Telemetry propagation tests

With my 20X2 RFM42 lost Model Locator as a tranmitter and the matching RFM31 receiver built into a portable unit, I have been able to carry out some comparisions of the effectivnesss of the various transmission modes these Hope RF units are capable of.

Power output of the transmitter was at the lowest possible, the data sheets are confused as to how much power this is, but its not important for the comparision. I beleive the power output is 1mW or less. Frequency was 434Mhz. Testing was done in an open space over the local park. Transmitter antenna vertical stuck to the back of a plastic chair.

At;

FSK and 10kbs data rates the reliable error free reception range was 15.5M
GFSK and 10kbs data rates the reliable error free reception range was 15M
GFSK and 10kbs data rates, with Manchester encoding, the reliable error free reception range was 7M

FSK and 2kbs data rates the reliable error free reception range was 15.2M

With a UHF tranciever as the receiver the Morse audio (same power) could be very clearly heard at 300M, I ran out of space in the park. Given the strong clear signal heard at that distance I would estimate the Morse audio range as arounf 1000M.

So Morse the very clear winner, at least 20 times better range, possibly more than 60 times the range. What I will next compare is the automatic Morse decoder I built (see the projects forum) with the data telemetry.

Morse rate?

Interesting... what is the Morse rate you were using? You quote for 10Kbit rate but if you were to take that down to 100bit rate you should get much further range... One time I had a reliable Morse communication over about a 1500 Mile [US] path with .0-15 milliwatts input power - perhaps 10 baud [the antenna was rather long - Windom style compared to model aircraft]. Data rate and frequency stability and receiver bandwidth are primary determinants...

Morse was 15WPM, easy to decode by ear. I might see what happens to the automatic decoder if you try to send at a much higher rate say 50WPM.

The RF module is specified as being capable down to 0.125kbs, I tried 1kbs but it did not work. And whilst I agree that in theory the lower data rates should perform better, there was no noticeable differance between 10kbs and 2kbs.

You infer that when listening to the Morse you were using as UHF transceiver, if so I'd expect better reception range as I'm sure the typical 0.1uV sensitivity exceeds the sensitivity of an RFM31 module (-118dBM or 0.28uv) and if so, then the comaprision is not truely valid, I suggest normalising the results for the UHF receiver to the basline sensitivity of the RFM31 to give at least a better comparison: 300M * 0.16/0.28 = 170M, plus were you using SSB or NFM for reception (widely accepted to be equal in demodulation efficiency), if WFM then add another (effective) improvement of gain by using SSB of 6dB, so reducing the normalised range even further.

The on-paper sensitivity of the RFM31 modules is very close to a typical UHF receiver, it's surprising that there is such as difference in performance, I suppose it's down to the relatively wide band-width needed to send data at 10Kbps; compared with for example PSK31 (31Hz) as these narrow-band modulation techniques like F/PSK are as good as (better in some respects) than Morse. See RADCOM Jan-2010 for more details!

I see what you mean about the data sheet, no definitive Power Output specfication, so assume what? the minimum or maxmium for the 868 Mhz band - probably 10mW in pulse form (actual range is 10mW to 500mW with varying duty cycle)

OT but - Is the Lost Model Locator waterproof?

@srnet - Here's video of my son flying his Bixler completely out of sight, 2000 ft out over water! May explain why I'm working on 'Go home' enhancements.

http://www.ilovefpv.com/

Hobby King Bixler FPV over Cold Spring, NY

g6ejd said:

You infer that when listening to the Morse you were using as UHF transceiver

Click to expand...

I did more than infer, thats what I said.

I was using NFM, deviation on the RFM42 transmitter for data set at 25khz, and for the morse audio about 8khz, that was the deviation that gave the cleanest sounding audio.

The large differance surprised me, I was expecting a differance of course, but it looks like the useable data telemetry range is only about 1/50th of that which can be achived with Morse and a cheap UHF tranciever. I will see if I can get the data telemetry working at the slowest possible speed, speed the Morse up and equalise the deviations.

The data sheets are a nightmare, lots of errors and some critical omissions, it took a long time to get the data telemetry working at all.

Armp said:

@srnet - Here's video of my son flying his Bixler completely out of sight, 2000 ft out over water! May explain why I'm working on 'Go home' enhancements.

http://www.ilovefpv.com/

Hobby King Bixler FPV over Cold Spring, NY

Click to expand...

