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

[T-Bouw]

Good find! I love those old screens, they have a charm of their own.

Reminds me of the old computer terminals. On that note, perhaps it has a simple serial connection to the crt somewhere inside?

[Cairol]

I like the "OH SH..!" Button in the first posts video.

Every spacecraft should have one.

[madlemur]

Please tell me I wasn't the only one totally distracted by the control panels in Dr. Pym's lab in Ant Man​...

[richfiles]

So, I've been looking at this CRT while I wait for my FDAI controller parts to arrive on a slow boat from China.

So I've done a little bit of reverse engineering on that CRT, but I have a LOT of work left to actually do anything with it. It has 3 signal wires. The three signals are fed through some inductors and resistors, and that leads directly to a programmable logic array being used as the word processor's "processor". The good news is that means It's a TTL driven video signal, and I can just desolder the couple analog components to recreate the signal conditioning circuit. The problem is in actually driving the signals themselves. Since the chip is a custom ASIC design, programmed into a logic array, it means I have no "video chip" to hack. I have to BUILD the video driver from scratch. It's almost certainly likely that with as small and simple as the monitor's PC board is (it's the small rectangular board to the right of the CRT, viewed from behind), that the three signals are horizontal sync, vertical sync, and the actual video stream.

In order to do this, I'm going to have to measure the signals using an oscilloscope. I'll need to measure the timing for both the horizontal and vertical sync, and measure the timing for the pixels too. Until I do this, I don't know the frequency it refreshes at, nor the scan rate for pixels. It appears to be at least an 80 column display, so even though it's reduced height, it'll be pushing a rather high pixel rate per horizontal line. Each character on the screen (with the original controller) is 8 pixels wide (including the gap between letters). That means this monitor has a 640 pixel horizontal line resolution. I'm guessing the vertical resolution is around 300, give or take a dozen or two pixels. Hard to tell from the large blank spaces between lines. The Arduino TV out libraries don't exactly come close to this resolution. Even if this updates at only 30 Hz, that's still over 5.7 million bit flips per second, over 11.5 million if it refreshes at 60 Hz. That's really pushing an Arduino. They only run 16 MHz anyway. I might look into an Arduino Due, as I understand they are much faster. I only need to deal with monochrome, but there is a lot of bits to deal with for 640x300-ish. I might also look into some older style hardware. If I can find an old chip/chipset from an 8-bit/16-bit era computer that actually handles generating the video signals and maintaining the video memory, then I might try to make something like that instead, so I can just send data updates to it, and have it handle everything. I might also try a hardware based setup. I have TONS of CPLD chips that I can program logic hardware in. I can also scavenge some SRAM chips from the Word Processor board and create a scanable video memory. That might be another viable way to go about doing this. If I do it right, I might be able to do a proper bitmap, and display graphics on it.

If I figure out graphics, I may skip regular text generation, and instead do something like an orbit visualizer. Now that the plug-in outputs SoI, I can potentially display SoI and render the orbit based on the already existing orbital data. It would not be particularly advanced. A circle to represent the body being orbited, and a circle/ellipse/arc representing the orbit. I don't think we have enough info to do SoI changes, but if we can get time to next SoI as a transmitted value, then that point could be plotted on the map. Since the screen is so wide, It might pay to do two angles: a top down, and an edge on view. What I can show all depends what's sent in the data packets.

so, yeah, that's what I'm thinking about while I wait for the FDAI parts to show up.

[Sputnix]

Wowsers! That's awesome

Not being able to see the monitor itself (and not seeing anything from the pictures you've posted) -- does the unit have any i/o at all? [eg. serial]

Would it be possible to leave the processing being done by the existing system, and send data to it to display what you want, rather than hooking straight into the display?

[stibbons]

Hey Mulbin, care to add my project to your post? I've been updating http://forum.kerbalspaceprogram.com/threads/102513-My-KSP-control-panel with build progress, and it currently looks like this: http://i.imgur.com/3INZrTm.jpg

[richfiles]

No, the word processor has no external I/O, and internally, it appears to be configured with a programable logic array as a processor, rather than use a traditional CPU. This would be a very late model stand alone word processor as well, as the 1994 dates inside indicate it was made very close to the end of the run for this type of non PC word processor. As for how a signal is being displayed, at a simple glance, I can tell you that the main board is generating ALL the raster data from inside the ASIC. The circuit driving the CRT itself is simply not complex. It's an incredibly small and simple board, with very few semiconductors even! It's literally taking external signals, and using those to drive the electron gun and the deflection coils. It breaks down like this: I need three signals. Video, Horizontal sync, and Vertical sync.

