Executing this command:
rtl_power -d 0 -e 1m -g 49 -f 102200000:104600000:300000
on both rtl-sdr v3 and KrakenSDR produces very different results.
For rtl-sdr V3:
For KrakenSDR:
KrakenSDR and RTL-SDR are both not calibrated in terms of power measurements. So they will be different, and probably even different from unit to unit
1 Like
What about the difference in frequency response?
The RF input frontend of the two devices are also slightly different. The V3 has modifications to allow it to work with HF signals, so the response will be different.
Also it’s not possible to measure this without a noise source being on the front end. Did you have a noise source connected when rtl_power was running?
1 Like
I don’t have a noise source for the rtl-sdrV3 so I will have to rely on the krakensdr.
However, I’ve now got bigger problem.
Basically I can’t get a power measurement anymore from the krakensdr because it won’t deliver signal from my PCB antenna.
SDR++ tuned to a nearby FM station (103.1MHz):
-54dBm rtl-sdrV3 with PCB antenna
-73dBm krakensdr with PCB antenna
-42dBm rtl-sdrV3 with kraken antenna
-50dBm krakensdr with kraken antenna
All of these measurements were made with the same SDR++ settings through ch0 id 1000 and the same cables from the antenna.
Something must have changed on my host computer (laptop) but I can’t figure out what.
I suppose I’ll have to try to run SDR++ on the Raspberry Pi4 that I use for DoA --somehow – to control for whatever went wrong on my host computer.
Although I wasn’t able to run SDR++ on the RBP4 I was able to run the rtl_power command in that environment and compare the numbers that it reported in my laptop environment.
There is no substantive difference. In both environments there is about a 20dBm drop when going from the kraken antenna to the PCB antenna.
The problem is not in my laptop environment (either that or it is in both my laptop environment and the ISO image for KrakenDOA running on my RBP4 which is quite unlikely).
So what would cause the rtl-sdrV3 to have only a 10dBm difference between those antennas and the krakensdr to have a 20dBm difference?
SDR++ aznd rtl_power don’t measure in dBm because the RTL-SDRs are not calibrated. It measures in dBFS instead, so you can’t simply compare the two. You can try to compare signal SNR though and they should be similar.
1 Like
I can listen to the station with SDR++ and rtl-sdrV3.
Changing only to KrakenSDR I can barely hear it above the noise.
I’m really at a loss for how to proceed. Must I return my KrakenSDR?
Unfortunately, I do not know why you are getting poor SNR on the KrakenSDR. It could be ESD damaged, or have a blown front end.
If you want to return it, just email thekraken@krakenrf.com with the reason and your order details.
I figured out one problem is that I was confused by the 3 different numbering standards for the channels:
When looking at the KrakenSDR case, CH0 is called device 1 by rtl_power and rtl_eeprom reports it’s serial number as 1000. However, the default device number for the rtl_* commands is 0, which points you at the SMA connector labeled CH4 which has serial number 1004. I’d been trying to debug a 2 channel PCB with 2 channels by hooking them up to the SMA connectors labeled CH0 and CH1 – which frequently ended up with rtl_power reporting what it saw with an open circuit SMA.
That said, I’m still seeing a problem with gain when I use rtl_power with KrakenSDR vs using the rtl-sdr V3 dongle with rtl_power and that includes looking just at S/N ratios with, for example, the GnuRadio MeasureFmSignalPower.grc block with the Osmocom source replaced by the KrakenSDR source block.
I think the problem may boil down to the fact that with the Osmocom block I can set more than just one kind of “gain”, but with the KrakenSDR source block there is only one kind of “gain”.
Unfortunately, I’m using Linux which doesn’t support SDR# (which can set other kinds of “gain”) – only SDR++ (which supports only the “gain” supported by the KrakenSDR source grc).
I’m not certain how to go about getting the two SDR environments set up so I can compare their SNRs.
The osmocom block probably lets you set the LNA, Mixer and VGA gain settings individually. Whilst other programs just use a single slider which selects fairly optimal values for each of the three gain stages for each increment, resulting in a total gain only setting.