Is it possible to use AGC on the rtlsdrs now we're using Soapy SDR?

Hello, I’m trying to set up a second station with a nooelec rtl-sdr and testing the signal, with various configurations, in gqrx. I get great results in gqrx by just using the Hardware AGC (which over rides the manual LNA slider in gqrx), but poor to no signal if I adjust using the manual LNA slider. Is it possible, now we’ve moved to Soapy SDR, to replicate the control I have in gqrx and set the satnogs client to use AGC alone?

Historically I’ve found the AGC in the RTLSDR to be fairly poor - tending towards too much gain resulting in lots of IMD products and a reduction in available dynamic range. It’s also designed to work with signals much wider in bandwidth that those we are looking at.

The vast majority of stations in the SatNOGS network are going to be limited by local noise. In these cases, the gain should be fixed to a point where the observed noise floor on the SDR starts to rise. This maximises the available dynamic range of the SDR, and optimises the sensitivity.
I’ve written a guide on how to determine this point here:

For stations with very low noise floors, AGC may work OK (by virtue of it pegging the gain to full), but for many stations it is likely to cause more problems.

What is your antenna set up, and what sat were you looking at?


Thanks for such a speedy reply! I have a home made QFH and I’m interested in picking up signals from 430 to 440MHz. I’ve sourced a number of LNAs and filters from China and am mixing and matching to see how well they perform. This LNA looks promising:

and performs as advertised according to my Nano VNA, and it’s this one that seems to work more successfully with AGC in gqrx.

I’d meant to add… it may have good noise levels. There’s only about 40cm of coax from the antenna to the LNA which then goes into the rtlsdr through a switchable attenuator. The rtlsdr is kept at the antenna end and the Pi inside the house using a USB over ethernet extender, powered at the antenna end by POE.

Thank you for the link to your wiki for gain setting. That looks useful.

Often stations are limited by noise from their surrounding area, though this is usually less of an issue on 70cm.
If you’re using a LNA you can run coax back from the antenna instead of running USB-over-ethernet, which could potentially produce noise.

Have a go at setting the gains per the guide - I’d be interested to know at what gain setting your noise floor starts to rise. Just note that you can’t turn the Bias Tee on via SoapyRemote, so if you’re using that, you will have to enable it separately using rtl_biast.

1 Like

Thanks. I’ll have a go at that over the next few days as well as make up a coax lead to test. I’ll post how it goes.

1 Like

OK, seems to be running nicely now. I tested the noise floor which was very similar to your example at ~ 20db. I moved to coax from ‘usb over ethernet’ as you suggested, but still had the same problem… blank waterfalls at all gain settings. So I tried a different LNA:

and it all sprang into life. This was one of the cheaper LNAs and behaved differently to the other LNAs when tested on a Nano VNA. The Nano VNA CH0 saturated this LNA and needed attenuation for a measurement to be made. The other LNAs I’d been using coped with the output from the Nano VNA. This LNA is also working with a turnstile antenna I made for 435MHz. I now need to fine tune gain for both the QFH and turnstile, which I’ll have a crack at using your instructions… instead of just plugging in another Raspberry Pi running gqrx.

Any suggestions for getting the LNA I referenced in my second post working? I’ve tried up to -35db of attenuation on the output, but that didn’t appear to be the problem. I’m not an electrical engineer, simply a purchaser of Chinese electronics.

Out of interest, I may have a go at switching back to ‘usb over ethernet’ to see if it does degrade the signal. The components seemed well shielded… but not the cables, of course.

Thanks again for your help, and for the page on gain setting.

I don’t know if you have resolve this, but I found it very helpful to set and fix the limits of the waterfall. You can do this under the satnogs advanced settings under waterfall. Set the high limit, set the lower limit and then change the adjustment from True to false. You can then adjust the gain and see exactly what is happening to your signal on the waterfall.