The Arduino TVout library only seems to generate composite sync signals (not to mention has a sad resolution). I also don't need any type of color support, so the tricks to send out extra bits with port writes doesn't help, as I still need them streamed sequentially. There is a VGA library, but again, I have no need for color, and I don't know if the timings can be readily adapted easily. It'd be better for me to try a NON arduino based means of controlling this, as pushing an Arduino this hard seems like a bad starting point for someone just learning C.

I've actually been looking into a few retro CRT driver chips. I've found one that can be had for under $10-20, seems common enough, can drive 80 column text mode, and do pixel addressable graphics, and has a rudimentary line, arc, box, and circle drawing capability as well. My problem, again, is programming inexperience. Nothing more, nothing less. It's gonna take me a good long while to even learn how to interface this chip. It's not as straight forward as I was hoping.

The NEC uPD7220 - The datasheet with TWO pages of just contents!

http://bitsavers.trailing-edge.com/pdf/nec/uPD7220-uPD7220A_User_Manual_Dec85.pdf

[richfiles]

Oh AWESOME! My trio of International Instruments 1251 dual edgewise meters shipped! About TIME!

The 1251 is the one on the right with double meters. It's 6 inches (14.7 cm tall)!

Got them 76% below normal cost! The company I got these from decided to try ebay. I got a bid in for $19.95 USD. They normally sell at $85 USD. I convinced them to send me THREE at the $19.95 price!

I'll get a new set up pictures when they arrive, showing the three 1251s beside my four GE 180 meters. That's 10 readouts worth of analog information! Combine that with the extra glideslope readout on my FDAI, and my 4 inch (9.8 cm) round vertical velocity gauge... That's a total of 12 analog readouts that I can hook up! There's also the two heading readouts on the FDAI, but those don't count... They are reserved for when we can get velocity/target/maneuver nodes transmitted in the packet. Haha! I'm talking about maxing out the number of PWM lines on a Mega here!

[richfiles]

Oh my God... I freaking LOVE the guys over at the EEVBlog... They found the SERVICE MANUAL to that word processor, and they inform me that I apparently have a tinkerer's dream. Not even a hacker's dream. The service manual documents EVERYTHING... There is LITERALLY nothing to hack. Just reprogram it. It's a run of the mill computer running an enhanced Z80 CPU with the program in ROM. The Service manual documents the ENTIRE machine, right down to even the memory map. All I'd need to do is learn how to program Z80 assembly, and write my own code for it. Just desolder the ROM, and replace it with a socket and an EEPROM.

Granted... That's a lot of work. I don't know Z80 assembly, but I HAVE learned that these old word processors are AMAZINGLY reconfigurable! It's basically a 12 MHz Z80 computer with a built in wide screen mono display, a floppy drive, a buzzer, a keyboard, three built in stepper motor drivers, a DC motor driver, a pair of solenoid drivers, and at least 2 inputs OTHER than enough inputs to encode the entire keyboard! The screen resolution is actually 819 x 240. VERY wide screen!

Looking at this thing... I almost feel bad that all I want to do with it is display a few circles and arcs! If I actually could learn Z80 assembly, I bet I could handle most of the freaking control panel itself! The only thing it doesn't have are analog outputs... and the machine has unused memory addresses! You could add DACs in those memory locations if you really wanted!

This thing is INCREDIBLY user accessible!

http://www.maisiemanuals.com/brother-word-processor-2450ds-parts-service-manual-p-1515.html

Even without knowing assembly, It might be possible to create a loader that loads data into the video RAM (which is a separate chip), and just let the machine run without input on the CPU. In doing so, it'd be possible to directly write bits to the screen... Though it seems SUCH a waste of a perfectly good Z80 computer... If only I knew Z80 assembly...

[Mulbin]

Hey Mulbin, care to add my project to your post? I've been updating http://forum.kerbalspaceprogram.com/threads/102513-My-KSP-control-panel with build progress, and it currently looks like this: http://i.imgur.com/3INZrTm.jpg

http://i.imgur.com/3INZrTm.jpg

Added! Looking great!