It can be, it would fit inside (with battery) one of those waterproof keyholders or mobile phone bags you can get for going swimming. The ones I use are just coated with conformal coating. If you dont coat them it only taskes a little bit of damp to raise the sleep current above the normal 20uA or so. The locator can be put to sleep for long periods to conserve battery power.

My Electric Gentle Lady went for a swim in the sea once (Lipo battery failure) The plane survived with little more than a broken read stab, most of the electronics never worked again.

I have RTH available via my Eagletree logger and OSD setup, but I have never used it.

In all honesty if it went down in the Hudson at this time of the year there's little/no chance of recovering it. If you look at the video you can see there's no river traffic at all.

My son has got some OSD chips, and wants 'us' (ie me) to look into adding GPS data to the video downlink - have to see how much the RTH 20M2 code takes. The downlink is 5.8Ghz, with range >2500ft, but we don't know how far yet.

FSK and 10kbs data rates the reliable error free reception range was 15.5M

Click to expand...

It would be interesting to know the range using Morse equivalent baud rates (say 5 baud ~300 bpm ~15wpm) and the impact of Manchester coding on that, if you were able to do the test. This is fast enough for a lot of purposes even with the overhead of a crc to correct some errors.

srnet: Your Morse findings aren't too surprising! I've made past tests with cheapie Keymark 433 MHz modules (of ~a few mW Tx power & ~ -104 dBm RX sensitivity). The Morse ID could be received (by ear) to 100s of metres LOS with the simple RX hack, and ~1km with a cotanga Yagi. It'd be interesting to try your Morse decoding software with this setup - can you point me to a link ? Stan.

geoff07 said:

It would be interesting to know the range using Morse equivalent baud rates (say 5 baud ~300 bpm ~15wpm) and the impact of Manchester coding on that, if you were able to do the test. This is fast enough for a lot of purposes even with the overhead of a crc to correct some errors.

Click to expand...

Well you can use the module for direct modulation so you could slow the data rate down as much as you like.

In packet handler mode there is no overhead to using a CRC, its generated and checked automatically. All a CRC does is tell you if a packet is consistent, it cannot correct any errors.

manuka said:

srnet: Your Morse findings aren't too surprising!

Click to expand...

Surprising to me, I did not expect the differance to be so stark.

I need a larger clear flat area to do further checks.

There are local beaches long enough (2-3km) at low tide, such as 'Bad Wolf Bay' from the Dr. Who TV series, but the tides are wrong until next weekend.

What a crc can do depends on the polynomial in use. But error correction is possible if you have enough bits. and a suitable polynomial. The crc can be computed on the fly in hardware, but the overhead is the extra data that you have to send.

I'm interested in use of simple radio channels with low power at 433MHz in the telemetry band, and how to maximise quality/range at ground level. One day I will do my own tests, but I am interested at the moment in Manchester encoding because claims are made that I'm not sure are borne out in practice. I haven't seen any test results and just wonder if these tests might cover that ground.

geoff07 said:

but I am interested at the moment in Manchester encoding because claims are made that I'm not sure are borne out in practice. I haven't seen any test results and just wonder if these tests might cover that ground.

Click to expand...

See the first post, adding manchester encoding cut the range in half, not what you might expect .......

srnet, I remember from my oceanographic days that telemetry tests over wet sands can give some very unrepresentative results. A buoy to shore link even seem to go through a null at certain ranges. All many years ago now but it did seem like black magic at the time.

I have just been able to test the portable magic Morse decoder with the same transmitter and power, the decoder started to fail at around the 250M mark.

So it would appear that for my application a cheap UHF tranciever and a simple tone decoder circuit (based on a 08M2 and NE567 PLL) will outperform the telemetry receiver by such a margin (a factor of 15 or more in range terms) that the telemetry is not worth considering. I have conducted tests in a 'real world ' lost model situation by placing the lost model locator in a spare model fuselage at ground level in the middle of a wood. The audible range under these worse case type conditions was around 1km, derating that for telemetry by a factor of 15 would would give a detection range of only 70M or so, not very useful. Of course the amount of data I need to pick up is small, as little as 8 characters for a position fix.