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

[richfiles]

To explain to those who do not know what a synchro is, a synchro is a type of control transformer. The primary winding can be rotated, and the secondary windings output a phase shifted signal based on the angle of rotation between the two. Typically, a pair of slip rings and brushes energizes a wound rotor, which generates a magnetic field aligned with the synchro shaft index (of course, it alternates, because you are feeding it AC). As you rotate the shaft, *the field angle rotates. The stator typically consists of a 2 or 3 phase set of windings (90° apart or 120° apart, respectively). These shifted phase angle stator windings pick up the amplitude of the magnetic field of the rotor at *different phase shifted angles (redundancy department of redundancy calling ) , allowing precise representation of the shaft position.

In the control transformer variation, a higher power motor drive performs the motion, driven by a servo amplifier. There is a receiver synchro mechanically paired to the shaft of the motor, and the servo amplifier will drive the motor in the direction that causes the phase differential between he control transformer and the receiving transformer to be nulled.

It also occurs to me, that between the DC bus driving the arduino and all the digital stuff, and the fact that an AC bus is required for the FDAI, I legitimately have cause to include DC and AC bus meters and switches on the panel, and it gives me an opportunity to trigger the gyro power fail flag, if the DC bus is on, but not the AC! Cool!

***EDIT***

I need to further research this... Speaking with an engineer the other day, He was saying that I may be confusing synchro operation with resolver operation. RESOLVERS are also I part I worked with... in fact, I'd say that I spent probably 95% of my time with resolvers, in comparison to maybe only 5% of my time with synchros. Resolvers use an inductively transmitted reverence and a pair of 90° sine and cosine phases to pick up a rotating phase shifted output.

It seems synchros may actually rely on amplitude modulation in both the positive and negative, but no phase shifting occurs. If this is the case, then that means I could simply take the PWM outputs of an arduino, and drive them through a filter to get AC out of it, and feed that into an amplifier directly. I still have to have the arduino doing some math, but it could end up being rather simple. If I just have it perform math on an analog input, I could theoretically feed a reference AC sine wave into it, and then generate my calculated amplitudes using a simple positive or negative multiplier based on math performed on the input reference (either an input, or internally generated).

If that's the case, I only need to generate ONE sine wave at 400 Hz, and then just generate 9 multipliers.

Edited November 23, 2022 by richfiles

[Antipaten]

I do apologize for the derail, but I couldn't help but respond to an RTL comment with an RDL reply!

You absolutely win!

[faeryty]

Just started my own attempt at a simpit. Not very far yet:

but I feel happy that I'm on the right track. I at least have an altimeter that reliable reports my height above sea level.... so long as that is less than 10Mm at the moment. I must get round to getting a 16 segment display so I can show an SI prefix and scale it.

[ThereIs0nly0ne]

Work in progress:

Link:

http://forum.kerbalspaceprogram.com/threads/102746-CP-1-Custom-Command-Pod-%28DIY-Sim-Pit%29

[richfiles]

I've had these buttons for a good long time, I think since around 2006-07, possibly? Not sure. I plan to use the wide black switches on my DSKY (I really should snag a couple more, but I might just simply do a truncated keyboard). The narrow colored switches are begging to be used for Action groups. I have a pair each of red, yellow, green, blue, and white switches. I think I rather like the idea of having action groups be on these pushbuttons, with LEDs to read out the status. I'll likely have the wider switches manage things like time warp, and other system level actions. I might do a lot of that via the DSKY. Anyway, as a bonus, these pushbuttons have a WONDERFUL mechanical click to them! The recessed part is also good for labels, I figure I can try to find white decal lettering or something (black lettering for the yellow and white buttons). Since that part is recessed, the lettering shouldn't wear off with repeated pressing!

I can easily mod the cap to have a pair of bicolor LEDS. ALL the switches are internally capable of having LED mounted. You just push back a dowel and pop the cap off, and you can drill out either the center for one LED, or two adjacent holes for two LEDs. The caps snap right back on.

To anyone who wants to use these, Digikey carries many of them, but not all (search for "5500 series", then click the "Pushbutton" category). They are spendy switches... $3.28-4.40 each... Even with price breaks, these add up. I got them for another project, but never used them. Some switches are available in single quantities, but most require a minimum purchase of 500. Selection is spotty, but they do have some in stock in single unit minimum quantities. I managed to snag a lot of the last single unit quantity items with colored keycaps.

These buttons are the one big non-Apolloish thing that I'm set on. I know toggles would be better, but i honestly don't care! I've had these for years, they cost me a small fortune, and I wanna use them for something cool!

I'll still have toggles, just not not he action groups. I do still think the wide black ones will look nice on the DSKY, even if it's not a perfect style match.

Edited November 25, 2022 by richfiles

[richfiles]

So... someone on another thread suggested using tiny viewfinder CRTs as a unique display option.

Specifically, they suggested this little project as an aesthetic enhancement to any control panel:
From Hackaday http://hackaday.com/2015/07/07/headphone-amp-features-a-tiny-crt/

That is one smexy little amplifier!