Some awesome projects on this thread now. I was looking at my panel in storage last week... won't be long now, had an offer accepted on a new house with space for a simpit

[Sputnix]

It's basically a 12 MHz Z80 computer

Pft! Arduino's are more powerful at 16 mhz!

You should be able to find a few guides on the internet... or possibly a way to compile 'down' to z80 from a higher language.

[madlemur]

GCC based compiler for "Small Devices", including the z80: http://sdcc.sourceforge.net/

Build Your Own Z-80 Computer by Steve Ciarcia: http://retro.hansotten.nl/uploads/z80/BuildYourOwnZ80.pdf

Edited August 5, 2015 by madlemur

[richfiles]

Pft! Arduino's are more powerful at 16 mhz!

That CPU may be only running 12 MHz, compared to an arduino's 16 MHz, but the difference is the addressability of this chip, and the fact that it has a custom gate array performing as a graphics chip (and more), offloading much of the hard work in creating video.

Just thanks to the addressability alone, this CPU can have up to 64K-bytes of I/O port mapping, and up to 512K-bytes of directly addressed memory (compared to a Z80's 256 I/O mapped ports, and 64K of directly addressable memory). it supports bank switching too. On this particular word processor model, it has a 128K RAM, a 512K program memory, and 32K of Video memory. The service manual states that the program ROM can be expanded (via bank switching) to 8 M-bits (1 Megabyte).

Video, the external I/O latching, and the keyboard matrix scan are all handed off to the gate array logic chip. The 32K video RAM is on the other side of the gate array (it handles it). The gate array chip runs at 15.33 MHz, and handles port mapping, video generation, and almost all the aforementioned I/O controls. Unlike a CPU running at 16 MHz, which is a serial sequential instruction executing device, the gate array logic chip handles all it's functions in hardware... It's performing all it's functionality at the hardware logic level, as a hardware state machine, and not as a serial instruction executor... at 15.33 MHz... An arduino takes almost EVERYTHING the Atmega's got to just barely generate a 128x96 pixel signal. This gate array hardware offloads all that work, and makes it possible to generate 819 x 240 @ 60 Hz refresh rate, without the CPU needing to even do anything, unless it wants to actually change what's on the screen! That's a thing of beauty, I tell you!

According to the datasheet for the CPU, it has 12 additional instructions over a stock Z80, a wider address bus (A0-A19), a built in memory management unit with a 2 channel DMA controller. It has two channel Asynchronous Serial Communication Interface and a 1-channel Clocked Serial I/O Port. These don't appear to even be used in the word processor, meaning they are AVAILABLE for communications. There are other goodies in it too. It features a two channel 16 bit programable reload timer, which is described as being useful for "counting, timing, and output waveform generation"... It's got enough memory, I could do straight lookup tables for EVERYTHING, rather than waste cycles calculating ANY sine values. I could literally drop the synchro multiplier formulas in a spreadsheet, make tables for every third or quarter degree tick (or whatever resolution I decide upon), and make a lookup table for the three multiplier values! One of the new commands, IS a multiply command! It also has new commands that can basically load a value into a register immediately on reading a memory address, or vice versa. It has rapid block I/O commands that can increment or decrement, or run once or repeat (another potentially useful tool for loading lookup table values, and for transferring data to video memory) This chip, combined with the gate array "graphics chip" is practically begging to do fully graphic video output and make me a sandwi... I mean make my FDAI "navball" sine waves!

Indeed, a Z80 C compiler makes sense, though it'll need to be an HD64180 specific compiler, as this enhanced Z80 variant has features the regular Z80 just simply doesn't have. I'd most definitely want to take advantage of those additional 12 instructions, and all those fun extra features! I mean, guys... The Z80 required you to loop "add" functions in order to multiply... the HD64180 HAS a native multiply command! With all those FDAI synchro multipliers I want to generate... Wow... I bet i COULD get double duty front his thing and drive both the graphic CRT AND the FDAI synchro sine wave signals!

This thing is more than "just another Z80", and with the added support hardware, it's got a fair bit of beef to it, with memory that would make an Arduino Mega2560 CRY! Stock setup is 2x the program memory (4x with bank switching), and 16x the RAM! Arduino has no native video commands or signal generating capacity. This puppy has a hardware gate array logic graphics chip that handles all the signal generation and memory management, and adds 32K of video memory... This thing has 4x more VIDEO memory than a Mega2560 has TOTAL RAM!

