, shrill, eyeball stabbing blue LEDs that are so typical. I have no idea what makes this blue special, but I love it, and I'm sticking with it. I may try to find a larger diameter red LED to match it, since one is a 10mm and the other an 8mm LED.

Seriously though... That blue is so beautiful! I love it! I have not reached the point yet, but there will be a tube (likely just black heat shrink) around the red and blue LEDs. This will prevent light spill from the backlighting, or spoiling the colors. Another option I could take, since I am satisfied with the size, color and brightness of the red LED, is to simply think heat shrink tubing slightly past the curvature of the blue LED, and make the opening beveled, so only 8mm of the blue LED is seen... Yeah, I think that makes the most sense. It's simple, and entirely doable with zero effort, and I like zero effort! 

Okay, I lied... I didn't mount the meter before, I was just test fitting it. I attached the LED wiring to the rear plate first. I had to drill out extra holes to accommodate the additional lighting circuits. Initially, it only had a pair of incandescent lamps that illuminated the front of a metal plate. This was just two terminals. Now I have backlighting, Rate Warn, and Contact Light, along with a ground for all three. I used some scraps of the copper board I used in making the light shields for my edgewise meters, and made terminals, because I could not find my terminals. As soon as it was fully assembled, I found my terminals. Oh well. 

Here, the meter movement assembly is properly mounted and is now wired in to the rear plate. The black electrical tape holding the red wire down is only temporary, I glued the red wire down, since it wanted to stick up.

Here, we can see the assembled internals, as well as the top diffuser plate. I still need to drill the LED house for the Contact Light and Rate Warn LEDs. I will fill in my errors with some plastic cement, and hope that they disappear, more or less, and that it reinforces the part... I really don't want to cut out another one... I still need to mount an intermediary diffuser between he LEDs and the diffuser plate. Unfortunately, the LEDs don't have much distance for their light to spread evenly, so I went with extra LEDs and will add more diffuser layers.

One aspect I had considered, was fanning the light out at an angle, so there is artificially more spread to the projected beam. When that hits the diffuser, it'll already be slightly wider, and be less of a hotspot. Early tests show this should work quite well.

I almost don't want to post this final image, just cause it looks so bad, at least in comparison to what the final product will look like, but this is where I ended:

Part of the reason this looks so very bad, is because this is just a paper label, and the intermediate diffusers are not installed. There's a ton of light spill from the edges, hotspots, and it's overall WAY too bright, due to it being missing the extra diffusers. There is also no cover plate for the meter's center, and the Indicator LEDs are not installed.

Yet to come, installing the intermediary diffusers, drilling the indicator LED holes in the main diffuser plate and installing those LEDs, printing the GOOD looking scale on the Nekoosa Synapse XM synthetic polyester based paper, and figuring out how I am securing the scale to the meter. That last bit is entirely up in the air. I've considered gluing it to the diffuser plate. I've also considered leaving "tabs" around the perimeter, and drilling a bunch of holes on the edge of the fiberglass plate, and strapping it down like a drum membrane using nylon floss. Haven't decided yet. Glue is easy, but I don't know if it'll hold forever. That Nekoosa synthetic paper is definitely strong enough to be held taught by tabs... It's literally how I did the previous batch of meters. I stretched out tabs till they were tight and screwed down on 'em.

Anyway, that'll have to wait. I'm at a nice, satisfying point right now. I can see the light (ahem) at the end of the tunnel, and I am brainstorming how I'll do the last few bits of work on this step of the project. In the meantime, I do need to get back to work on work. I'm probably gonna be hitting that hard till at least the start of July. This, however, is some hope that I'll have breaks during that span to work on this! 

Edited May 21, 2021 by richfiles
I stared back hard enough!

[richfiles]

Some minor thoughts on mounting the scales and the top diffuser plate...
Turns out the diffuser plate is just thick enough that with the scale sheet in place, the pointer just skims on the surface...
This could be an issue, but I have a few options though...