So then... I have a few of these puppies laying around, so I decided to fire one up and send it a source signal. Here's my first camcorder viewfinder CRT powered up...

There are no bounds to my random epic parts supplies!

Check out this itty bitty screen! That is one itty bitty battlebot!

I wanna say I have at least two or three of these screens laying around, but I'll have to dig up the other ones. I'm thinking of mimicking that waveform project that was shared with me, and displaying the game audio (should be VERY cool, since I use Chatterer) as a waveform display. On another CRT, I might create some sweep generators to produce random lissajous patterns. I think it'd be REALLY cool to take a third CRT and loop some Kerbal facial expressions through it. I wonder if Gene Kerman's headset would be visible? It'd be a cool "depiction" of Mission control!

It just occurred to me... that ENTIRE SCREEN is the same size as the space between a single whole number division on that ginormous meter!

Edited November 25, 2022 by richfiles

[Sputnix]

That is one smexy little amplifier!

So then... I have a few of these puppies laying around, so I decided to fire one up and send it a source signal. Here's my first camcorder viewfinder CRT powered up...

There are no bounds to my random epic parts supplies!

Check out this itty bitty screen! That is one itty bitty battlebot!

I wanna say I have at least two or three of these screens laying around, but I'll have to dig up the other ones. I'm thinking of mimicking that waveform project that was shared with me, and displaying the game audio (should be VERY cool, since I use Chatterer) as a waveform display. On another CRT, I might create some sweep generators to produce random lissajous patterns. I think it'd be REALLY cool to take a third CRT and loop some Kerbal facial expressions through it. I wonder if Gene Kerman's headset would be visible? It'd be a cool "depiction" of Mission control!

It just occurred to me... that ENTIRE SCREEN is the same size as the space between a single whole number division on that ginormous meter!

Hah! That's awesome!

I have a few old camercorders lying around (and recently took one apart too); was just this morning trying to work out if there was a way to use the LCD module.

It'd be a pain to work out what wires go where. Did you use some service guide, or were your boards marked appropriately?

[richfiles]

Hah! That's awesome!

I have a few old camercorders lying around (and recently took one apart too); was just this morning trying to work out if there was a way to use the LCD module.

It'd be a pain to work out what wires go where. Did you use some service guide, or were your boards marked appropriately?

Reverse engineered.

CRTs are really easy. LCDs can be hit or miss... Sometimes, they reduce costs by keeping it digital between the CCD encoder and the LCD. If that's the case, you won't find a composite signal, and would have to reverse engineer the entire digital protocol AND find the pinouts... Not easy unless you do actually find documentation. If you're lucky, it'll be a composite signal, just like the CRTs. composite LCD modules were used, but they became more scarce as the technology advanced. A combo camcorder, that has an LCD and a CRT viewfinder will likely use a composite LCD, since it already has to generate composite for the LCD.

The process I go through for reverse engineering a CRT module, is as follows.

Start by looking up your chips. Most modern ones have only a single chip, and right there, you can get a lot of your info. My chip's datasheet suggested in the app notes that the video signal needs to be AC coupled across a capacitor, and that prevents a continuity test. The board was also too dense to trace the wire traces. I'll get back to that.

Only 3 wires are really ever relevant. Ground, Positive, and Composite video (Luminance, or "Y"). That signal is the same as the yellow wire on the back of most any standard definition video device. Usually... I'll come back to that too.

You'll want to REALLY get started by determining your power connections. The chip datasheet can often help, but not always. If a regulator is in circuit, you will have to trace it out

Ground is super easy though. do a continuity check between the negative lead (always marked) of an electrolytic capacitor and each pin of the connector. This finds your ground. If you have the chip datasheet, you can confirm this further by doing a continuity check between your newfound ground connector and a lead on the chip marked as ground.

Next, look for the other lead of the capacitor, and see if it traces back to either a small semiconductor, or directly to the connector. if direct, or if the small semiconductor has one wire to the positive of the capacitor, one wire to the ground, and one wire to the connector, that will be your positive. To test, start with a current limited power supply and turn up your voltage, starting from about 3 volts. pause around 4.8 volts. Many newer models only require 4.8 volts to run, some use 5, some 9 and some very old ones take 12 volts. The idea is to start low, and see when the screen comes on. The chip datasheet, if you find it, can often help you determine the operating voltage. If the power connections go straight to the chip, then feed it what the datasheet says, or only what makes it come on... More, and you may burn it out. If you see distortion on the screen, and it was fine at a lower voltage, you probably went too high, and should kill power and go back to the lower voltage. Be VERY conservative with this... ASSUME it is a lower voltage device, and if it doesn't work, try higher. ALSO, note the current limited supply... well, if your current limited supply is actually indicating that it is limiting, that means there is a large power draw, and something is likely wired wrong, or the device is damaged. You shouldn't see much more than 200-300 mA on newer ones, and rarely over half an amp on bigger ones. Some older ones might draw a little more... but 200-300 mA is a good rule of thumb to start with.

