This is something that might interest some of the long-range FPV pilots.

Its a project that I've been working on in the background for some time, and finally took the time to put the preliminary code together and flight test it... it flew

So what is it?. Its a (potentially) long range replacement for the standard 35/40/72MHz R/C uplinks, using bi-directional modem modules.
For use in Europe, an 868MHz module (AC4868-250) may be used, with claimed 15km range (using std. whip antenna). For use in the US, a 900MHz module (AC4790-1000) is used (drop-in replacement), with claimed 20 mile (32km) range.
The US system has the advantage of performing FHSS (frequency-hopping, spread spectrum), with up to 32 channels.

So where's the catch?. Latency... a standard R/C receiver has a very short time delay between a stick being moved, and the servo starting to move. This is in the order of 20ms (the frame time).

For the modems, as I have them configured currently, this time is 100ms, 5x slower, with a potential to 1/2 this easily, and maybe the potential to reduce it a bit futher.

Now, does this matter? For a 3d heli, yes, for a 200mph jet, yes.. but for the kind of relatively slow & stable planes that are interesting for long-range flight, I claim no.

The first test flight was today (after writing the firmware last night). I installed the receiver on a slowstick that I've used in the past for FPV.
The system worked as expected. The latency is noticable, but certainly won't get in the way of long range flight with a stable model.
Even acrobatics, low to the ground, takeoff, landing, etc. where the latency would have the biggest impact, were quite possible.

I'm not interested in commercializing this design, but once some of the bugs are out of it, and the next board design has been tested, I'm considering publishing the design into the community.

AnthonyRC,

So are you using them in Transmit only or full duplex to get that rate?

Camship

For the moment, full duplex. Hop time is 53ms, so full duplex doubles this.
So far I haven't got them to run correctly in Tx-only mode, but from reading their docs it appears to be possible.
Ideally I want to run them in 1/2 duplex ~95% of the time, with an occasional directional reversal to allow the Tx to sense Rx RSSI, and other telemetry.

Originally Posted by AnthonyRC

For the moment, full duplex. Hop time is 53ms, so full duplex doubles this.
So far I haven't got them to run correctly in Tx-only mode, but from reading their docs it appears to be possible.
Ideally I want to run them in 1/2 duplex ~95% of the time, with an occasional directional reversal to allow the Tx to sense Rx RSSI, and other telemetry.

Hmmm, asynchronous transmit with firmware data pairing/checking and offload/download buffer from remote during low data transmit periods or trigger re-transmit for x-number of lost packets? Am I getting this correct? Am I making sense?

Camship

uh... yeah... what he said...

System is real simple (for now, after a couple of hours of coding), Tx decodes PPM using the CCP hardware, generates a packet of abs. positions which get squirted into the modem synchronous with the hop cycle. Rx generates 6 channels of PWM with a 20ms frame rate (to keep the servos happy), and when (minimaly error checked) packet arrives, positions are updated.
Button on Tx lets the Rx lock in the failsafe servo positions during the flight, once the plane is nicely trimmed, to handle the odd lost packet.

In order to come close to the european '10% tx. time amortized over 1 hr.' the 'nothing changed' packets need to be skipped... unless flights are kept below 6 minutes and I just tx all the time . Not sure whether this tx time 'rule' applies in the US, probably not since its FHSS.

In case some more people want it, and you consider producing it for selling, I would be really interested in one. It would take the hardest step out of my long range monitoring system I am trying to make.

Originally Posted by AnthonyRC

So where's the catch?. Latency... a standard R/C receiver has a very short time delay between a stick being moved, and the servo starting to move. This is in the order of 20ms (the frame time).

For the modems, as I have them configured currently, this time is 100ms, 5x slower, with a potential to 1/2 this easily, and maybe the potential to reduce it a bit futher.

I'm noticing a slight delay in the Maxstreams with the converter I'm using, so don't feel bad.. It's not horrible, but it is noticeable. So far I'm up to 1000' with 900mhz. This is one way communication only.

I've done pretty much the same with other modules, latency is normal but it's update time that suffers and needs to be limited at half the standard 50Hz rate. That's enough to see the servo making "steps" if moving the stick quickly from one end to the other, but as with Anthony's it would be totally flyable for typical FPV planes.
I just can't be satisfied with that "average" performance, and wouldn't be able to increase it without module firmware mods. And I still have to reach the range limit of my new Futaba RX, so it's not that much needed anymore either. Might take it out again when I have nothing else to do...

I think Dave Jones at auav.net did something similar. Maybe you could ask him for some advice? http://auav.net/DSS.html
-dave

( edit ) BTW, I started a thread back on this a while back, https://www.rcgroups.com/forums/showthread.php?t=652328 and ultimately found this: http://www.ringolake.com/pic_proj/se.../rc_trans.html any thoughts on that? Do you think it would have the same latency problems?

THAT'S Who did it.. I knew his name was Dave, just forgot his last name!

Originally Posted by dalbert02

I think Dave Jones at auav.net did something similar. Maybe you could ask him for some advice? http://auav.net/DSS.html
-dave

Interesting site!, many thanks for the link, some interesting stuff up there.

In some ways it is quite similar, my Rx has a PPM input for a std. R/C receiver, as a backup, and to reduce latencies for the critical take-off and landing.

The first FPV test flights with it were this evening, and went very well considering...
First flight had a LiPo problem, looks like a dying cell on a pack that I don't use much. About 1km out the power disappeared, but the uplink enabled me to pilot the thing all the way down to the ground (lucky, since the plane was over two fields with tall crops, one sunflowers, and one corn, and I was able to drop it neatly between them).
Second flight, with a new pack, was perfect, I didn't stray much past the 1km mark, but the system was rock solid. Need to try a longer flight with a chase-car as backup within the next few days.

Next step is to hook up the output power control to the button on the Tx, to allow range testing with very low power, and increase it during a flight.

Very professional and innovative job,
If you are looking for someone that can run a beta-site for you, I’ll be glad to do that. I’ll pay for it of course!
Itzik

Originally Posted by dalbert02

and ultimately found this: http://www.ringolake.com/pic_proj/se.../rc_trans.html any thoughts on that? Do you think it would have the same latency problems?

That's the circuit I'm using.... The building blog.. Although I've kinda hacked the code a bit.

After I fiddled around with the Maxstream settings I lost some of the 'stepping' that Kilrah is talking about. I've flown that unit plus a 900mhz card up to the 1000' mark so far. There is a noticeable lag with the PPM to Serial converter. Strangely it's about the same lag I get with my DX6. If you'd like after I get the kinks worked out of the system I can send you a board or something.

So far this setup would work VERY well for the head tracker. I'm still cautiously optimistic about flying a plane with it. Although the original author of the code said he flew a UAV 2 miles away with it.

Thank you Scott! Any assistance you could lend would be much appreciated. I have been following your thread blog with much interest but I missed the part of what circuit you were using. Please keep up the good work and keep us posted!
-dave

Interesting... this little project just jumped from something that I would trust to fly my FPV slowstick, to something that I'd consider putting in the cularis...

A few code changes, a bit of digital filtering, and I think this is quite usable.
Video attached.

http://www.aeropix.ch/videos/EzModemLink1.wmv