A: I cut the center out of the plate so it slips around the meter movement assembly. I then print the scale sheet, and cut out a square shape, and put holes where the four cover plate screws go. I cut a second sheet of diffuser plastic with a hole like the diffuser plate, to fit around the meter movement assembly. It too would be cut to match the square shape and the holes that line up with the cover plate mounting screws. I would sandwich the rigid diffuser plate between the lower diffuser sheet and the scale sheet, and the scale sheet would then screw to the two front mounts of the meter movement. The two sheets secured tightly from the four corner screw posts should tightly sandwich the rigid diffuser plate between, and screwing the scale down to the meter movement assembly makes sure it is tight at the center too, and flush with he top of the assembly.

B: The second method will likely still have me mount the scale sheet in the same manner, but would skip an extra diffuser layer sandwiching the diffuser plate. Instead, I would make a new diffuser plate altogether, this time mounting it to my mill (lathe would be better, but mine is not set up, and I can cheat with the mill). What I would do is mill out half the thickness of the diffuser plate, just at the center where it mounts over the meter movement assembly. By halving the thickness there, I make enough clearance for the plate to mount solidly and securely to the meter movement, without it being thick enough to interfere with the pointer's movement.

Option A is the easier method, but it puts 4 diffuser layers between the LEDs and the viewer's eye. Not sure if I want that much light reduction. Each layer softens hotspots and spreads the light, but also dims it by a little bit too. Option B takes more work, and is slightly inconvenient because my lathe isn't currently set up, and my mill needs the space around it cleaned up a bit, plus it requires me to do some wonky tool mounting to do properly. As for the diffuser plate itself, I got *plenty* of material, so I can even afford to botch a few till I get it perfect. That's no big deal. All the diffuser sheets and plates are salvaged from my Grandmother's old TV... Sadly, it was being stored in a trailer parked beside my father's barn, and a sheet of snow "avalanched" off the roof last winter. The impact of all that snow against the trailer knocked the TV over and shattered the LCD. Unfortunate too, as it was a really nice TV. At least I'm reusing it's parts for a good purpose. I'll likely use more of the diffuser plate as the backing for the horizontal control surface as well. It'll make something nice and solid to support the illuminated legends for the various controls. That's something I'll get to when I finish the instruments and begin that stage of construction.

[richfiles]

Went in a slightly different direction today...
This was my first real day off in nearly 2 months...
Got my work done, got my sleep in, and yesterday morning, I actually woke up early. 

I still question why they didn't just make the digit the full height of the module, and stuff that tiny little chip inside the bottom half of the "8" instead of having it sit below... 

Anyway, I wanted to test the HDSP-960 smart LED displays that I had bought all the way back in May. First off, I wanted to verify they all actually worked, which they did! I also really wanted to give the seller positive feedback before the response window closed. These are some pretty specialized and uncommon parts, generally only seen in very high end gear, or in military or aerospace applications. I wanted to test before giving feedback, but if good, I absolutely wanted to give the best feedback for making these available. The cost of these parts was rather high... About $10 a digit. That really adds up when you have an 8 digit display, but the visual of the HDSP-960's 4x7 element matrix just looks too good for the role of a singular readout. I absolutely love the retro look and feel of the display. Pretty much all modern dot matrix LED displays use a 5x7 matrix. 

Part of the delay in testing stems from the fact that, while I have breadboards, over the years, my wire supply got used up. I ended up ordering an 840 pack of precut and formed breadboarding wires. They worked quite nicely, although I'm starting to realize this old breadboard (I got it as teen, back in the 1990s) is showing it's age... Many holes contacts are warped and bent, and it's often very difficult to insert parts or wires. Might be time to put ol' faithful up on a shelf and buy a nice new reliable breadboard. Easier to reminisce about old memories, than try to insert some new old stock memories into those pesky holes! 

