No receive for OFDM sounding package

[Yiyang Li]

Hi,

I've tested OFDM sounding with dense deployment: Tx from EBC, receive with guesthouse, moran and ustar with shout long range experiment profile and pc04-meb.emulab as orchistrator. After some debugging, the last run still failed to produce meaningful received waveform, despite a simple link budget test showed that I should get something at least with the shortest link (EBC to guesthouse). I wonder if someone could help check what happened/what might be the reason, either I made rookie mistakes in certain setups, if I should just crank up TX gain higher than 45, tx and rx windows are not in sync, or there is some system/hardware problem such as the antennas are not facing each other/NLOS that makes receive impossible. Full details are attached in the zip folder.

Thank you so much for the help!

Best regards,

Yiyang

[Kirk Webb]

Hello Yiyang,

Since the SDRs you are using at the Dense Deployment sites are USRP
B210s, you will need to use gain values much higher than you are
currently using. Since OFDM signals have a PAPR of ~10 dB, I would
suggest trying a TX gain of 80. Try ~72 for the RX gain. You may need
to play around with these gain values some. The values I've suggested
are near the maximums. You may find that you can (or need to) back
them off some.

The transmitter and receiver should be roughly time synchronized, as
set up by Shout, and IIRC, the transmit code should send out several
repetitions of the waveform to give the receiver a reasonable chance
of capturing it.

-Kirk

> --
> You received this message because you are subscribed to the Google Groups "Powder Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to powder-users...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/powder-users/4c6099c2-354b-42cf-be60-a8f7129edc77n%40googlegroups.com.

[Yiyang Li]

Thank you Kirk! 

I had assumptions that higher gain is needed but I didn’t want to just crank it up that high without some permission. In this case I’ll try to set up the experiment again with said gains, will follow up with the new results.

Yiyang

Kirk Webb <kw...@cs.utah.edu>于2026年10月1日 周四13:17写道：

To view this discussion visit https://groups.google.com/d/msgid/powder-users/CAFKO4CHzqvFXzHu-%2BHwW2AAO6hgryB4CR-xVK3vv_M7KqZz60g%40mail.gmail.com.

[Dustin Maas]

Hi Yiyang,

Agree with Kirk about the gains… there is an RF front end here, but its gain is not particularly high, and this is not a narrow band signal.

One further suggestion: a good smoke test is to start up uhd_siggen/uhd_fft at your transmitter/receiver pair and just see if you can see a tone in realtime. If so, start up your transmit loop with your wideband OFDM signal then see if you can see it with uhd_fft. Adjust transmit/receive gains such that you are getting seemingly good received signal in uhd_fft (good SNR; not clipping/saturating). You can do this for each pair of radios that you plan to use in order to calibrate gains.

-Dustin

To view this discussion visit https://groups.google.com/d/msgid/powder-users/CALmxb8MOUUuEYPx-oghuwKLp_SjTnhQ4VirNT-RG2Wocm6X2Qg%40mail.gmail.com.

[Alex Orange]

Yiyang,

     Just a note, the uhd_fft/uhd_siggen in the SHOUT profile are not patched to work with the frontends. To do as Dustin is suggesting (which I think is a great idea) I'd suggest using our SDR-outdoor-GNURadio profile. You can use uhd_fft/uhd_siggen as suggested to get a feel for what gains to use and double check everything is working as expected. Then you can even transmit and receive your waveforms using the same profile/experiment. As explained in the docs, the suggested way to do this is to extend your transmit waveform using extend-waveform (installed in that profile) to some bit longer to give some empty space before/after. If for instance you have 10ms of waveform you might extend to 50ms. Then you transmit on repeat and receive at the other radio with 2x this extended length to guarantee you get at least one full sample.

     A note about gains, you can transmit with as much as you like. I would just keep going up until you find there's too much distortion. Given that you're operating above 3 GHz you may not even see a non-linearity problem at the max of 89.75 dB. For the receive side you want to basically run at the maximum gain where you don't see distortion. Use uhd_fft while you're transmitting, go to the time tab, zoom out on the Y axis, set the trigger to auto, change the trigger level and then reduce the gain until you don't see it going over something like 0.3-0.5. Sometimes you may need to go lower if there's some out-of-band noise, but 0.3-0.5 is a good place to start. If you don't get that high don't worry about it, just don't go higher than that or you'll have distortion problems.

Alex

To view this discussion visit https://groups.google.com/d/msgid/powder-users/8646F97D-D6D5-41FF-8781-FCA55CA52608%40gmail.com.

[Yiyang Li]

Hi all,

Thanks for the help and feedback! I'm getting good results from Moran TX and some reasonable results from Ustar TX and Moran RX. There are hardware aliasing and some time sync issues, but overall I could observe the OFDM I was transmitting now.

Best regards,

Yiyang