View Full Version : bendix king HSI ki525 working with Arduino and Prepar3d

12-02-2016, 11:06 AM
Finally, I manage to get this real HSI fully working with prepar3d.



I have swap the two internal resolvers with absolute encoders.
Each sensor (absolute encoders, photoresistors, is read with Arduino and them move dc motor, d'arsonvals, electromagnet....

-read absolute encoder
-compare with prepar3d heading
-move dc motor.

HDG flag:
-read simulator main bus voltaje
-If enough voltaje, then move d'ársonval

NAV flag:
-read simulator radio status.
-If nav1 with valid vor or ils, then d'arsonval

To/from flag:
-read status in simulator.
-D'arsonval. Positive, Negative, or off.

CRS Needle:
-read crs absolute encoder.
-Compare with vor/ils radial.
-Calculate deviation and move D'arsonval.
It is posible to be done reading the direct value of deviation in the simulator. But simulator only uses INT number to CRS set. So the needle move is not smooth. Very better result reading and calculating the exact deviation with FLOAT numbers directly in arduino. It can be seen in the vid. Much more smooth the real hsi than the simulator one.

HDG bug:
-It is read with a led light and two photoresistors. Their values change between -30º to +30º from lubber line.
-I have found a equation to know his exact position (-30 to 30º) with those values.
-It is NOT need to know the position out of this range, because it is used with auto pilot and only needs the value in this range to smooth heading.
This is why is not always synchro with simulator. But works perfect.

Glide Slope:
The most dificult one.
The exact position is known like the HDG bug. Reading two photoresistors than see a led light. When needle moves, one photoresistor gets more light and the other one less light. With and equation, I manage to know its exact position.
The needle move is made with an electromagnet. The equations to rise or low the voltaje of electromagnet are quite difiticult. But finally works perfect.

RMI and Whisky compass seen in the vid are very easy comparing to hsi.

Thanks for reading!.

12-03-2016, 08:20 PM
Awesome work! I have the same HSI and have sifted through the maintenance manual a bit, but have not started working on the interface yet. Are you using Hall encoders? I've used Halls to piggy-back onto some CX synchros, they work perfectly.

I actually figured out how to read resolvers recently, so I'm thinking I'll be able to interface the HSI without modification...we'll see. What's the story on the compass? Bar magnet underneath, driven by a stepper?


12-04-2016, 01:10 PM
Hi Matt. Im Arturo, from youtube!

In compass and Crs select I have used absolute encoders.
Printing 3d pieces to fit the shape of the original synchros.

Like this in VOR:




One of them wasn´t so easy. The shaft of one of the the synchros is very long, and I had to use a coupler and ball bearings. But this is the idea.

And for the compass I have use the same motor was installed. In my case, a DC motor.
I think the original real HSI use a stepper motor, (as i have seen in the mantenience manual) but was change to a DC motor for the ast300 sim.
It is the same. Both of them can do the job.

12-04-2016, 03:44 PM
Oh, hi Arturo. :) I really like the plastic adapter, that would be a fabulous excuse for me to get a 3D printer...or a lathe, I want a lathe too! Got a data sheet on those encoders? Are they Hall-based? The Hall chip I've been using requires that I attach a magnet to something, which is perfect for synchros, but not always convenient for other tasks (but man, they work amazingly well). The guy I'm doing the VOR for picked up one of the self-contained Hall encoders and used it to replace something (a resolver, I think) in one of his instruments, looks like it is about the size of many synchros and resolvers.

You mention "original synchros", but those look like resolvers, no? Do they have a black and a red wire, then four wires together (red, black, blue, yellow)? Speaking of resolvers, if you're not going to use them for anything, I wouldn't mind taking them off your hands, though I suspect the shipping from Spain to US might be prohibitive. I'm seriously considering using resolvers even for real instrument conversions (for instance, converting real pitot-static instruments to sim use).

I started back this weekend to re-building the electrical system on my sim, but now I may have to take a break and work on that HSI earlier than expected. Your progress has got me excited. :)


12-04-2016, 04:28 PM
Resolvers... (I call everything synchro... :roll:).

