[WIP] The REAL Nav Ball Project Thread

[Patupi]

I suspected you'd already considered it, but just wanted to check. Glad to see you're making progress. I am a CNC machinst/programmer so I understand how much a degree off can be! Usually even sloppy tolerance on angles is maximum of half a degree. I also was a clock maker/repairer a while back and got to know about gear meshing and running on centers (or pressure after center would be more accurate I guess) so they don't grind the gears. Doesn't take much to begin to wear everything so I hope you get that sorted out. It's not just the angular drift!

I'll be interested to see how things develop with this and to see how you implement the tags like prograde, retrograde and nodes. Colored cross wires like in the apollo one? Would definitely work, but might be awkward when they cross each other. One needs to be thicker, but shorter, so both are visible... etc etc.

[NeoMorph]

The needle system is pretty cool... and sooooooooo darn easy to implement. All you need is an Arduino plus a little Adafruit 16 channel servo board (each needle uses a single servo).

For example, look at the original...

As you can see, there are two servo motors on each of the three sides which has a wire that leads to the indicator needle painted on the front of the display. One needle shows the rate of change (ie the speed of the axis - it's the outer indicators on the above FDAI picture) and the other needle is drift from a set angle (the inner needles on the FDAI pictures). So when you set the node up in KSP the needles show which direction you will need to go to get to the correct attitude. They are actually BETTER than the KSP icons in fact because that system doesn't tell you which way you have to turn. Cool thing is all you have to do is send the signal to the servo and it will deflect the needle.

The only hard part is initially calibrating the needles. Actually, not even hard... just time consuming to make sure the deflection isn't too much that you bend the needle. I've made some jigs so that I can hot bend the wire without stressing it and making them good duplicates so they match the three sides.

Another thing that is different with the needle system is that you can set the needle's scale factor. This allows you to use 2x, 1x or 0.5x the true value to narrow in on the correct attitude. You would start off at 2x so that you know which way to go but when you are getting near to the correct attitude you would use 1x and then 0.5x to make it more sensitive. I'm planning on using this in my system. At the moment it's not part of the current prototype but should be easy enough to set up once everything else is working.

The weirdest thing about designing this nav ball is that there are SOOOOOO many different versions for the Apollo program. For example the version in the picture above has white needles but in other versions there are yellow needles. There are soooo many different marking layouts for the ball that I don't even know which one to choose from. Some are from training rigs, others from previous versions so it makes kind of sense. Then they used the same system on the shuttle too (with even MORE mods). I just have to choose one I guess... after all this is Kerbal and not Human.

In fact out of everything in my simulator build there is only one thing that I have absolutely ZERO idea of how to make.... and that is the tape altimeters and speed indicators (see image below). I could easily make the tape and use stepper motors to move it up and down but the big problem is... how would I do the markings. They would have to be printed onto the tape somehow. Oh well, one problem at a time.

Left hand tape is the radar altitude and the right hand tape is the rate of descent/ascent.

[tom1499]

Yay for pretty pictures! Um, wow. Just looked at tape altimeters and ,JEEZ, thats some amount of gears. Not a bother though if you have modern motors and Micro-controllers. Have you considered just going with the roller-disc-thing ones? Depending on the range and number size it might be quite big, but may be easier to make and less fragile?

My god those navballs are sexy. Someone actually picked one up at an auction, what a lucky dude. Looking at your earlier picture, this might be a fun little project for me to investigate, any comments on that? Just looking for any info on the Gemini FDAI equivalent. Barely anything on it. All I know is it had 4 gimbals to avoid gimbal-lock and that it was used in Gemini.