Anyway, I did eventually get a test circuit wired up, and I was able to cycle through all the binary inputs... At first, I thought I had things wired wrong... I was fine... I had just misread the datasheet, and it turns out the binary inputs are active high, but the decimal point is active low... Go figure. I had the binary wired so a switch turned on would output a 0, which is why when it displays all 0s, the relevant switches are all on. All binary lines high is a blank state, which explains the lack of digit displayed. I also could not get the latch data line to work... Normally, you set the data on the binary input, then pulse the latch data line, and it updates the display to reflect and hold the new data. Once I got it to finally light up, it would update the display as soon as I changed data, and not when I pulsed the latch data input. That one was entirely my fault... I was initially powering this with a 5 volt, 400mA power supply... NOPE! These are toasty little power gobblers. It's easy to forget that these are NOS... New Old Stock... These particular ones were made in 1992. It IS a vintage display, after all. These were first manufactured in 1972! I find these displays to be nice and bright, and very readable. I guess, to get an LED to be that bright back then, you had to use a bit more current than we do with modern LEDs. Turns out, I had to switch to my bigger bench power supply. These suckers draw a combined 1.16 AMPS when displaying the "all pixels on" test character! That's 201mA per digit! When displaying a more normal "all eights", the current draw is a combined 824mA, or 103mA per digit. Obviously, some of the heat is being generated by losses in the on chip LED drivers... I notice with all pixels on, the current per pixel calculates to 9.57mA, while with only 15 pixels on (the number required to display an eight with a decimal point), the current draw calculates to only 6.86mA per pixel. This tells me that between 15 and 21 LED pixels, there are an additional 2.71mA of losses, per pixel. Don't know what vintage tech those "smart" driver chips are using, but that's an extra 57-58mA being burned by the chip, just between having 21 vs 15 pixels lit.

Fun fact! This is not a video! It is a sequence of still photos, taken by carefully holding the phone in the same place, taking a shot, and then setting the DIP switches for the next shot, all strung together as a GIF, specifically pronounced the correct way. 

Actually, digging even deeper, with my DMM, it seems that the LED pixels might not be drawing nearly that much power... The chip on the other hand... WOW! 378mA is drawn from all 8 digits, even when not a single LED pixel is lit! That's 47-48mA quiescent current, per digit! Turning the decimal on draws an attentional 55-60mA, about 7.5mA per digit, regardless of what character is displayed, even if no other character is displayed. The negative dash symbol (b1011 / d13 / hD) is the lowest pixel count, non single pixel character (4 LED elements), and it draws a combined 544mA, or 68mA per digit. Displaying a 1, which lights 7 LED elements results in 667mA combined, or about 83-84mA per digit.

It's abundantly clear that the driver chip is burning a lot of the extra energy. These modules are made from glass and ceramic, and actually get pretty hot to the touch. They got up to 157°F / 69°C after letting them run few minutes. Nice and toasty... Since the bottom half of each module is ceramic, what I may end up doing, is to use an aluminum block behind them as a heatsink, and add a bit of thermal paste, to help draw out that heat, through the block, and into the aluminum front panel. I can have header sockets above and below the aluminum block, and can wire and secure those as required. I only need 5V+ and Ground, 4 binary data lines, a decimal line, and 8 latch lines, one for each digit. I can likely just hand wire the header sockets and then epoxy them onto the sides of the aluminum block. All I gotta do then is apply thermal paste to the block, and plug in the displays. I want to find a bar magnifier lens to give some extra height to the digits, and that can rest directly on top of the displays. I can easily find clear ones, and I can slip a green gel sheet between the displays and the magnifier, to provide them with green filtering. So long as my bezel ends up fitting perfectly, it should apply just the right holding force against that lens and the displays, so they can never loosen or free themselves from the aluminum block. That ought to help keep them nice and cool, and the bar lens will give it that nostalgic 1970s LED display look!

LEFT: Remember bubble LEDs... I certainly do... I could try to find a bubble lens for my displays, but getting a bubble lens sized and spaced just right for the HDSP-0960s would be near impossible!
RIGHT: The alternative, is me feeling that cylindrical bar lens vibe for this project... I feel this vibe, because It's actually not unobtainium! 

Edited July 5, 2021 by richfiles

[richfiles]

I don't get it!!! 

