A few years ago it looked like GA avionics will move to “glass” and everything will be connected up with ARINC 429 which is a digital connection (2 wires plus a ground/shield – not unlike RS422 i.e. from A to B, with 1 driver and multiple receivers).
However the average plane contains a whole pile of smaller instruments e.g. flowmeters, RPM sensors, etc and the technology in these has not moved since the 1980s. A lot of the companies’ names are to be found in the 1960s Apollo programme! They froze their technology in the 1980s. There are just very few exceptions e.g. Castleberry make an ARINC 429 version of their electric horizon for the autopilot connection – but virtually nothing is STCd to connect to it.
ARINC 429 is great for pressure altitude, where it replaces the 10 or 11 gray code wires, but half the equipment out there is not mutually compatible – e.g. Avidyne and Garmin won’t talk to each other.
So the average plane is still full of wires…
Garmin now own the GA universe but aren’t interested in expanding their product range sideways to e.g. fuel totalisers.
You sound like ARINC429 is somehow universal, yet at the lower end of the GA spectrum, there seems to be a lot more RS232 being used. Just look at the GTNXX IM, the RS232 compatibility list is way larger than the ARINC 429 compatibility list.
I’d say ARINC429 has no future, because it is an aviation only technology, which means volumes are very low and thus cost is exorbitant. Just enter “ARINC429” into the Digikey search box and cry. There’s only one transmitter chip which costs a fortune. If you need a receiver, you’ll probably have to build it using discrete components (and comparators). And then how do you connect the decoded ARINC429 signal to a microcontroller? No micro I know of has an ARINC429 interface. You could probably hack something using two event capture channels and decode the bits in software, but this is cumbersome and limited to very few (1) channel since most micros only have a limited number (such as 2) of event capture channels. Compare this to Ethernet, where you just write the packet to somewhere in the memory, and then tell the Ethernet DMA controller to send it. It’s similarly simple for RS232 (or 422).
All of Garmin Glass IMO interconnects using “HSDB”, which uses an Ethernet PHY (10baseT if I’m not mistaken). Unfortunately, the higher level protocol isn’t publicly known, so every manufacturer wanting to interoperate would have to reverse engineer it, or make an agreement with Garmin (not sure if this is possible).
So I see something ethernet based for higher data rates, and RS232 and possibly 422 for lower datarate stuff.
[formatting fixed]
My setup (GTN and Aspen) uses ARINC 429 for all communication between the PFD and the GPS. This includes GPS position, flight plan data, annunciations as well as horizontal deviation for GPS, VOR and ILS. It works very well, and is massively simpler than using individual cables for all that information…
The Aspen is quite happy with any ARINC429 data you throw at it, from just about any “standards compliant” source. This surely should be the future for interconnection – but of course it works against the vendor lock-in so beloved of [almost] all avionics manufacturers…
The vast majority of the avionics cabling in my stack is to tie the KNS80 and ADF into the Aspen and connect all the audio sources together into the Audio panel. This was completely redone when I had the GTN and Aspen installed, and is very neat – but its still a lot of cables.
Does anyone know how the GMA35 gets cabled up when combined with a GTN? Does each individual audio channel have its own pair of wires, or are the signals digitally muxed onto one line?
> Garmin now own the GA universe but aren’t interested in expanding their product range sideways to e.g. fuel totalisers.
Actually, the integrated (G1000 etc and G3x) solutions all talk to Garmin-sourced fuel totalisers (presumably white labelled from someone else though).
Yes – I was going to add that ARINC 429 is damn expensive to implement.
There are a few people who make the controlled slew rate driver interface chips for it, and there are some chips which implement the whole 32-bit packet stuff (which one would do with a microcontroller anyway, if one had one in the product anyway and it had spare time). There are also FPGA macros which implement the 32-bit part – that RS232 to ARINC converter I mentioned previously uses Altera FPGAs and it’s pretty obvious this is what they did (because an FPGA is a total overkill otherwise). I looked at this a few months ago, regarding a product I was going to develop.
If I was doing it I would do a discrete circuit for the driver and the receiver anyway and that would save something like $20-40 in parts.
The reason IMHO why the “little instruments” are RS232 is because (a) the interface is just a MAX232A which costs nothing (and you need it anyway, for e.g. factory config) and (b) the companies are mostly frozen in the 1980s, with the designer(s) having moved on long ago. For example I don’t think Shadin have any designers there anymore. It’s a nice business to be in – lots of cash cows, crying out for somebody to PMA them 
Does each individual audio channel have its own pair of wires, or are the signals digitally muxed onto one line?
I am sure it is analog audio. I don’t have the GMA35 IM though (just looked).
I think the GMA35 support three audio channels, of which only one can come from a GNS or GTN box. And the outputs have to go to normal headsets anyway.