It's all about the support hardware. Arduinos make great I/O and sensor platforms. There are lots of peripherals that do those things... This machine that I have, is specialized in graphic video generation and pulse timing (for operating the printer mechanism). Since I don't need the printer, I could basically "steal" those programable hardware timers to handle my waveform generation!

Tinkerer's dream indeed...

You know what's sad... I have always been a hardware guy. I design stuff using logic gates. I've built Arithmetic Logic Units and stuff like that in Little Big Planet 2/3. LBP is a "platformer" type video game that lets you create your own levels. It has a cool "circuit" feature that lets you wire up digital logic to control the levels that you create... I only ever just make circuits!

GCC based compiler for "Small Devices", including the z80: http://sdcc.sourceforge.net/

All considered, I'm "getting" assembly faster than I'm "getting" C. That's scary. I still get each about as well as I get quantum physics though.

Once the task of programming scales up though, I know C will be easier to understand. Just need to find a compiler that takes advantage of the extra commands of the HD64180.

Looks like SDCC does have at least some of the extra codes. Hopefully all of them. I need to get that to run on my Mac now.

Edited August 6, 2015 by richfiles

[Sputnix]

I was only joking, of course...

[richfiles]

I was only joking, of course...

I know, but it's a great excuse to brag about what I've found!

Man, this is the little CPU that could!

[richfiles]

I got my DACs in the mail today, along with my switches. I think I'm gonna try something else with the DACs. I need 10, and the max number you can address is 8. I was reading on another forum, that you can use a bus multiplexer to toggle between two I2C busses. I can also get the uniquely addressed chips from a local company (Digikey is located in the state where I live, so their shipping, even UPS ground, is lightning fast). I'm gonna desolder the DACs already mounted to the boards, and drop in the uniquely addressed versions. I'll do 2 banks, 6 addresses in one bank and 4 in the the other. That will give me my 10 addresses, and leave me able to access a total of 16, for future expansion. Since the chips are under $1 each, I might just get enough to do all possible addresses in a dual bank setup. For my own needs, I can leave 4 chips alone, meaning I only need to change out 6 chips on the 10 boards I bought, and I'd be buying 12 chips total (each chip supports 2 addresses, so with 2 banks the stock chips would support 4 addresses out of 10 needed). It would also mean only having to toggle a single digital out vs strobing 10 digital outs in time with the sequential I2C transmits.

I also got my three EIL 1251 Dual edgewise meters... I have a short video of one doing about half range with a sine wave generated by an Arduino Mega 2560 R3. These are 0-10 volt meters, so I'll have to use a transistor or MOSFET to drive them the full range. My four GE 180 meters are actually 4-20mA industrial range meters, so i have to drive them through a voltage to current transmitter. I got the chips for those... WOW, they are tiny... Spent $10 to build it myself, cause the transmitters were $8-9 each... I might have been better off just spending the money on the pre-made ones... I might have to deadbug these chips under the microscope with some 30 gauge wire and my pencil tip soldering iron... I don't think I can etch boards fine enough for them, and ordering proto boards is a no go... They don't have the heatsinking die pad. I may just suck up the $10 loss and order the actually $36 in preassembled boards.

https://www.youtube.com/watch?v=OplHSnYCxvo?version=3&loop=1&playlist=OplHSnYCxvo

Sadly, one of the three meters has the right needle broken off...

They make the long bat from what appears to be coated graphite a thin hollow aluminum tube, to save on mass, so the meter has a faster reaction time. Light, but brittle.

There IS a lot of space inside, so I might be able to repurpose it for something else... Maybe I could play with making a tape meter using a fabric ruler (long flexible plastic tape meant for measuring clothing). It'll have an incremental whole number scale with fractional divisions, and is flexible... If I can motorize a pair of take up reels, and then index optically off the markings on the back (it's common to have inches on one side, and cm on the other), then I could possibly make it work! The curved front face adds an element of uniqueness as well, as long as it doesn't rub the markings off the back. My only issue, The motors probably won't fit inside, because of the thickness, so they'll have to stick out the left side... I was TOTALLY gonna have my meters be on the right of my control panel, and I hope the motors don't interfere with stuff I planned to mount to the left of the meters.

I don't know... It's a lot of work, and I don't know how durable it would be if it does work. Still, I pretty much managed to get 3 meters for the price of 3/4 of one.