I have double, triple checked every wire, and I am having issues with his assembly! The craziest part, is if I preset the value, then unplug and replug a display module into the socket... IT LOADS THE CHARACTER!!! Literally "have you tried turning it off, then on again?"!!! I'm noticing that the latch buttons and the pull up resistors I'm using are all resulting in an unusually high voltage for ground (1.41v), and that is definitely not normal... Yet these are the SAME resistors I used yesterday, and the SAME resistors used on the SAME breadboard to pull up the data bus signals as well, and those pull down to <1v when low... I'm actually starting to suspect the breadboard itself might be defective... It's the same breadboard I used as a teen, back in the 1990s, and it's warped and old, and the portion I used for the latch pushbuttons is a different area than what I used before... Interestingly, the black module loaded characters just fine. Anyway, I either need to retire this old breadboard, or build a buffer circuit to make sure my low signals are 0 volts and my high signals are at 5v, and that there's no ambiguity at all in the signal integrity. It's just frustrating to have issues like this, and know 95% it's probably just worn out tools, and have that lingering 5% wonder if I'll end up finding an error this time if I just do a quadruple check... 

This is probably one area where building from mental plans probably doesn't help... If there are any errors in the construction, actually having it laid out of paper would certainly help with troubleshooting... A design like this though, I almost never put to paper though... It's just frustrating when things don't work... This doesn't actually happen to me often, so it's easy for roadblocks like this to feel really intense...

Anyway, the way this is configured, there are 7 data bus lines (the number only digits only use 5 of the 7 bits). Each of the eight number digits has it's own unique Latch Enable signal (this is the signal I'm having issues with on my test setup right now). Furthermore, the black alphanumeric module has 2 address inputs to select which of the four digits to update, and it has a Write Enable to write the new digit to the display. This comes out to a total of 18 digital signals to control this... And of course both the Arduino Uno and Pro Mini have only 14 Digital I/O... D'oh! 

This leaves we with a minor dilemma... Do I upgrade to a bigger variant of the Arduino to control this, or slap together some jellybean logic to reduce my pin count. Either way requires that I order parts, and there's no real benefit to one method vs the other. An Arduino Mega might be necessary to have up to 18 Digital outs, and still have room for serial, but that really feels like overkill! In that sense, I feel like some jellybean logic would make more sense:

If I feed 3 Digital I/O to the inputs of a 74HC138 3-to-8 decoder, I can decode three address lines into 8 individual one of 8 outputs. If I feed those eight outputs to a bus buffer chip with one Digital out feeding the Enable, then I can stack some functionality... The 3 address digital outs will feed the two address inputs of the black alphanumeric module, and the three inputs to the decoder chip. One digital out will go to the black module's Write Enable, and when activated, will load a character from the 7-bit data bus into the address selected by the first two bits of the three address Digital outs. The second digital out will be the Latch Enable, and will activate the enable chip connected to the output of the 3-to-8 decoder chip. That will pass one, and only one selected output to the one selected numeric display, updating just that one digit. This two chip solution compresses my Digital out requirements to: 7-bits for the data bus, 3-bits for the address bus, and 2 select outputs. That's only 12 wires, and even leaves 2 Digital I/O free for serial communication. With only 2 chips, I'll be able to get an Arduino Pro Mini to be able to handle all the I/O to control this display.

That's another thing... I think having clean 0v/5v signals from a proper controller will clean up these weird unexpected mystery voltages that I'm seeing, trying to mess with breadboards and pull up resistors, and what not... At least I hope so. Those digits were expensive! $10 per digit! I don't want to have some kind of fault that damages them. They already get pretty toasty, so I made sure to keep the rear clear, to allow for a heatsink to make contact with he rear of the displays.

Also... one more thing...
I might have found unobtanium!
If this works, it will require hand grinding to make it fit, but I found some hemispherical lenses meant for LED backlighting. I have no idea if they will show digits clearly through, or if they were a waste of $10... But the idea is, if I grind the sides of these lenses to form 10mm wide rounded rectangular lenses, then I may be able to place them over the displays and magnify the digits, like an old school bubble lens display. Again, no idea if it will actually work or not, but I'll give it a shot. It was a $10 gamble... $1 per lens is better than $10 a digit!