Finally, you need to find your video signal. Many viewfinders have indicator LEDs to show when you are recording, etc. Some have serial data lines for on screen displays. Some have horizontal or vertical sync signals that allow an external OSD chip to overlay text not he video signal before it reaches the CRT driver. Sometimes there is a microphone, or switches. You can ignore any wires that lead directly to such parts, if they are obvious. The video signal will be one of the easier things to find... scratch the unknown wires against ground, one at a time... You should see the screen flicker a little bit when you do that. If that doesn't do it, you can always just try feeding each wire a video signal.

Mine was a Samsung unit. It had two connectors. One was for the mic, and I just removed it. The other had 7 wires. I found 2 grounds, one directly wired positive power connection, and 4 unknowns. Here's another tip... Video signals require a ground. They can share a ground with the entire circuit, but it you have two grounds on your connector, there is a highly likely chance the other ground is meant to be paired tot he video signal. Check an adjacent wire first. Sure enough, for mine, it was right next to the second ground!

A few more things... First, if you REALLY have a hard time getting a good signal, you may need to add either a 75 ohm resistor in series with the signal, or tied between the signal and ground. not all these units have 75 ohm termination built in.

Second, If the viewfinder was the type that reflected the video off a mirror, then without the mirror, the video will be backwards. You can fix this easily by flipping the horizontal deflection wires around. These are a pair of wires that go to a coil around the neck of the tube. There will be 4 wires. two for vertical,a dn 2 for horizontal. Fortunately, mine were marked. If not parked, you need to use an ohm meter to determine which wires are part of one coil. Wen you find one coil, swap it, and see if it was the horizontal. If it flips vertical, undo the change, and flip the other two. Do not cross the streams. Make sure the wires are front he same coil. If the wires, when removed front he circuit, do not conduct, they are from separate coils.

So, that's the trick to these things. They are dirt cheap, if you can find a thrift store that doesn't over charge. Some thrift stores haven't figured out that the video tape based SD camcorder is SO BEYOND obsolete, thanks to smartphone cameras... I have a place still asking $20 for them... I've bought them for as little as $4. A camcorder is fun to take apart! Not to mention, if you get the OLD ones... as in the over the shoulder type, from the early 80s, you sometimes find very nice gear motors that you absolutely must send to me! Trust me, it's for a good cause, and also insect like robots!

Now we need software that flips through a few frames stored in memory, like a little arduino gif. I hear there is a TV library, that absolutely violates all sanity on the poor little chips, and SOMEHOW makes it generate a somewhat usable video signal. Saw a japanese site that was doing 4 bit greyscale... More than enough for Gene and his coffee mug! I hope!

Edited July 13, 2015 by richfiles

[Sputnix]

Reverse engineered.

*snip*

I love that you don't assume any prior knowledge

Thank's so much for sharing this information. It's very useful (and I'm keen to try it on at least one lcd module I pulled from a camera).

The only difficulty will be navigating around the SMCs (surface mounted components).

Edited July 13, 2015 by Sputnix

[jotapeh]

This thread blows my mind. Incredible projects and some of the most helpful electronics hobby advice I've read in years.

[richfiles]

My FDAI arrived in the mail on Thursday, literally MINUTES before I was leaving town to go to a robotics meet up that I have not been to for 5 YEARS... On a side note... another member... an Engineer named Bob... Hmm... Musta traded careers with Bill... He moved out of state 8 years ago, and has been half way across the USA since then... He was back in state for family stuff... and randomly decided to show up again, after 8 YEARS... at the same meeting I showed up to after 5! Talk about chance encounters... Anyway, I edited my synchro post earlier in this thread... He shared some interesting information with me. I may have been mistaken, and had been confusing some elements of synchro operation with resolver operation. What he told me suggests emulating synchros may actually be MUCH easier than previously anticipated... I need to only generate one 400 Hz sine wave, and then simply calculate 9 multipliers to that reference, and output the reference and the 9 derivative outputs... No phase shifting, just amplitude shifting... Nice...

Anyway, I've not had time to do anything yet... but, you know... pics!

Look at my ball, my ball is amazing!

That ball turns smooth as butter! OHCRAPOHCRAP!!! Neilmon, you're getting mighty close to gimbal lock... Oh wait... Kerbals DID get an extra gimbal for Khristmas!

The reason a Kerbal Pilot can hold vectors and nodes and such, is the Scientists and Engineers ask what direction "Z" and "M" are. Jeb says "You're kidding, right?"

That Raytheon "chip in a can" has a date code of 1974! This puppy is 5 years older than ME!