Still sucks, but hey, I at least have four single and two dual edgewise meters, plus my round vertical velocity meter! That's still nine enormous analog readouts, even if I set the half broken meter aside. The left meter DOES still work, so I could use it for a 10th meter, but it's just odd, with it being literally only half a meter. I had considered doing a vertical line of annunciators, but I don't want to distract from the DSKY themed annunciators I wanna build.

One option may be to use it for orbital/targeting/node annunciators. I'll have operational annunciators on the DSKY clone, but the FDAI can sit right beside the broken meter. Since you can only display a single vector at a time using the crosshairs, it's necessary to be able to select them. I can have 10 small annunciators to indicate Prograde, Retrograde, Normal, ... Target, anti-target, etc. I know that getting these values calculated and incorporated into the packet are an important thing. We don't have them yet, but I plan to build under the assumption we will eventually get them. The navball is not nearly as useful as an attitude indicator without having the flight director functioning as well.

***UPDATE***

WOW!!! The guy already got back to me and they are sending a replacement for the broken meter. Told me to keep the broken one for parts!

Edited November 25, 2022 by richfiles

[Sputnix]

That's so freaking awesome!

[richfiles]

Well, pics of all the meters below. I took out the scale inserts on the damaged meter, and fed a fabric tape measure through just to see how it looks. I will definitely need to get some teflon sheet to help it slide better, and I'll need two rollers at each end, and two reels. I'll look for a tape with horizontal numbers.

I have to admit, I'm a little shocked by the amount of space this takes up. I'm seriously considering mounting the meters on either side of the central controls, on a second tier of desk cutout. I definitely want the FDAI navball centered, and I want the DSKY clone near that. The vertical velocity meter needs to be close to front and center as well... That means DSKY, FDAI, VVM... I only planned the primary cutout to be as wide as a central keyboard segment (letter keys only, no tenkeys or page keys, with arrows arranged below, not to the side), and a pair of joysticks on either side, plus the throttle. May have to be wider? Who knows...

This thing's getting big! I'm definitely throwing the VFD displays as an overhead display. I may take the broken meter, and do horizontal scaled for it. If I make a tape meter on the other half, maybe I can make it into a radar altimeter??? The fabric tape I have has a scale of 0-150 on the cm side, and 0-60 on the inches side. Ideas? Of course, there's also still the idea of vector annunciators... Ugh... I think I prefer that to the complexity of building a tape meter. I dunno...

Edited August 10, 2015 by richfiles

[richfiles]

I have to share this little gem of a document I found... I MUST!

A lot of people seem to enjoy modeling their Kerbal controllers, or at least taking inspiration from, the Apollo program.

What I've found is a NASA document that was buried on NASA's website. It has INSANE details about the little things... I mean, You just have to read it...

https://www.hq.nasa.gov/alsj/tnD7919DsplysCntrls.pdf

They don't just describe what was used or how it was put together, but often why!

Did you know the prevalence of vertical edgewise meters is actually because of crew preference? They said the vertical meters were just easier to read at a glance. They also could be packed tightly together, efficiently using precious panel space. The LEM contains no dual scale semi-circular meters, because by the time they got to designing it, the crew had made clear that those things were awful to read, and so none were used any further in the design. Vertical edgewise meters and single scale circular meters continued to be used on the Space Shuttle, thanks to the easy readability. They also considered designing their meters so the needle moves completely off scale in the absence of power or an input signal. This provides positive verification that the meter is functioning, and if what it's measuring fails, the needle disappears off the bottom of the scale, to indicate the extreme fault, as opposed even, to just reading 0. Not sure if this functionality was actually implemented, but it is discussed in the document. At the very least, all those types of meters would drop their readings to at least 0, in the absence of power or signal. The few servometric meters that were required included a fault light. A servometric meter, in the case of a failure, will read what ever the last reading was... Not a good thing during a fault. Tape meters and scroll meters were types of servometric meters used.

The dial faces of all the meters were back lit with EL lighting. The transilluminated markings of the command module meters were made with an etching technique, while the LM meters used a film overlay. In both cases, where color bands existed on the scale, those colors were painted (using translucent paint) directly on the meters. They actually had problems with this paint flaking!