[automcdonough]

very cool

[automcdonough]

good idea with the ebay. some tinkering around getting some of these patched into that servo board might be a nice shortcut. and they are the real deal.

http://www.ebay.com/sch/Aviation-Parts-/26435/i.html?_vxp=mtr&_nkw=&_from=R40&_dcat=26435&Part%2520Type=Avionics

http://www.ebay.com/sch/i.html?_odkw=avoinics&_osacat=0&_from=R40&_trksid=p2045573.m570.l2632.R2.TR12.TRC2.A0.Xaltime&_nkw=altimeter&_sacat=26435

http://www.ebay.com/sch/Aviation-Parts-/26435/i.html?_nkw=vertical+speed+indicator&_frs=1

[Patupi]

In fact out of everything in my simulator build there is only one thing that I have absolutely ZERO idea of how to make.... and that is the tape altimeters and speed indicators (see image below). I could easily make the tape and use stepper motors to move it up and down but the big problem is... how would I do the markings. They would have to be printed onto the tape somehow. Oh well, one problem at a time.

It certainly looks cool. I'm sure you could find someone to make the tape for you if you can't yourself, but what would be cool is if the tape were two layers, one transparent (or rather translucent), one not. Make holes in the one that isn't and then illuminate from behind with an LED stack and a diffuser. In fact if there is any way for the ball to do that, illuminate the grid markings and numbers on it via LEDs from inside... very cool. Rather tricky to get even lighting inside though. A series of segments for the diffuser? Multiple LEDs that go around with the ball? Then you have the problem of electrical connections to the outside when it has no part of the ball that can't be seen from the front. Induction? Maybe way too much to handle.

... of course they could be run by batteries... but still would be a pain to replace them!

[Javster]

Well, I wish you all the best in this endeavor.

[NeoMorph]

God... I'm having sooooooo many problems with this build. Half the problem is not having the right tools. I know WHAT to do... just don't have the tools to do it accurately enough.

Yup, the real problem is accuracy in my fabbing. I'm currently saving up for a small CNC machine which I've been yearning for for years. Once I get one of those it will be dead easy. My latest version works but there is a problem with the pitch gears again. The fact I ordered two gears that were listed as having 10mm bore holes arrived looking great but when I went to fit them to my new 10mm shaft I found... that the holes were actually 8mm. I turns out I missed one sentence in the description...

"Bore size is subject to change from table below, if bore size is critical then select 'Custom Bore Size' with bore required when making purchase."

To get them to drill the correct size that the table shows clearly is 10mm... (I ordered the 25 tooth version)... THEY WANT TO CHARGE ME £2.75 EXTRA FOR A GEAR THAT COSTS £5.76.

Here is the chart of the gears... The bore hole size is D1...

http://www.technobotsonline.com/gears/mod-1.0-steel-spur-gears.html

So yeah... Kind of peed off with the whole project at the moment. I'll get it done but will get easier when I have the right tools.

[automcdonough]

aw yeah that is tricky. a CNC costs a lot more than a nice set of drills!

I've made gears by hand before, they were good enough for the low-speed application I made them for. Printed them 1:1 on a sticker and laid it flat on the metal then got to work with a saw and files. It sounds tedious (and it was) but like most of the things I've made by hand, took way less time than trying to get a 3D printer to do it right. (and in that case I wanted metal..)

[Patupi]

Using a center drill with a pillar drill would probably work if you're careful. It should self-center if the center drill tip is close to the 8mm dia. Not sure if you can get a custom center drill with 8mm tip and 10mm shank though. Usually they are more disparate sizes than that. In theory you could even do it with a pistol drill... but you'd need a very steady hand and a good eye to keep the drill perpendicular to the gear.

[Mulbin]

Any update on this? A navball is way down my list of things to make for my simpit build... of course if you made a few to sell...

[HoustonDave]

You, sir, are a minor demi-diety for pursuing this. How awesomely groovy.

[NeoMorph]

Just looking for any info on the Gemini FDAI equivalent. Barely anything on it. All I know is it had 4 gimbals to avoid gimbal-lock and that it was used in Gemini.