Anyway, I'll have a rotary switch that has several positions. I'll be able to select prograde, retrograde, normal, anti-normal, radial, anti-radial, node, anti-node, target, and anti-target, and that particular vector will then appear on the ball by way of the deflecting crosshairs. Might do two rotaries... one to select mode (grade, normal, radial, node, target), and a second switch for pro, and anti/retro. I have 3 flags on this unit that read OFF, LOC, and GS. maybe i can use LOC to indicate a target is LOCked? GS... It stands for glide slope. Any other ideas? I can tie the OFF to the inverter/amplifier that powers the unit. If that's not on, then OFF will be displayed, and my AC bus will have no voltage. The inputs show signal inputs for the glideslope pointer, which is the arrow not he left of the unit, right above the GS flag. That's an analog deflection meter, same as the two heading crosshairs. I'm considering using that one for *atmosphere. It's unobtrusive, and the divisions I think would work out well. Not sure if there is an input for the "fat T" shaped bit at the bottom, and that ball at the bottom, directly above the afore mentioned "T" is literally a fluid filled ball level.

*I need the maneuver ΔV to display somewhere... I might use this for THAT purpose instead. Or I might just get another meter. Who knows...

Edited November 25, 2022 by richfiles

[Sputnix]

That ball turns smooth as butter! OHCRAPOHCRAP!!! Neilmon, you're getting mighty close to gimbal lock... Oh wait... Kerbals DID get an extra gimbal for Khristmas!

Haha. Love the reference

This is seriously amazing + cool.

I'm keeping an eye out for one too; If ever I manage to snag one, I'll be sure to get in contact with you (or just review this whole thread )

Edit:

I just saw how much velcro is on that FDAI. You may want to address that before you go to space

Edited July 18, 2015 by Sputnix

[richfiles]

Sputnix said:

Haha. Love the reference

This is seriously amazing + cool.

I'm keeping an eye out for one too; If ever I manage to snag one, I'll be sure to get in contact with you (or just review this whole thread )

Edit:

I just saw how much velcro is on that FDAI. You may want to address that before you go to space

I saw another one pop up on ebay. They ARE expensive... Just so you know... I've seen them sell for over a grand before, so the one listed is... not terrible...

http://www.ebay.com/itm/RARE-USN-USMC-A-4-Skyhawk-Attitude-Indicator-Gauge-Instrument-Type-AN-AJB-3-/371357712620

The good news, is you also see them sometimes pop up for the $150-300 + shipping mark. You just have to be on top of them and react fast.

Also... MAKE SURE it included the servo amplifier in the back. That's the little "backpack" thing attached to the rear of the unit. It needs that to function. It's probably possible to recreate that function as well... The circuit isn't complicated, but you'd have to reverse engineer the ENTIRE UNIT to do that... Not a simple task. It's better to pass on a real cheap unit and wait for one that has everything. If you find a sub $100 3 axis unit that's missing the backpack, it may actually work, but the electronics package may have failed. The controls are actually quite simple. You have servo motors and 3 synchro receivers. If you can figure out the pins to those, you could control the servo motors directly from an arduino, while reading the analog value of the synchros. You'll still need a 400 Hz sine wave for reference, but that's no problem. It's POSSIBLE, but more complicated. not that much more complicated, but you have to aim for a completely different direction of control (reading synchros instead of emulating the outputs of them). You also need to be aware that some units might be sold to collectors for show... and might not actually work at all.

Pay special attention to what you're getting... There are VERY MANY 2 axis units on ebay. Practically ALL the units with a blue sky are 2 axis units. Sad too, as it'd match Kerbal nicely. 2 axis units only have pitch and roll. No yaw. A lot of those types of attitude indicators actually have built in gyros too! Either and electrical power source or a vacuum line spins the gyro and the "ball" simply reacts to the local gyro... No electrical command signals at all to control those types of navballs. If you look inside a 2 axis navball, you'll find it's actually a nav-donut. If you want to fly a lot of planes, you can get away with one, but it'll be of limited use in space. Patience is a virtue. I've been looking for one of these since the beginning of the year.

The one I bought left me with under $20 left in the bank account for two straight paychecks (after bills). This was after also getting my renter's tax refund as well!!!

Also... The US tax system is idiotic... Tax you for stuff... then say, oh, I guess you can have some of this back, cause we asked for too much... How 'bout you have a system that actually gets it RIGHT.

Okay, enough ranting about that... Onto important things!

Everyone knows velcro is a space age fastener used to secure checklists and snack holders to your instrumentation panel!

See! it's legit! Velcro everywhere! Adds to the authenticity!

Edited November 25, 2022 by richfiles

[richfiles]

Hmm...

I'm considering, after finishing interfacing my real navball... Possibly seeing how hard it'd be to do a homemade navball, using off the shelf components. I used to work in manufacturing aerospace/defense motors, synchros, etc, so I'm very familiar with a lot of precision motor tech. I also have a complete understanding of the mechanical makeup of a navball. I even came up with a means to build one without the complex slip ring mechanism (you provide power with the left and right "wishbone" feeding power into the right and left side of the disc through very simple wipers. No date is sent this way, only power. You then have a phototransistor/LED pair at the pivot on either side to transmit data into and out of the ball. A few capacitors filter the DC bus voltage, and an arduino pro mini or something small like that sits inside the ball to communicate with the outside of the ball, and to run the two internal motors (for yaw and pitch). Externally, you have another small arduino drive the wishbone motor, optical communications to the ball's internal arduino, and the flight director/rate meters.