Here it is the datasheet of the absolute encoders Im using:

Living in US, I would consider those ones. Which I like a lot. They are SMALL

Better the pwm version than analog.
10 bits resolution is perfect (1024 steps by turn).

I had a little problems at first with some noise, jitter, readings.
I use an array to smooth reading. Works perfect.

My resolvers will not leave my home ;) You never know when you need them.

I like the pwm absolute encoders solution because I only need one pin in arduino to know the exact position. And it does not have to be the analog ones. One digital pin, thats all.

I have working at this moment HSI, VSI, RMI (lot of pins for this one), Whisky compass and Turn coordinator with only one arduino mega. There are free pins for somthing else, but the rest of gauges (ASI, Attitude, Vor2, Altimeter) will go with a second arduino.

A third one for switches, leds.. etc.

Come on! Lets go for this HSI! I want to see it!

And... I have to work with my attitude indicator.. ;) earlier than expected..

05-09-2017, 01:59 AM
Well I have tackled the HSI. Everything looked fine until I attempted to interface the resolvers. These particular units are designed to operate on MUCH lower voltages than the other resolvers I've worked with (the course resolver calls for about 0.2 volts...yes, POINT TWO volts). My resolver interface is working (theory of operation is the same), but I'm squaring off the signals in the sine wave from the stators...I think. I really need an oscilloscope to sort this out. I'm experimenting with putting some resistors in series with the reference signal driving the rotor, but then I get asymmetrical outputs from the stators...how odd!

At this point, there isn't all that much gratification in getting the resolvers going anyway, since I've already figured out the basic theory. The main advantage would be that I don't have to hack up the HSI. But I may end up just getting some Hall encoders, or some units like yours, and yanking out these low-voltage resolvers.

The glideslope flags looked like an interesting project, then I accidentally hit the little LED with 5vDC....ooops! It's glued in there pretty good, too.....this is one situation where I think a small servo may be just the thing. :)


05-09-2017, 03:21 AM
Hi matt!

I am glad to know about your HSI progress!

RESOLVERS: I remember you use Arduinos, isnt it? I suppose you know, but in Arduino you can use the voltage reference that you want (from 0v to 5v). You only have to use the command analog Reference (EXTERNAL) and feed the AREF pin with 0.2v. In this way you shouldn't have any problem with those resolvers.

GLIDE SLOPE: The most "funny" part :) In my unit, the glide slope led didn't work. I had to change it. It is glued, but with a little drill I wear the old led, and put and glued a new one. It is a 3mm led. It is easy to change it. Try it! you don't have anything to loose. If you change it, yoy can try to use the internal electronics. Im not good enough to undestarnd this thing, so I interfaced each sensor direct with arduino:
-the 3mm led
-in the opposite side, two photoresistor. Reading and comparing both you can know where is exactly the needle.
-electromagnet. You put more or less voltage (I use L298n motor driver) to get the needle going up and down. But its behaviour is not exactly the same like a D'arsonval (one position for each diferent voltage). It is quite more complicated (and funny).

HDG BUG: how are you reading it? There is no resolver for this one. It uses the same led and photoresistors way. Comparing photoresistors readings, you know how many degrees is clockwise or counterclockwise from the lubber line. (with a maximun +-30 degrees).


05-09-2017, 11:50 PM
Hi again Arturo,

I forgot about the AREF, thanks for the tip! I'm not sure if this is the solution or not, I think one of my problems is that the little rotor behaves oddly with so much voltage getting thrown to it. But I may try voltage dividers to drive the rotor (instead of resistors in series), and putting something like 0.3v to the AREF pin.

Maybe my GS LED was already toast, like yours. :) I too will not bother with the internal electronics, even if I end up replacing the LED. One difference between little general aviation instruments and all the 400hz "big iron" gear is that many of our little GA instruments are designed to work directly with very low voltage, analog waveforms and signals, because they are being driven directly by NAV radios, remote compass units, etc. There's really no point in trying to mimic all these signals when you can just hook directly into the electro-mechanical devices inside the instrument.

On the HDG bug, mine is oddball. I looked at it only briefly, and it appears to be going to a synchro. I didn't dig into it much further, because I have absolutely no use for it (no autopilot, and no flight director). It will serve only as a reference for the pilot. :)