Fixing up the Hamgeek pluto+ AD9363 SDR (libresdr/zynqsdr clone)

[MoA-Shaun]

Hi, recently I got one of the Hamgeek pluto+ AD9363 SDR and it seemed like the clocking was not even close to 0.5ppm accuracy, and the external 10 mhz input didn't work, reading the other threads it seems like other people with the hamgeek had the same issue and said theirs came with a TCXO and not a VCTCXO. What I hadn't seen was someone actually replace theirs with a proper VCTCXO yet and thats what I did.

Images:

First, besides checking known frequencies and seeing a shift on both receive and transmit, I also determined a way in software to confirm that you have a TCXO and not a VCTCXO

With a fake VCTCXO when you change the xo_correction to anything more than +-20 you will get an error, for example:

# iio_attr -d ad9361-phy xo_correction 40000100
ERROR: Invalid argument (22) while writing 'xo_correction' with '40000100'

And after I replaced with a proper VCTCXO you get:

# iio_attr -d ad9361-phy xo_correction 40000100
40000100

On a side note I think it would be extremely useful for tezuka to do a test during bootup or something that informs the user that the clock isn't a VCTCXO on all the different versions of this board, as I think a lot of people have just a TCXO.

I replaced the 40 mhz clock labeled MT217 (no idea what brand or model this actually is, if any one knows it would be appreciated) with a TXEAACSANF-40.000000 that I got off digikey for $2.70. I believe the TG2520SMN 40.0000M-CCGNDM5 should also work fine and is $2 cheaper, but its also smaller and I'd rather make my life a bit easier with the placement than save $2

After replacing the clock, 10 mhz input now works fine, but I'm still having issues with transmitting things with tight timing constraints like LTE and GPS (its possible this is a software issue since I've never tested it on an AD936X before though) and I haven't debugged it with a separate SDR yet.

On the receive side I seem to be getting a dc bias in the center channel frequency that I can't seem to get rid of that the AD936X should be able to handle (From what I hear at least, this is my first AD936X project and I could just have way to much expectations of this cheap sdr). I tried various calibration options, but if other users have a recommendation I'd like to hear it. Or if you think theres another hardware issue causing this, let me know. Additionally I'm getting a much wider bias on the top and bottom of the bandwidth range when receiving on a frequency without broadcasts (Another thing that should be able to be fixed with calibration, unless its a cheap hardware issue).

Additionally on certain frequencies I'm getting a quick pulse every 2 seconds or so were everything is pushed up 20db, I tried a whole bunch of commands to disable every single automatic gain control I could find but to no avail, and I have no idea if the AD9363, FPGA logic, or something else is causing this. Any ideas would be helpful.

[MoA-Shaun]

One thing I also noticed is when you remove the 10 mhz clock while its running, the frequency jumps back to what it was before you connected the 10 mhz clock. Shouldn't the VCTCXO be provided the same voltage after the 10 mhz gets disconnected? Also I just noticed that mine came with a 12 bit dac instead of the 8 bit (assuming it isn't counterfeit).

[davemwave]

Hi I have this same LibreSDR clone. It seems to have the VCTCXO and DAC because I can move the frequency around with "iio_attr -d ad9361-phy xo_correction 40000100" etc. How to I enable the 10 MHz ext reference tracking? When I connect an external 10 MHz the blue LED comes on but I don't see any small change in frequency. I am running "Universal Zynq70x/AD936x firmware builder
Version tezuka-0.2.2" from SDcard. Is the the right firmware? Merci beaucoup Evariste, F5OEO for excellent work! Thanks, Dave W1SZ