Alternately, you can actually get away with only one motor inside the ball but then you have to place the pitch motor not he side of the wishbone. This requires you counterbalance it on the other side, as well as add more side clearance to accommodate the added width. You still have to have one internal motor to handle the yaw. Probably best to keep both internal.

You can drive the motors using a low resolution encoder and a PID library for arduino. Not certain if you can run more than one PID loop on a single arduino though. They use interrupts to handle the encoder ticks. You also need to have a motor with a lower encoder rate, to prevent interrupts from interfering with the serial in. Serial between the external and internal arduinos should be handled with a single wire interface, so you can feed it via LED. Note, that the wishbone needs a second set of wipers at the roll pivot to transmit power, and a second set of LED/phototransistor sets for communication.

It really ought to be possible. I know there is a thread where someone was trying to do this, but I've seen no progress on that, sadly.

If I play is careful, it can be kept on the cheap. Cheap enough for anyone to put together.

[faeryty]

I've been thinking about how to kludge together my own navBall today too, although as I have no idea how the real one works (your explanations flew straight over my head) I'm starting from scratch. My first thoughts were comparing it to a mouse/trackball but my investigations soon brought me to this ball balancing robot. Turn it upside down and you've got a controllable ball with 3DOF.

I still need to work out feedback to compensate for any drift on the ball - my first thoughts are magnets and hall effect sensors, but I have to learn more about electromagnetism and I have no idea if this plan has any merit to it whatsoever, especially on the scale I'm planning.

I only had this idea today, and I'm sure it's full of fatal flaws. So before i buy a load of omni directional wheels and stepper motors tell me - why won't it work?

Edited July 23, 2015 by faeryty

[richfiles]

The Apollo navball diagram.

Here are my basic sketches of how a homemade navball might come together. This isn't complete, nor tested, nor set in stone.

The basic premise, is you have to transfer power INTO the ball. To do that, you use slip rings, or you can try to make it work with something like maybe spinning it on earphone plugs. You need to get data into the unit too. If you can get a 1 wire serial interface working, then you can use the third conductor of the earphone jack, or you can have a hollow tube and shine an LED through with a phototransistor at the other end to pick up the signals. The data sent is real simple, just 2 floating point numbers, in degrees, probably with a checksum, since the signal might not be very clean. You could even send the data three times in rapid succession and do a 2 of 3 selection for each bit, to try to filter out electrical noise.

 

All the motors need to be gearmotors. These are simply small motors with built in gearboxes at the end. The motors also all have an encoder attached. 1000 - 1200 count is probably more than plenty. That gives almost a 1/3 degree resolution, I think. Too high a count, and the arduino has a lot of trouble reading the counts if you use an interrupt based PID library (that's for speed control of a motor). What happens is that the motor is controlled just fine, but it misses incoming data on the serial. The yaw motor (My) has to have a shaft on both ends and is secured to the central disc. It spins the two half spheres to show yaw. The Pitch motor (Mp) tilts the central disc to show pitch. The motor is attached to the disc, and the shaft attaches to the wishbone. The other side freely spins against the wishbone. This free spinning side is actually a very good place for the slip ring or for the hollow optical tube, or for the headphone jack based slip ring.

If you go with optical transmission, you'll need to also add a small signal amplifier to power the LED on the wishbone. With slip rings, you don't need that, but you risk a noisier signal, since you have contacts rubbing against each other. The roll motor (Mr) needs a slip ring. If the motor shaft is hollow, you can feed your wires THROUGH it, and hook up a headphone jack slip ring to the other side of the motor. If the motor shaft is not hollow, the only way to transmit power is with a slip ring between the motor and the wishbone.

You can get a better slip ring signal by having multiple wipers. Wipers at the springy metal contact that touches the slip ring. There is more than one way to make a slip ring, but you'll just have to figure out what works. The apollo used a GIANT slip ring, and sent at least 12 signals into the ball! With a tiny arduino, you only need 2 power wires and one or two signals. 1 signal if you can utilize a 1 wire serial connection.

***NOTE***

Looked on ebay, slip rings can be had for UNDER $10 each, and you only need two. If you buy ones with extra conductors, you can double up wires for extra signal noise reduction, and even do full SPI serial interface... Looks like it'll be WAY easier than I though!

6 wire:

http://www.ebay.com/itm/12-5mm-300Rpm-6-Wires-CIRCUITS-2A-Capsule-Slip-Ring-240V-AC-for-Monitor-Robotic-/171247673771

and 12 wire:

http://www.ebay.com/itm/300Rpm-Capsule-Slip-Ring-12-Circuits-Wires-22mm-2A-AC240V-Test-Equipment-/281270446236

Two more things... If you can find a larger bearing of some kind to support the weight of the wishbone, it's better to turn the wishbone, by a gear, or even a rubber belt. That takes the weight of the ENTIRE assembly off the motor shaft, and lets you have room for the slip ring int he middle, with no problem.