This document also led me to find the original switch manufacturer (and the fact that the switches are still manufactured, even if ridiculously expensive), thanks to the descriptive detail of the "wedge-shaped tab handle that provided the crewmen a large purchase area with which to actuate the switch while wearing a pressurized glove". Exact quote!

The locking lever switches had "large bat-shaped handles" and both the lever-lock and wedge-shaped toggles had radio-luminescent tips that glowed under low lighting conditions, such as when the cabin lights were dimmed for the crew to take star readings. While they are not locking levers, there are a lot of toggle switches on ebay with LED tips. If you use a larger than standard resistor in series with green LED versions of these, you can recreate the appearance of the lever-lock switches very sufficiently! Basically, you create a VERY dim green glow at the tip of a cylindrical bat shaped switch. A perfectly fine cheap analog of the originals, without devastating your wallet! If you really want to get the look right, you can try to find an aluminum tube with an inner diameter that closely matches the outer diameter of the bat. The true lever -lock switches had a larger handle at the tip. It's a cheap and easy mod, if you wanna put in the extra effort.

Other details, include the fact that NASA referred to the switch guards as "wickets". I previously mentioned that the recessed "troughs" that the switches sat in were the result of surrounding the toggles with the plastic nomenclature panels. Those panels were backlit by EL (electroluminescent) panels. Turns out the spill lighting from the edge of those plastic nomenclature overlays was used as the means to light up the toggle switches as well.

The document goes into intricate details like the exact dimensions of pushbuttons. They went with slightly larger pushbuttons than were typical on most aircraft, and even the Gemini capsules. They found 0.8 inch (2.03 cm) square switches to be better suited for gloved hands, and also allowed larger, more easily read legends. They state that the master alarm switch is 1 square inch (6.5 square cm), and is actually rectangular, vs square. They even describe the acceptable range for the actuating force (3 - 21 newtons) of the button.

All the rotary switches had 30° detents, and featured a round skirt. The knob had either a hole in the skirt or an actual slot in the knob that let the EL backlit panels shine through the knob and illuminate a position marker to indicate the selected position, or in NASA's words "to allow for transillumination of an integral pointer indictum". They liked the rotary knobs for weight savings, panel space savings, and simplification of operation, but were always worried about the fact that if you have a failure, more functions than just an A/B toggle could be lost. All rotary switches were pre-set to the most frequently used or most critical positions pre-launch, just incase of failure during the mission. They even described some manufacturer supply incidents where they failed to indicate the rotational torque desired for some rotary controls, and got parts that were completely unsuitable. They ended up fixing the low torque items with friction washers, and the high torque ones went into sims and prototypes... Aka, non flight applications. The knobs were changed from plastic to metal, to reduce flammable materials inside the cockpit after the Apollo 1 fire.

Sadly, I STILL haven't found an equivalent knob. I'd settle for just the pointer portion, as I can always glue a disc to the bottom. That's easy enough to do. They really do have an incredibly distinctive style. I'm a little surprised no one has tried to copy it. It looks good to me!

The circuit breakers were generally accepted as the actuator most susceptible to breakage. As most of us realize, this actually happened on the Apollo 11 LEM, where Buzz Aldrin famously used a pen to fix the problem. The circuit breakers used a black plastic push/pull actuator. They originally had a white painted ring that was exposed when pulled into the open position. That was replaced with an aluminum band, because the paint proved prone to flaking. They used a combination of recessed panels and long wire guards that went over the top of several switches, and secured to either end of the row. Sometimes both methods were used. I swear I have seen the shape of the recessed panel dividers... I wanna say I've seen aluminum/vinyl siding use that shape for structures used in either eaves or something like that... I KNOW I've seen some commonly available material in that EXACT shape before!!! Maybe I'll run down to the hardware store this week to job my memory... It DOES make for a rather cool switch arrangement, making it look like "tiers" of switches. Looks good, to me! I'll get back to this later, if I find the material I'm thinking of.

While annunciator lights were commonly used for alarms and many status events that were of immediately critical importance, they actually used a good number of status flags. These were mechanical parts, constructed very similarly to an analog meter. Instead of a needle though, they had a "flag". In the off position, it might show white and black diagonal stripes, and in the on position, the flag would cover the stripes, and show solid grey. Some flags had three positions, where the third position would show red, to indicate a fault. They preferred the flags for general status indicators, because it kept the control panel from looking like a Christmas Tree, a major issue if you need everything dim to take star sightings. The problem with flags, is if there is a fault, they are easy to miss, since they don't actively illuminate, thus the reason warnings and cautions used annunciator lights.

