thUndeadMod v3.4 - openLRS S**T just got real :)

[papa_lazerous]

What Mihai is suggesting is to set the frequency to 435Mhz in the configurator.  It's a relatively quiet area of the band and not as naughty as being on 433Mhz.  Lets also remember that the great range checks we all did recently where at 439mhz and nobody seemed to mind

[Splashy]

Yes, I agree that this is not as naughty and at 439 there were no incidents but I would not like to see this detract from the quest to get 459 working properly.

I have been taking a look at the datasheet and will try a few things out. Can anyone see a problem with using the spektrum analyzer RX as a way of comparing output power? From what I gather this should give a true report of the comparative power output.


I also see that HopeRF have a higher powered transceiver chip http://www.hoperf.com/rf_fsk/fsk/RFM23BP.htm this could be an option for a new board design.

[Coyote]

Ive done that with 433, 435 and 459, the spectrum shows exactly the same values with each band flashed. All peak around 227

[papa_lazerous]

Yes, I agree that this is not as naughty and at 439 there were no incidents but I would not like to see this detract from the quest to get 459 working properly.

I have been taking a look at the datasheet and will try a few things out. Can anyone see a problem with using the spektrum analyzer RX as a way of comparing output power? From what I gather this should give a true report of the comparative power output.


I also see that HopeRF have a higher powered transceiver chip http://www.hoperf.com/rf_fsk/fsk/RFM23BP.htm this could be an option for a new board design.

Hang on a minute I mentioned the other chip that HopeRF do (there are only 2 chips that are more powerful in this band) and said about a pro-mini and you called it a bodge in your post.  Now you are suggesting the other chip and a new board design.....Its like an echo.  You've suggested in essence the same thing that you called a bodge ten minutes ago.

[zitron]

I just went outside in the dark and suspiciously walked around the neighbourhood with my RSSI measurement box and did some experiments.



Setup:

Antenna: 15.2mm cut for 459MHz
Board: my own DIY RFM22B running custom code, should perform the same as OLRS
Settings: single frequency (no hopping), 38400 Baud, 19.6kHz frequency deviation, 40-50Hz PPM from er9x radio

Procedure:

1. Set frequency to 459.5MHz, output 1dBm
2. Walk with gps to ~50m
3. Write down RSSI
4. Repeat for 20dBm
5. Repeat for 435.5MHz
6. RSSI is calculated: RSSI  = raw value/2 - 130

Results:

I have attached the results. I would say my measurement uncertainty is +/-5dBm, and +/-5m for distance.

Conclusions:

From Dennis Frie's previous results, I would expect a ~3dB difference. I did not see a measurable difference between 459.5 and 435.5 using my crude method. It is likely that if I used proper 435MHz antennas I will be able to see a bigger difference. I have used 433 antenna on 459 before, and saw 3-4dB difference, so this is in line with Dennis Frie's predictions.

I have to say that the RF22 module works just fine at 459MHz, maybe with a small penalty. No where near as dramatic as you guys seem to be seeing. Clearly something else must be the problem. Have you guys tried disabling frequency hopping? 459MHz is very clean where I live I simply don't see the need for FH.

[papa_lazerous]

Well this could be the best news we have heard tonight (I say could with baited breath) How different is your setup from OpenLRS?  You stated you are running custom hardware and firmware right?  It would be interesting to learn from your setup if you are willing to share.  The plot is really thickening here

[webe]

Hey Coyote,

according to your post « Reply #76 on: March 12, 2012, 04:52:13 PM » you did range test with all the same setup but 459 UHF band?
I suggest you to repeat range test with 459 AND 433 bands each after another. If both bands will show same results as previous tests, then I see only difference - antennas. I dont think that would be a tranceivers problem...

Right, so I went out and performed the range test :)

The results, were to say the least, bad, worse than bad, terrible :(

1) This was a ground to ground test
2) TX was at 459MHz UK legal 100mw output power ( Or supposed to be but I seriously doubt it )
3) Tx antenna was a Diamond SRH771 SMA 433MHz
4) Hopping was 5 channels
5) A GoPro was on and recording, but then turned off to see if range came back
6) Gopro was 12" from the OpenLRS Rx
7) My OpenLRS antenna was not fitted as recommended directly to the Rx but on a 24" RG318 tail
8) 5.8ghz vtx was running the whole test with a linear antenna so to cause as much effect on results as possible. It shouldn't, but anything that was gong to be switched on while flying was.