A large, heavy duty bearing with a pipe through the center is actually a VERY good option. You can mount the slip ring tot eh end of the pipe, and feed the wires through the pipe to the wishbone. You have a cog or toothed pulley around the pipe, and that leads to your motor. The motor can have the same size gear/pulley, or it can be half the size and then you just run the motor twice the speed of the other ones. You COULD put the pitch motor outside the wishbone, and have the shaft attached to the central plate, as it looks like it'd be easier, but it really isn't. It's easier, AND more compact to have both motors inside the ball, and to drive them both with the arduino inside the ball, and just communicate via the slip rings. If the slip rings on ebay are reliable enough, I'd just say forget the optical thing all together. Good quality slip rings will have at least two wipers, and should transmit a fairly clean signal. I was thinking homemade slip rings would be likely used.

Also, have capacitors inside the navball, on the central plate. Attach them (in the correct polarity) to the (+) and (-) power supply. That lets the slip rings keep the capacitors well charged, and any time the motors draw extra power, they pull from the caps. You can also position the capacitors to try to balance out the weight of the pitch motor.

***EDIT***

Well... I took a close look at that schematic, and Apollo used resolvers, instead of synchros for the motor feedback... Well I'll be... I used to work with resolvers all the time! It's irrelevant to this project, since a custom home made one won't use them, nor does my synchro based F-4 Phandom FDAI/ADI... But it's just interesting, I think! I strongly suspect they had finer resolution of the ball, as synchros have a dead zone near the peak amplitudes, where the signal can be... imprecise... Only by a tiny fractional bit, but I suppose in space, that extra bit counts.

That's just interesting!

Also... some NSFK smexy-smexy topless pics of an Apollo FDAI!

Needless to say, a homemade one will not be such a... precision instrument, but it IS possible.

Edited November 25, 2022 by richfiles

[Antipaten]

If you're using a stereo 3.5 mm jack or 3 slip rings you can discard the optical communication gizmo completely.

The simplest form of 1-wire communication you can do is to connect TTL-serial to RX on the controller inside the ball using the third connector to get one-way communication. You don't need two way communication or error correction if you send the positioning data as a constant stream and filter out values that differs too much from what could amount to a reasonable rate of turn/roll. If there is a glitch in communication it will in worst case show like a small shudder of the ball and the position is corrected almost immediately at the next successful transmission.

[richfiles]

The 3.5mm jacks as slip rings was more or less an option for power transmission. Optional signal transmission too. Optical does solve the noise issue as parts wear...

Buuuut...

I was just going through my Epic Piles o' Crapâ„¢... And discovered I have many of the necessary parts to build the majority of a navball... Four or five times over!

If I bought some more slip rings (I already have a few of the 12 conductor slip rings, the wishbone yoke and wishbone already), and if I go with an external pitch drive motor... Then all I'd need to build home made navballs with precision gearing is... Wait for it...

Mount one motor (with encoder) to a disc (the ONE mechanical part I'd need to make, and it's just a static mount), Mount that disc to the hardware I already have, wire it up, and slap a half ball to each motor shaft on either side of the disc... That's it. The ball would be nearly 3 inches in diameter (7.35 cm). If you want a larger ball, you could also discard the existing wishbone and replace it with a wider part (it's just a flat piece of metal bent into a "C" shape). Granted... That's the simplified version!

Part pics are below...

This, however, would be an encoder based drive setup, using DC motors. FAR more EASILY controlled with an arduino, using standard libraries. This WILL be easier to make work than the real deal, like my ARU/11-A FDAI! Unfortunately for this concept, I'm still gonna focus all my work on getting the real deal working, as that will be BEYOND epic!

Still, I have MOST of the parts to build a few navballs! I'd probably have to order the motors, not just for inside the ball, but for pitch and roll too, cause I don't have any gearmotors that have shafts on both ends, and I would need that for yaw. I also would want motors with very low gearing ratios, so if a capsule is in a rapid spin, the ball can actually keep up. Still, most of the parts are actually here! I have matching gears that mesh up with these too. I just have to pop them off the old motors. The original gearmotors are geared with a higher ratio, and are way too slow. Besides, I plan to use them for robot leg motors! Interestingly enough, the encoders of this assembly actually fed their data through the slip rings, so the slip rings must be clean enough for fine digital signals. The encoders are, however, only 128 count. With the roll gearing, that's 256 counts per revolution. JUST barely tolerable for roll. I'd also consider looking for 256 or 512 count encoders, just to get better resolution, or else you won't even have one encoder transition per degree. A 1000 step encoder is probably best. It provides nearly 1/3 degree resolution, and will improve the jitters. More line fonts than that, and monitoring three of them on one arduino might get hectic. I don't really know what an arduino can handle, in terms of encoder speeds and line counts. I'd also say you NEED encoders with an index. These encoders do NOT have an index. Can not reiterate enough, this is NOT optional for a position seeking device. These exiting encoders will NOT work. When you power up, all the axes have to do a spin until they hit the index, so the arduino can get a position lock on the ball. It can then count steps from that to know where the ball is, and can double check the count and reset it if necessary any time it passes an index again. Since the slip rings seem to be smooth enough to pass the encoder signals, that actually simplifies things, cause it means I only need a single arduino to control it, and don't need one inside the ball. Just need to feed 2 wires to the DC motor, and a minimum of 5 wires from the encoder. the pitch and yaw encoders can share power wires, so... Motor-Yaw x2, Motor-Pitch x2, Encoder Power x2, Pitch Encoder x3, Yaw Encoder x3. That's exactly 12 wires, exactly what I have for the yoke slip ring.

