usrp-users@lists.ettus.com

Discussion and technical support related to USRP, UHD, RFNoC

View all threads

Adjusting VCTCXO of B210 with use of AD9361 DAC output

PK
Piotr Krysik
Mon, Apr 22, 2019 5:45 AM

Hello all,

I'm trying to implement better synchronization for SDR base GSM mobile
station (wiki of the project:
https://osmocom.org/projects/osmocom-bb-sdr-phy/wiki).

Currently the mobile station computes clock offset and applies
correction in software to a fractional resampler, in order to achieve
more accurate sample frequency synchronization in relation to a GSM BTS.
However doing this in software is still a bit awkward. It would be much
simpler (in terms of implementation and in terms of computations) to
have the correction applied at the hardware level. For example Lime chip
based designs (like LimeSDR or XTRX) have a DAC attached between FPGA
and VCTCXO, so it is possible to directly adjust the clock source frequency.

I noticed that AD9361 has DACs for controlling stuff and started to
think how to connect it to the VCTCXO tune input. Then I looked at the
B210 schematic and noticed there actually is a path from one of these
DAC outputs to the VCTCXO, but disconnected with use of missing R118
resistor:
schmatic: https://imgur.com/a/cgTfJ5G
pcb: https://imgur.com/a/NYhYxAb

My questions:

  1. Is it possible to put a resistor (with some correct resistance) in
    R118 place and not break AD9361 or ADF4001 or any other chip on B210?
  2. If not - do you know what other circuit modifications might be needed?
  3. Does anyone have experience with adjusting VCTCXO frequency though
    AD9361 DAC (this is probably question to B210 designer)?

Best Regards,
Piotr Krysik

Hello all, I'm trying to implement better synchronization for SDR base GSM mobile station (wiki of the project: https://osmocom.org/projects/osmocom-bb-sdr-phy/wiki). Currently the mobile station computes clock offset and applies correction in software to a fractional resampler, in order to achieve more accurate sample frequency synchronization in relation to a GSM BTS. However doing this in software is still a bit awkward. It would be much simpler (in terms of implementation and in terms of computations) to have the correction applied at the hardware level. For example Lime chip based designs (like LimeSDR or XTRX) have a DAC attached between FPGA and VCTCXO, so it is possible to directly adjust the clock source frequency. I noticed that AD9361 has DACs for controlling stuff and started to think how to connect it to the VCTCXO tune input. Then I looked at the B210 schematic and noticed there actually is a path from one of these DAC outputs to the VCTCXO, but disconnected with use of missing R118 resistor: schmatic: https://imgur.com/a/cgTfJ5G pcb: https://imgur.com/a/NYhYxAb My questions: 1. Is it possible to put a resistor (with some correct resistance) in R118 place and not break AD9361 or ADF4001 or any other chip on B210? 2. If not - do you know what other circuit modifications might be needed? 3. Does anyone have experience with adjusting VCTCXO frequency though AD9361 DAC (this is probably question to B210 designer)? Best Regards, Piotr Krysik
SM
Sylvain Munaut
Mon, Apr 22, 2019 4:17 PM

Hi Piotr,

The main issue is that the AUX DAC output swing is between 0.5v and 1V
(i.e. VDD_GPO-0.3v).
That's not the right range at all to control the VCXO meaningfully.

Cheers,

Sylvain
Hi Piotr, The main issue is that the AUX DAC output swing is between 0.5v and 1V (i.e. VDD_GPO-0.3v). That's not the right range at all to control the VCXO meaningfully. Cheers, Sylvain
PK
Piotr Krysik
Tue, Apr 23, 2019 12:00 PM

W dniu 22.04.2019 o 18:17, Sylvain Munaut pisze:

Hi Piotr,

The main issue is that the AUX DAC output swing is between 0.5v and 1V
(i.e. VDD_GPO-0.3v).
That's not the right range at all to control the VCXO meaningfully.

Cheers,

 Sylvain

Thanks Sylvain for the answer. It's probably not worth the effort to
create some circuit that will amplify that DAC output - especially every
user would have to do this as well.

USRP B200mini looks much more appealing in comparison. Just checked the
schematics and it has separate DAC connected to the FPGA, intended for
controlling VCTCXO. I need to learn how to control it.

--
Best Regards,
Piotr Krysik

W dniu 22.04.2019 o 18:17, Sylvain Munaut pisze: > Hi Piotr, > > The main issue is that the AUX DAC output swing is between 0.5v and 1V > (i.e. VDD_GPO-0.3v). > That's not the right range at all to control the VCXO meaningfully. > > Cheers, > > Sylvain > Thanks Sylvain for the answer. It's probably not worth the effort to create some circuit that will amplify that DAC output - especially every user would have to do this as well. USRP B200mini looks much more appealing in comparison. Just checked the schematics and it has separate DAC connected to the FPGA, intended for controlling VCTCXO. I need to learn how to control it. -- Best Regards, Piotr Krysik