I repeated last times test, the tx was sat at exactly the same place within 12" of last time, the weather was exactly the same as last time and I covered the same route for the test.

Expecting the same results as last time I drove straight to the 2.4km first check point and got out the car, not a thing, RSSI 0.04v no control :( I then just for the purpose of the test continued to the 4.3km point that 433mhz code has passed fine with no glitches no problem. Got out the car, RSSI 0.04 no control :(

Very disappointing

So I started from scratch, I went back to the tx to make sure everything was fine, it was, so I did a much shorter range test.

To my horror my max distance with no glitches was only 300m !

At 470m if I banked the plane to induce dbi loss I lost control at 90 degree bank.

At 870m if I banked the plane anywhere past 10 degrees, total loss of control :(

My conclusion :

1) Cat your not doing something wrong :)

2) I have no doubt Thundeads code is in no way at fault, the pins were moved up one, the RSSI code has nothing to do with the connection

3) I believe, very much, that the 433 chip just can not take the 459mhz, I believe it drops the power that much it renders the system useless. I also don`t believe the output is anywhere near 100mw, not even close. 459  and 433 are that close in frequencies that if they were using the same power, the ranges would match within a few meters of each other. Which suggests the chip is running way short of anything close to 100mw

4) The only way this is going to be improved is by using a different chip

[webe]

BTW if weather doesn't get worse I'm going to do range tests today using 435band with stock and Nagoya NA771 on TX, stock and DragonLink dipole on RX. I'll get back here and report my results.

[Coyote]

Hi Webe, yes im going out today armed with the laptop too, to flash to test it with 3.3 on 433, 3.5 on 433, 3.5 on 435 and 3.5 on 459. It will be a chew on, but will yield some conclusive results

[Devonian]

@ zitron

Great work, thanks for sharing, I'm sure that between us we can figure out what's going wrong here.

You say you used an antenna 152mm long - for 459MHz - I calculate 163mm.
Formula is:
300 (speed of light in Mega metres per second) / Frequency / 4 (for quarter wave)

300 / 459 / 4 = 163mm

I also have no need for frequency hopping as out in the sticks in rural Devon, it's a pretty quiet band !

Would you care to share your code and maybe some idea of hardware?

I'm toying with building a simple RSSI meter with a pro-mini board + RFM22B module and simple meter
http://www.dpcav.com/xcart/product.php?productid=16562&cat=252&page=3
I have a spare meter already from another project and it works very well.

Nigel

[thundead]

morning all :)

Coyote, be sure to test out 3.2 again because that gave us the best range. Maybe rssi on pin 1 still messes something up donno.

[Coyote]

Ok, thats gonna be a long test, 6x 5km tests lol

[thundead]

no point testing out 3.3 as its the one with broken rssi. try the one on pin 8 instead (3.1 or 3.2)

Ok, thats gonna be a long test, 6x 5km tests lol

[zitron]

Well this could be the best news we have heard tonight (I say could with baited breath) How different is your setup from OpenLRS?  You stated you are running custom hardware and firmware right?  It would be interesting to learn from your setup if you are willing to share.  The plot is really thickening here

There really shouldn't be any performance difference between my setup and OLRS, you can find details about my setup here: http://arduino.cc/forum/index.php/topic,93777

@ zitron

Great work, thanks for sharing, I'm sure that between us we can figure out what's going wrong here.

You say you used an antenna 152mm long - for 459MHz - I calculate 163mm.
Formula is:
300 (speed of light in Mega metres per second) / Frequency / 4 (for quarter wave)

300 / 459 / 4 = 163mm

Hmm.. I'm using the wikipedia formula 143/F, which gives me 156mm: http://en.wikipedia.org/wiki/Dipole_antenna

I did some experiments and found that 152-156 seems to be best for 22awg wire covered with heat shrink. But I'm not 100% sure if this is right.

[Coyote]

Live update, I'm at the range test location.

Test 1

Thundead v3.2 433mhz 5 channel hop 100mw >8.04km ground test.

Control loss only if 26 db loss incurred

Next test

Thundead v3.5 433 MHz same as above so it willprove RSSI and pin changes are fine