The 8-Ball in the Gemini was pretty much the same as the Apollo one - but it was connected to a 4 gimbals IMU system rather than 3 in the Apollo. It's the same as saying that a Rev counter is an engine... The Rev counter is just indicating how many revs the engine is doing. In the spacecraft, the 8-Ball is the rev counter and the IMU is the engine.

The Gemini system was way better than the Apollo one as it never got Gimball Lock (the red areas of the ball on the Apollo version).

A Fourth Gimbal for Christmas

About two hours after the Apollo 11 landing, Command Module Pilot Mike Collins had the following conversation with CapCom Owen Garriott.

104:59:35 Garriott: Columbia, Houston. We noticed you are maneuvering very close to gimbal lock. I suggest you move back away. Over.

104:59:43 Collins: Yeah. I am going around it, doing a CMC Auto maneuver to the Pad values of roll 270, pitch 101, yaw 45.

104:59:52 Garriott: Roger, Columbia. (Long Pause)

105:00:30 Collins: (Faint, joking) How about sending me a fourth gimbal for Christmas.

(Armstrong - "This is Mike at his best. We had a four-gimbal platform on Gemini.")

Edited March 28, 2014 by NeoMorph

[NeoMorph]

Any update on this? A navball is way down my list of things to make for my simpit build... of course if you made a few to sell...

Been working on the electronics and coding for the past few weeks... will probably have some vids to show in a couple of weeks... if all goes well that is.

Good news was that a problem I had been seeing over the last couple of weeks turned out to be a wire that had melted, stretched and broke... but the insulation looked fine still. Trying to trace the system wasn't working because when I move the board around the contacts reconnected... and the tests went through fine. Put it all down again and... wouldn't work again. Was driving me nuts.

Was only when I said "Arrrgh... Restart from scratch" and I pulled all the wires out of the breadboard that the one signal wire pulled and snapped really easily. Plugged it all back in and it is working again.

I'm using a 20x4 as a test readout system to track the values that the navball will have to replicate. But the system will have to handle things like roll rates, rate multipliers and the such so its having to have a lot of extra alpha code in there keeping track of everything and output it to the display.

[AntiMatter001]

i'm really excited to see this and i would just like to ask a few questions in and around this topic.

1) why on earth did you start this (just curious)

2)where are you gonna put it afterwards?

3) anymore projects (like eggbot) are you or have you done.

4) where can i sign up? (joking i'd probably be more of a hindrance than a help)

love what you're doing and i think my dad (robotics/electronics nerd) would love to see how your doing it

[NeoMorph]

1. I love making stuff (especially simulators)

2. Probalby put it in my walk in cupboard

3. I'm going to make my own Eggbot chassis but using the Ebb electronics and software.

4. Not really sure what other parts to do atm. Have been off the last week due to a bug.

I still haven't ordered my CNC machine because I want to clear some space for it first and the weather hasn't been brilliant.

[cyberKerb]

Just letting you know that I'm also following this project and find your process great to read. It's one of those thing that I'll pretty much always check in on as it's an exciting project for someone to do

Also just saying that for sharing your progress so far and looking forward to any updates in the future.

[AmeliaEatyaheart]

This is an amazing project, been loving watching it develop

I have to ask though - how are you getting your yaw/pitch/roll values out of KSP? I'd love to make something like this (if I can) - I'm good with arduinos, but no clue on how to connect to the computer for the data!

[jerome.tremblay]

I have been wondering about the spinning disk inside the ball. I don't think there should be one in a simulated navball. In a ship, you want the ball to remain stable and rotate the enclosure around it as the ship moves. In a simulated navball the situation is inversed, the ball must move and the enclosure remains fixed. The LAST thing you want in that case is a working gyroscope.