Edited July 8, 2021 by richfiles

[tsaven]

Man, this whole thing is just unrealistically good.  You're making us all look positively lazy by comparison.

[richfiles]

Interesting development in something that has sat on the back burner even longer than any recent developments! Waaaaay back in 2017, when I did the diode ROMs for the DSKY units readouts, I decided I needed a 4-to-16 decoder to take a simple 4 bit value and decode it into a single output to send to each character select of the ROM... I found some 74HC154 chips that would do the logic decoding, but forgot about the basic issue of driving the LEDs. A digital logic chip won't put out enough current to light the whole display. I figured that I could do a simple transistor amplifier, but basic transistor amplifier circuits are inverting circuits, meaning high signals become low, and low becomes high. This complicates things, as now the 4-to-16 decoders outputs would be inverted, which is not what the diode ROMs require to be driven.

Just so happens, I came across a sale at a surplus parts supplier I frequently buy from, and they had packages of 74HC4514 chips for a very good price. I was unfamiliar with this part, so I looked up the datasheet. The chips I had previously bought were active low outputs, which is how the diode ROMs were built to be driven... The problem with the 74HC154 chips, is that when you add a transistor driver to those active low outputs, they will become active high driven outputs The datasheet revealed that these 74HC4514 chips have active high outputs, meaning when fed into a single transistor inverting amplifier circuit, each driven output will become... drumroll please... active low! This is exactly what is needed!

Notice how the A0-A3 columns are identical, but the Output columns are flipped opposite of each other? THAT is exactly what I needed!

What's even more, is that this 74HC4514 chip has an additional feature... The inputs are latched! I removed the latch information from the truth tables, as it made certain parts not line up, and the purpose of the tables above is for comparison of just the outputs, but basically, these new chips can latch the data on the four input lines (A0-A3). This means that I can now run these chips on a bus, just like the velocity display chips, instead of have dedicated lines for every single input of every diode ROM. Now, since the chips have latched inputs, I can send data over the four wires, to all the diode ROMs inputs at once, and then select just the one that I want to latch with a single latch signal. If I had 4 diode ROMs, the old chips would require I have 16 wires to control them all (4 wires per ROM) and all those wires would have to be held in their states. Now, I need only 4 data lines and 4 chip selects to enable the latch on each. That's not bad, and quite frankly, after learning how straining the velocity readout's latch selects were, I'm actually glad this chip makes this so much simpler!

Edited July 16, 2021 by richfiles

[cyberKerb]

Hey Rich, something for reference / idea on using your Navball

 

[richfiles]

Oh man, so he rebuilt the ball... Wow... That's impressive work! Me, I'm dead set on keeping it original, and given that my old employer used to manufacture synchros, motors, and motion controllers, I have the right background for it. The code for the controller intimidates me more than anything... Except maybe the inverter. That's gonna be interesting to build. I need to find an inverter circuit that can take a sinusoidal input and output a true sinusoidal AC supply, at up to 115vac at 400 Hz. My waveform generator is an 11-bit DAC with 2048 steps of amplitude... So it is for all intents and purposes, a legitimate true sinusoidal waveform. Many inverters do modified sine, which is just crudely stepped square waves that have a brief rest at the zero crossing point. I don't wanna use a modified sine wave, cause square waves have terrible harmonics, and a lot of these 400Hz aviation instruments "hum" at 400Hz... Last thing I want is harmonics coming through! Furthermore, I gotta find a circuit that expects an input in those transitional regions... Some types of circuits draw undesirable current when not fully on or off. Whatever inverter circuit I settle on, has to be a true analog amplifier, and not something that expects effectively digital switched states. I've really put zero time or effort into researching that bit, and am only running off vague memories of analog circuit dabbling from many years ago. I'll get it figured out eventually...

