Menu Sign In Contact FAQ
Banner
Welcome to our forums

Mysteries of ARINC407 and ARINC429

From here

The MD26 is indeed junk. The output is almost a square wave. It also runs really hot and frequently packs up.

If you have one “illegal” converter for sell- I’m interested :)

It would be a straightforward project, to take in ARINC429 heading and generate two sinewaves with a phase shift, relative to a 3rd sinewave (input; I am not sure how these systems work since you need only two to deliver a bearing, although the "classic synchro uses 3 wires) but I wonder if there would be any demand. I am pretty sure I have seen these boxes advertised. Perhaps not cheap; none of these oddball converters are cheap. Anything “arinc429” is expensive anyway. Those HI3593 chips are close to $100 each which is outrageous, and ensures that any product using it has to sell for at least a few hundred, but they are very convenient.

Shadin AIS450 at a mere USD 4200 Is there a market these days, at say 1k or so?

Administrator
Shoreham EGKA, United Kingdom

That’s really funny; clicking on the BUY link there takes you to an MD26 page, but it says “pure sinewave” so it looks like the MD26 has been quietly updated

Maybe they used the same circuit as the SG102 uses.

Administrator
Shoreham EGKA, United Kingdom

I am now successfully decoding the ex-SG102 label 320 heading into degrees:

So generating two sinewaves out of this is… just some trig, and two D-A converters

I am trying to work out the typical XYZ wiring. AFAICT you accept a 400Hz reference signal from outside (and in some installations this comes from a powerful source e.g. an MD26) and you generate two sinewaves relative to that.

The phase shift between the two sinewaves is the heading.

But I now see a wiring diagram above which appears to use just one of the two sinewaves to deliver the heading, which implies one of the two must be phase-locked to the reference, so you can use ref+sine1 or sine1+sine2 equivalently.

Looking at some circuit diagrams of old King gear, the sine1 and sine2 inputs present very small loads – of the order of a few k. But the ref input can be a big load, hence the power available from sources like the MD26 or an SG102.

The sine1 and sine2 signals also swing only between -5V and +5V. The reference needs to swing a lot more – 24V RMS or so.

If you don’t have to generate the reference, then it is a compact product which doesn’t need much power coming out of it.

Administrator
Shoreham EGKA, United Kingdom

Anyway https://easyeda.com/kenjewell/Arduino-ARINC is interesting project.

The examples there were extremely useful. Thank you for finding that!

I now have code running with both transmit and receive – label 320 at the moment. With the help of another pilot here who amazingly found an old MAXT PCMCIA interface card on Ebay in Germany (€150) and – with the help of the manufacturer who is still around – me managing to find some software to drive it, I have a great debugging setup on a winXP laptop:

ARINC429 is a pig’s ear to code, but with a modern fast CPU (I am working on an ST 32F417 which does 168 single precision megaflops and has a 32 bit barrel shifter) it is plenty fast enough.

There are ample opportunities for bugs. For example north is coded as 0, east is coded as +90, west is coded as -90, and you don’t need a PhD to spot that south is both 180 and -180 at the same time But these are 16 bit signed integers in the actual packet so it can be +32767 to -32768. The units are specified as 1/180 of a degree which is 0.00555555 but that overflows to 32400 so clearly everyone has “worked around that one”. To add some fun, some equipment uses 0.055 which underflows to 32727 so you get a “little jump” around 180 (a dead zone 0.4 degree wide) which few will notice. Using 0.00549333 gives you 32767 and -32767 and probably the best way, if the input angle is a float, is to test for > 179.99 or < -179.99 and fix these at +32767 and -32767 respectively, so exactly 180 never actually appears but the dead zone is just 1 LSB at 16 bits which is 0.01 degree.

I never found a connector for the PCMCIA card (Hirose stopped making them 10+ years ago, and any cables with them on were €400+) so I did a little custom PCB of the right thickness which pushes in

This is what ARINC429 looks like on the wires

That is the slow one – 12500bits/sec. The fast one is 100000bits/sec.

Everything to do with ARINC429 has an extra zero on the end of the price. That PCMCIA test setup was about 5k to buy new. The chips are expensive; I am using an HI3593 which is best part of $100 and that alone ensures anything using this interface will sell for 1k+. However, for receive-only uses, one doesn’t need any of this and one can do it with a fast micro for next to nothing. This, no doubt, is why a lot of kit has loads of receive channels and only 1 transmit channel

BTW someone sent me this which has useful docs.

Administrator
Shoreham EGKA, United Kingdom

This could be option to feed old wx1000+ with heading form g5/gad29b.

http://www.Bornholm.Aero
EKRN, Denmark

That shows that receiving ARINC429 can be done very cheaply, with a “software UART” approach.

The output interface isn’t useful however; you need a USB Controller to extract the data. To output ARINC407 (XYZ) you need to drive two DACs.

Administrator
Shoreham EGKA, United Kingdom

I found out how you generate XY from an external reference:

You don’t phase-shift anything. You just vary the amplitudes of the two outputs.

In reality, the 400Hz is usually neither 400Hz (the primitive MD26 is a free running oscillator, for example) nor is it a sinewave, but it works well enough when driving primitive mechanical instruments. If you want to do it properly you need to internally generate a pure sinewave which is phase locked to the reference. Garmin obviously took a shortcut in requiring the installer to provide that

If you are building a box which generates the reference that doesn’t change anything; you can just do it better. Sandel generate a pure sinewave by using a Class D audio amplifier, and the frequency is exact because it is done digitally.

Never managed to find a copy of ARINC407… even though the current version is 1958.

Administrator
Shoreham EGKA, United Kingdom

Peter wrote:

However, for receive-only uses, one doesn’t need any of this and one can do it with a fast micro for next to nothing.

I don’t see why (at least for testing/development purposes) you can’t do the output with a cheap microcontroller, either – the ARINC429 electrical characteristics aren’t rocket science.

Andreas IOM

Can anyone think of a moving map GPS which can accept the position data via ARINC429 labels 110, 120, 111, 121?

I need to independently verify whether the data I am generating here is correct

Administrator
Shoreham EGKA, United Kingdom
32 Posts
Sign in to add your message

Back to Top