The majority of the annunciator lights were for cautions and warnings. Cautions were amber, warnings were red. The alarm tones for a CM fault and an LM fault were also different. The CM fault tone was an alternating 750 Hz and 2000 Hz tone. The LM fault tone was a steady 3000 Hz tone. The LEM used several "layers" of annunciator lights to indicate all subsystems gated into each single caution or warning light. The CM used flags to specify what sub system a particular caution or warning came up on. The reason for the LEM going all out, was because if a fault happened on final decent, you didn't have time to look for flags. You needed immediate information.

The Lunar Contact annunciator was unique, in that it illuminated blue. It would illuminate if the 3 meter (about 10 foot) long contact probes on the landing gear made contact with the lunar surface. Pretty nifty bit of info. I suppose if you have a blue LED light up any time the radar altimeter reads 3 meters or less, you could actually recreate this functionality in a Kerbal controller! The lights were round, and were 1 inch (about 2.5 cm) in diameter.

After the Apollo 1 fire, an addition to the design, were nozzle holes in the instrument panels. They could shove a fire extinguisher nozzle into those holes to extinguish a fire behind the instrument panel. I think those are the unlabeled holes with the red ring around them. I never knew that.

One part of the document describes how they had considered using radioactive sources on one side of the RCS tanks of the LM. scintillator-photomultipliers would be placed on the other side, and would measure the quantity of propellant based on how much absorption/scatter of the radiation occurred. They had planned to use nixie tubes to display the results, but the entire method/project was canceled in favor of indirect quantity gauging, based on temperature and pressure. They specified that that system, had it been completed, would have been the only time NASA had ever employed the use of nixies onboard a vessel. Alas... it never happened. Still, radioactive RCS nixie tube fuel gauge. Wow...

ORDEAL (Orbital Rate Display Earth And Lunar) was actually developed so late into the design, that it exists as a literal black box attached to the wall... There was never any stage of the design where it was integrated into the instrument panel. It was literally hung on the wall and wired into the FDAI, as an afterthought! It was the device that generated their horizon, relative to the orbital body. Without it, they only had inertial reference. Turns out, it was a very much NEEDED afterthought!

The LM design resulted in the development of a nifty little cross pointer meter that simultaneously reported both vertical velocity and horizontal velocity on the same readout. Interestingly, they had considered an EL panel grid (pixels) to display the cross digitally. They didn't build it, because it would have required a lot of extra hardware back then to convert to a digital readout, and the technology in those days generally only had a refresh rate of 10 samples per second. By the time the document was written, they had determined that the technology that currently existed (circa 1975) now made it feasible. Analog was just simpler to pull off back then. Of course, you can easily recreate this with a small LCD or OLED display, or even a grid of MAX7219 LED drivers. Wire up a square of either 4 or 9 8x8 LED matrix displays, and you have a digital readout for both horizontal surface and vertical velocity! Personally, I kinda like the idea of the old analog style. Man... to this day, I regret throwing out an old stereo system that had a cross pointer meter... Dang it... Would have been GREAT for such a readout.

They actually brought up a genuine NASA certified kitchen timer (aka an egg timer) at the recommendation of the Apollo 7 crew. The Apollo 8 crew used it frequently!

Don't forget to have a kitchen timer at your Kerbal controller... They are great for timing long maneuvers, AND NASA approved!

Finally... Some eye candy...

Look at that cool, EL glow... I have GOT to figure out how to make those backlit panels! They are just too darn cool!

Have some really beautiful pictures to inspire you all!

Edited August 10, 2015 by richfiles

[Sputnix]

Always a font of information. Thanks for the find + share!

Someone linked me to the EL cable stuff recently (was it you? )

The Apollo Experience report is another fun read (your find is amazing too!).

I also found a heap of documents on Mission control stuff (for when I get around to building one of those ).

But seriously - you need to build a full on simpit (and you could fill it up quite nicely with those dials).

One way to do the back-lit panels is to get some clear pespex (Acrylic?), stick a cold cathode light on the bottom / top edge, and paint in everything but the lettering. So you'd have to do some masking/ stencil so that when you took it off you'd be left with the clear perspex lettering + lines. Then, the lights would bleed thru the perspex, lighting up the lines / text...