Edited November 28, 2021 by richfiles

[cyberKerb]

Using syncros would be quieter wouldn't it?

You could also see if a specific youtuber might want to use it for a content series / show & tell. You know of CuriousMarc? While the Navball is interesting and space related, it my not be vintage enough for him. However, if he's interested, it could be a win-win in that he gets a view videos out of the device restoration, and you get a circuit to help power it in your project. 

 

[richfiles]

I am so jelly of CuriousMarc and Fran Blanche, who have both had wonderful opportunities to handle real Apollo hardware!
It's always a joy to watch their videos! 

Honestly, most recently, I have loved his HP 9825 repair series. I have an HP9825 myself, and last I used it, it worked, but now I realize I had better install a crowbar circuit into the power supply (if it doesn't already have one), so if the power supply ever fails like his did, it doesn't absolutely wreck my machine. My workshop isn't a nearly as decked out as his is, and honestly, I don't think I have all the equipment to easily fix mine if it ever releases the magic smoke... At least if the repair is as complex as his repair was.

As for noise, syncros would actually be noisier, as I'd need servos to drive them, and they would likely exhibit a small amount of their own 400Hz hum, adding to the hum of the synchros and servos already inside the navball. My synchro emulator circuit is actually a thing that already exists... Problem is, flight simulator hardware is expensive and proprietary, but I know how it works... I literally used to work at a company that manufactured synchros. I want to design my own version that is open source, so people can make synchro emulators on the cheap, without being tied into expensive, closed source proprietary systems. All the parts are electronic, so no noise. The only noise will be what the navball's mechanical components make. I do want to use a proper sine wave generator for the purpose of avoiding the many harmonics of a square wave. My worry is that those harmonics would be audibly amplified by the mechanical hardware in the navball. The digital oscillator I have is old enough to be considered retro tech, but it's a high enough resolution output that even the tiniest filter capacitor will certainly make the output smooth.

[richfiles]

I think I've found a promising power amplifier circuit for powering the Navball. It's even period correct!

Quote

Developed to increase output power of digital-to-synchro converter systems while providing stable and accurate output and overall gain even with reactive loads, and includes overload protection.
Delivers 95 VRMS 400 Hz continuously into 500-ohm load.
Power bandwidth is about 20 kHz.
Foldback current limiting drops short-circuit current to 200mA when load exceeds 300 mA.

F. H. Catter-molen and J. A. Pieterse, Digital/Synchro Amplifier Features Overload Protection, EDN Magazine, Nov. 5, 1977, p 107-108.

The power supply is kinda nuts... +/-150 volts! Yikes! 
Well, if it does the job, then it does the job!

It comes from a 1977 issue of Electronic Design News, but it is designed to do EXACTLY what I need to do... Provide the powered high voltage 400Hz reference/power output for driving synchro based loads, using digital synchro emulation or conversion. I'll have to order the parts and build it up, and see how well it works!

[richfiles]

Small update, which may likely lead to a major update soon.

The TL;DR is that I posted the schematic above to just the right place... Turns out it caught the attention of an individual with far more extensive analog design experience than I have. They got curious enough about the project that they seem to have taken a genuine interest in coming up with a solution... For the fun or challenge of it. So far, the guy has already reworked the 1977 circuit into a modernized circuit concept, has run the circuit in a SPICE simulation to confirm operation, and verified it can be put together with commonly available modern components. The most expensive part is likely to be a transformer... 120vac primary, 230vac CT secondary. I suppose just a 230/240 to 230CT for other regions of the world. That's probably only a $20 part. I was thrilled just having a circuit that ought to work. Later today, after I got home, and I asked how he preferred to be credited, he tells me that now he's trying to design a PC board for it as well! Like... wow... 

All this started over a humorous post of how many tabs I had open in Chrome (All of them) 
That lead to the inevitable question of... my sanity, but then to the obvious question of why... That led to "Interesting..."

I'm beyond grateful. even if things don't pan out, I've gotten some great insights out of this! I wish him all the success, cause that translates into success here!