Edited November 21, 2014 by jerome.tremblay

[richfiles]

Have you ever considered another option for the ball... Make the ball with a pattern of north and south oriented magnets inside, and place a set of electromagnets around the perimeter of the ball. The ball would not be attached to anything at all, just spinning on some ball casters or possibly fluid suspended. Your control device would operate the electromagnets to create the opposing poles of the magnet pattern across the electromagnets, much int he way a sinusoidal brushless motor pushes and pulls on a permanent magnet rotor by creating a magnetic field that rotates around the stator. creating a hemispherical magnetic field that pushes and pulls the ball around was beyond the tech back then, but I can't imagine it wouldn't be possible to create a patterned magnetic field with, say a hexagonal array of coils covering the rear of the ball (like the retina of an eye), and three caster balls to hold the ball stable against the rear of the housing.

I'm thinking that if you get creative, it might be possible to create a pattern of magnets that could create a self calibrating start up sequence. Possibly an expanding spiral of north south regions in the ball, so if you start the field confined and then expand the field int eh opposing spiral, presumably, it'd drive the ball until it spins itself tot he aligned unique magnetic pairing with the controller generated field. Rotate the field, and you roll the ball, translate in one axis, and you get yaw, translate the field pattern in the other axis, and you get pitch.

i realize it'd be a start from scratch effort, but it'd probably be the cheapest possible ball design possible. just a bunch of electromagnet coils in a grid, and a bunch of permanent magnets glued to the inside of a two piece ball.

Presumably, if the magnetic pattern would not be a completely uniform, repeating pattern, and the field that is generated itself becomes the aligning factor through a sequence of generated fields designed to take advantage of that non uniformity to drive the ball into known predictable paths from an unknown start point. Another option available is to have a very uniform magnet pattern, which would allow for (I think) an easier software control if the fields, but have a CCD on the back of the ball image the ball to determine where it's at. That'd require image processing though, so that more or less negates the easier control pattern... I still theorize that a magnetic pole pattern could be created that forms a self aligning startup sequence, and maintains ball position based on the rotation or translation of a polarity field pattern over the electromagnets that matches the unique, non-uniform magnetic pattern of the ball. Yet another possibility is to place some hall effect sensors to detect the pattern of the magnets passing below them. Finally, the use of a laser barcode scanner MIGHT be able to be adapted to recognize the spacing of the graduations of the ball along different scanning axes, so that it can determine the start point of the ball.

I myself am considering getting an old arcade style joystick (two I suppose), and a few pushbuttons and toggles that have safety flip top covers and making a simple keyboard replacement controller. It would literally be a gutted USB keyboard, with switches hard wired to the key matrix to recreate the keyboard keys in the appropriate switches... Not exactly rocket surgery... LOL

But man, I so want a nav ball in it now! XD

[richfiles]

And I found a paper on spherical motors.

http://www.jhu.edu/news/home01/jan01/pdf/mechatronics00.pdf

The image is just one concept of a spherical motor. It's a very large variant, but the general concept is definitely there... Eliminate ALL the complex mechanicals by simply turning the ball itself into a static, free moving permanent magnet multi axis rotor, and having a ring or grid of electromagnets to manipulate it. I mean... LOOK at that ball! It's ALMOST the 8-Ball / navball already!

**EDIT**

Hmm... The link/images are gone. Oh well...

Edited July 18, 2015 by richfiles

[RobotsAndSpaceships]

This looks pretty cool.

[Rebreok]

Absolutely fascinating work you have here. I am in day 1 of doing the research to build a full simulation of a KSP cockpit, or at least the controls, and stumbled on this thread. Tell me, how much, if at all, did you consider a purely digital recreation? My first idea would be to simply project the KSP image from the navbal to a spherical screen of some sort. I actually found an interesting video of a device that looks like it could be recreated out of an arduino, some LED's and a stepper motor:

while not a tenth of the challenge you have undertaken, I hope you find it interesting. I will continue to monitor this thread as I work on my project as well. Good luck!

[gudenau]

Have you made any progress?

[cyberKerb]

Another option here:

http://eclecti.cc/computergraphics/snow-globe-part-one-cheap-diy-spherical-projection