Although super-bright LEDs might be the way to go these days.

does that make sense?

[richfiles]

Oh, it makes sense, but the worry is longevity. The Apollo CM and LM only had to function for a few days to a couple weeks at best. Painted surfaces are a liability... NASA even points out that paint was a continuous thorn in their side in the article I linked. I'm thinking that laminations are the only way to go that will preserve paint, but I kinda wonder what going with vinyl something would cost. Laminate it to the back of a thin sheet, then lay that on a thicker sheet. I'm concerned about costs and feasibility, vs durability and longevity.

And thanks for that other document! That was awesome!

I only mentioned the use of electroluminescent light as being the light source used by NASA on Apollo. I didn't link to any mentions of EL cable though. Of course, LEDs are probably sufficient if you have uniform enough coverage.

[madlemur]

No, not mine. This is the beginnings of my hardware controller. I've had an Arduino for some time, but I just recently ordered various switches and displays for building it up. So far, I have the core indicators: master alert (bi-color led), SAS, RCS, Brakes, Gear, Lights, Abort, and connection (the sole green led), and an LCD display I'm trying to get to display Ap and Pe. Waiting on some push buttons (haven't hooked up any toggles yet), and shift registers to work on a full annunciator.

From my local blog: http://forum.kerbalspaceprogram.com/blogs/119785-madlemur

Still working on the Ap/Pe display. I can get the correct Ap on the launchpad, and for a few thousand m after liftoff, but at some point, it begins wrapping around to negative numbers, and the Pe is never right. I'm assuming it's a float to string conversion issue, but I haven't had enough iterative cycles to nail it down; spent most of them getting all the leds to light correctly.

[Mulbin]

OK... just had the most amazing VR experience!

If anyone else on this thread has an Oculus Rift, you really need to back and download the latest version of the Apollo 11 Experience. I was just sat in the command module peering at the minute detail, every switch laid out and labelled. Can't get a much better visual source than that!

It's also really got me thinking, I've had some success in getting KSP running in full 3d VR (some bugs to iron out but close).... so maybe my simpit will now have an airlock. With an oculus rift inside is. Should make EVA's interesting

Anyway, here's a screenshot. Only in the rift you are actually there, fully surrounded and feels totally real.

[richfiles]

OK... just had the most amazing VR experience!

If anyone else on this thread has an Oculus Rift, you really need to back and download the latest version of the Apollo 11 Experience.

I don't have an Oculus Rift... I crossed my eyes instead!

[richfiles]

So, I'm trying a product out, made by a Wisconsin company called Nekoosa Coated Products. They make a cavitated polyester "paper" that is compatible with laser printers called Synaps Digital XM Synthetic Paper. I ordered some, with the hopes of seeing how it looks under acrylic. This may be a viable option for those without access to a CNC mill or a laser cutter to create their panel art by printing it on a home laser printer or going to a copy shop to do so. The 5 mil thickness, I'm told by their sales rep, offers excellent diffusion for backlighting.

Color reproduction on my printer is crap, considering I need to replace all three color toners, possibly the transfer drum, more probably the whole dumb printer...

Dumb thing leaves streaks.

My solution is going to simply be to go to a professional copy shop and show them the product, and see if I can run the meter faces and panel art on the specialty paper one of their machines. ALWAYS ASK. Don't just feed specialty paper into a $10000+ photocopier/printer without authorization... You don't want to be liable if something DOES happen!

Anyway, the method would be simple. white text, labels, lines, etc for backlight shine through. I don't know how opaque the toner will be, so I may experiment with printing solid black on the back side, with squares left open near where text or art must shine through. It's all an experiment for now.

If this actually works, and produces good looking results, then this should be a reasonably feasible way to do good panels without needing CNC or laser hardware! I'll let you guys know how it turns out once the materials arrive.

GAH!!! My "Green" MAX7219 LED displays just showed up today... Every single one of them came in RED... Grrrrr...

Maybe I can use this as an excuse to buy larger .56 or .8 inch Green digits and wire them up instead. Maybe do two sizes. Large for the main stuff, Altitude, Radar Alt, Ap, Pe, etc. and then do smaller displays for the other things. I dunno. We'll see. Any that I go bigger with, will need me to wire manually. Only .36 inch digit displays will fit the clearances of the sockets.

At least the LED displays are socketed!

Edited August 13, 2015 by richfiles