[hpsdr] openHPSDRj on Linux with a new ANAN-100D - no tx power on most bands

[Darren Long]

Hi all.

I have a new ANAN-100D to enjoy and only have Linux computers available
for use with it. After pottering around with using other code for Rx
only, I thought I'd try using OpenHPSDRj to fire up the transmitter.

Mindful of the apparent need to perform a per-band tx gain calibration,
I attempted to go through such a process, for a 20W output power level.

This worked fine on 80m, 60m, 40m and 20m.  However on 160m, 30m, 17m,
15m, 12m and 10m there was no indication of any output power at all.

Here is some observed data. The SWR readings came from openHPSDRj, power
readings were from an external meter, which coincided well enough with
the tx power reported by the s/w. PSU current measurements are very
approximate.

80m 20W OK @1.0:1 SWR with PSU current about 9A
60m 20W OK @1.0:1 SWR with PSU current about 9A
40m 20W OK @1.0:1 SWR with PSU current about 9A
20m 20W OK @1.0:1 SWR with PSU current about 9A

160m 0W @1.7-2.2:1 SWR with PSU current about 4A
30m  0W @1.7-2.2:1 SWR with PSU current about 4A
17m  0W @1.7-2.2:1 SWR with PSU current about 4A
15m  0W @1.7-2.2:1 SWR with PSU current about 4A
12m  0W @1.7-2.2:1 SWR with PSU current about 4A
10m  0W @1.7-2.2:1 SWR with PSU current about 4A

For the good bands when I activate tune in openHPSDRj, I see the PSU
current kick up to about 4 amps initially, for the PA bias, then the
carrier appear at > +10dBm  in the power spectrum for the duration of
the transmission and the current kicks up to 9A.  This signal's strength
seems to be unaffected by the per-band PA Settings slider and the drive
level slider.

For the bad bands, the carrier appears very briefly at a little below
-70dBm and then drops off almost immediately below the lower axis. The
PSU current remains steady at about 4A until tuning stops. As with the
good bands, this signal's strength seems to be unaffected by the
per-band PA Settings slider and the drive level slider, although that is
mostly a guess, as I can no longer see it in the power spectrum display
once its level drops.

My primary concern is that there a hardware fault in the ANAN-100D.

Another possibility is the new 4.9 f/w for the Angelia which came out
the day before I ordered the rig and came pre-installed.

So far I've been following up on the notion that this might be a
software issue in openHPSDRj. I have attempted to determine if this is
likely, by reading the USB protocol documentation and reviewing the
openHPSDRj code.  I don't really 'do' java, but I can't see any obvious
issues whilst rummaging around in the source.  I'd love to run it in a
debugger, but I installed NetBeans and it crashes on startup. I don't
fancy diving  into that can of worms right now, as I'm already deep
inside a more interesting rabbit-hole :)

In the openHPSDRj code, the RF gain is scaled by the drive level and
per-band PA gain for Penelope, whilst for other transmitters, the
normalised RF gain scaling of 1.0 is consistent with the docs.  I see
the C1 drive level byte being scaled by the product of the per-band PA
gain and the Tune drive level.  I think this is ignored by Penelope, but
not by other transmitters, i.e. the Angelia in the ANAN-100D.

I suppose the per-band PA gain setting obtained by calling
configuration.bands.get().getDrive() might be the prime candidate for an
anomaly like this.  The attribute band.drive seems to be initialised to
0.3, which I think is a reasonable starting point. I suppose there could
be an issue with the serialisation/deserialisation to/from the binary
config file, but I don't think that's likely.

So I'm left with the remaining possibility that there's an issue with
the IQ tune samples that only affects these specific bands in a
repeatable fashion. I haven't looked into the code for that yet.

I'm now thinking about how I might analyse some UDP traffic captured in
wireshark to see what is happening with the drive values and the TX IQ
samples.  I wonder if anyone has a dissector for the Metis UDP and USB
protocols.

I'm also wondering about the transverter tx output from the ANAN-100D.
I recall seeing somewhere that it might be at a 0dBm level, but I'm not
sure if it needs enabling somehow.  If this is likely to work, then I
was thinking of perhaps hooking that up to my USRP B100 and BasicRx with
60dB's of attenuation so I could see what is going on there.

Another option for me might be to have a play with the gr-hpsdr blocks
in gnuradio, to see how the transmitter behaves.  I've built gr-hpsdr
and have the blocks available in GRC, but haven't had a fiddle yet.

Thanks for reading. All suggestions welcomed.

Cheers, 73

Darren, G0HWW