Here are closeups of the assembly.

If anyone is interested in getting a jump at some part sets to experiment with (the MAIN thing is the slip rings and the yoke), PM me. I'd consider letting some of these go. You'd need a second slip ring, new encoders (with a better count, plus index), but both of those things are pretty cheap on ebay. The slip rings are just under about $10 for a 12 conductor part. Stand alone encoders are about $10-20 on ebay for 300, 360, 400, or 600 count per rotation units, though getting a motor with a built in encoder is probably better. You can also go with a much LOWER count if on the motor side of a gearmotor, vs the gear side, since the motor (thus the encoder) will turn multiple times per one output gear revolution. If you wanna experiment yourself, you'd be doing just that... Any additional parts you gotta make or buy yourself, and all code is on you... But yeah, you you want to play with a yoke and wishbone gear assembly, PM me and let me know. I only have like 6 of these things, though one is an un-disassembled unit that I'm keeping for use for a robot someday, maybe. These were originally security camera gimbals, meant for those cameras that sit under a plastic or glass dome. They were high end units originally, using Northrop Grumman slip rings and Escap motors. No motors would be included though... As I said, i have plans for these motors, AND they are unsuitable, as they move too slowly. I'd say a low gear ratio, like 10:1, maybe no more than 20:1 would give good, speedy motion. It's a balance between torque and top speed. Those numbers are just off the cuff guesses.

Truth is, I had the idea, but I've also got my real FDAI, so I may find it hard to motivate myself to make something I already have, and my priority is making the interface for the real deal. If I do ever get around to doing this on my own, it could actually be quite a while.

Edited November 25, 2022 by richfiles

[richfiles]

Oh. My...

So these are the Apollo toggle switches:

Guess what I found...

These are what some epic part number searches and some boring, photo-less digging around found. At least I got a nice photo at the end of my search!

These are made by Honeywell.

http://sensing.honeywell.com/product-page?pr_id=35890

Sadly, these are also not what you would call cheap...

I think a few people have posted small kerbal kontrollers that have costed less than ONE of these switches!

The reason for the cost, is these are entirely sealed, and rated to military specifications.

That translates into $$$.$$, and yes... Some of the switches actually use all those $ places in their actual price!

It's an unrealistic goal, but hey, maybe China will step up to the plate and make a cheap look alike someday. Honeywell's site refers to this style as a "Tab Lever" toggle. If that helps people in their search for more authentic Kapollo parts, then awesome!

For the time being, however, China does sell these:

http://www.ebay.com/itm/10-Pcs-12VDC-20A-SPST-ON-ON-12mm-Mounting-Thread-Dia-Black-Toggle-Switch/130901360723

They are a FAR CRY from the real deal, but they are close enough in size and throw, and the general-ish shape, that these are what I'm settling on for my toggles. Went and ordered 20 of them for $14-15 USD, shipped.

For those looking for a slightly better cheap analog, but are willing to spend a little more for a MUCH better shaped lever, this is an option:

http://www.ebay.com/itm/54100-01-25A-Series-On-Off-Toggle-Switch-/171867822485

http://www.digikey.com/product-detail/en/GTS447C301HR/CW161-ND/2349756 <--And a similar style from Digikey.

That'll be it, for today's episode of NASA uses REALLY FREAKING EXPENSIVE AND HARD TO FIND PARTS!!!

Just a thought...

Hmm... I wonder how hard it would be to make molds that you could dab some grey epoxy into, and then just "smoosh" onto either side of a toggle lever... Basically MAKE the tab lever style switch. Grey epoxy, I KNOW exists, and would hold color through wear and scratching. It'd be something fairly available, and doesn't require special equipment, like injection molding of plastic would... Hmmm...

Edited November 25, 2022 by richfiles

[Sputnix]

Oh. My...

Hmm... I wonder how hard it would be to make molds that you could dab some grey epoxy into, and then just "smoosh" onto either side of a toggle lever... Basically MAKE the tab lever style switch. Grey epoxy, I KNOW exists, and would hold color through wear and scratching. It'd be something fairly available, and doesn't require special equipment, like injection molding of plastic would... Hmmm...

The switches I used had a slight taper - which is why I chose them; as a sort of replica

Could you look into 3D printing? I've actually designed and printed toggle-switch guards (similar to what you see on shuttles).

Of course, this would require you to have access to a 3D printer

Though maker-spaces are increases, so there could be one around you somewhere.

[richfiles]

Closest maker space is a 2.5 hour drive for me...

It's all what I can afford/fit in my apartment, or what I can safely leave at my dad's farm, in the unheated barn/shop (no animals, just a barn converted into a shop/garage... But it get COLD here in winter). I tend to do woodwork at my dad's place. I can weld there too. Considering that welding IS an option for me though, I want to find some appropriately sized bits of angle and dowel steel to try to recreate the recessed switch grouping with the switch guards that were on apollo. u-bolts are great, but I'm really curious to try for something a bit closer to the real deal. It's more of an experiment is seeing if i can actually make something reasonably usable. It's not a huge deal though, if it doesn't work. If it proves too troublesome, I'll just get u-bolts and call it good enough.

Actually... U-NAILS... not bolts... Those come to mind as roughly being the size and shape I may be looking for... straighten the ends with a plier and vise, so they all are nicely lined up, then grind the points off, and tack weld them into place in the rails. The u-nails should be cheaper than u-bolts as well, I would think! Kouston, we've had a solution!

These are typically sold as "fencing nails"

Note the recess to the switches, and the shape of the switch guards.

AAAND...

For those who built wooden consoles... Get some nice long ones, and tap, tap, tap with the hammer, and you're good to go, just paint!

Edited November 25, 2022 by richfiles

[Sputnix]

Yeah, I noticed that 'in the days of ye olde' they were basically welded loops --

Some guy was selling a 4x on ebay a while back - unused apollo. Just steel, basically.

I like the idea of the uNails. Seen the Ubolt idea, but this is definitely a nice, cheaper alternative.

If you got access to a welder, and are neat enough with it, you could go all out and do this

[Cost of the materials might be a bit high though, so sticking to ply + uNails will probably be the cheapest option ]

[richfiles]

Yeah, I have access to a welder. Even if I am no good at it (it's been YEARS since I touched it), my father has been welding for most of his life. I've seen him do everything from cut hole patterns out using nothing more than an arc welder, to him putting together the craziest of thin stuff together. This won't even be difficult to do, I think. It looks like all it actually is, is a thin sheet of steel, two long straight steel dowels, and the u-nails. It'll basically be a tack weld between the dowel and the u-nails, then the metal sheet is laid out and tacked not he outside the edges to the dowels, I would imagine. At least, that makes the most sense, construction wise, to me.

Aaaaand... That's EXACTLY how they were made! This will be INCREDIBLY easy. Heck, I could make a TON of these things in one go!

The reason the switch guards and switches are recessed: The panels with the lettering are thick plastic panels that have electroluminescent backlighting. It's the thickness of the panels that creates the recess.

Edited November 25, 2022 by richfiles

[Sputnix]

The reason the switch guards and switches are recessed: The panels with the lettering are thick plastic panels that have electroluminescent backlighting. It's the thickness of the panels that creates the recess.

That's nifty to learn!

I always thought they were just 'recessed' because of the panelling (not realising it was thick plastic, and had EL cable behind) and that it was another way of minimising accidentally flipping switches.

I love the stuff you learn on these forums

[richfiles]

Yeah, the EL lighting was used because it was lighter than incandescent, or even fluorescent light bulbs. LEDs were not available back then. They existed, but only really as pin-point dim lab curiosities.

So, I went to the hardware store. Most of those u-nails are generally referred to as "fencing nails", but that particular type is also much narrower at the bend than at the points. Not ideal for our application. What I did find that appeared to be ideal, were the u-nails from gate/door hook latches. They are a perfect "U" shape, and the right size. If you find a set that uses them instead of eyelet screws, then you'll get two per hook set. They were under $2 per set, and the store I went to had over a dozen in stock... I didn't buy them, however.

What I bought were these:

These are used to suspend pipes from above. I think they are a little bigger than the official Apollo switch guards, but the look is definitely right. I might still go back and buy some of the gate hook latches, and try those, but these, I like them. They also came in 5 packs, and they had 4 in stock. That gives me 20 "U" shaped parts to work with, till they restock.

Bonus points... Since they are so long, you get the long horizontal bars for free!

Also, on a side note, I found this AMAZING CRT:

It's 8 x 3.5 inches (19.6 x 8.6 cm), and I want to DO SOMETHING with it SOOOOOO BAD!!!

It's currently residing in a fully functional Brother WP-2450DS "Portable" word processor. I was thinking of issuing it an eviction notice.

* Unfortunately, it doesn't use any processor I'm familiar with. I'm hoping it's just a standard NTSC type composite signal, but it might not be, considering the odd aspect ratio. I'm gonna have to TOTALLY reverse engineer this one. Man, I don't even know where to put this! I'm thinking, if I do use it for ANYTHING, it'd end up overhead.

* (It uses an enhanced Z80 CPU!)

Edited November 25, 2022 by richfiles