[hpsdr] ANAN200D RX1/RX2 Phase Coherent and Phase difference (2)

[Jan Simons]

Hi Joe,
Thanks for the explanation. I understand it is not a simple modification.
There is more to that than I thought.
I agree that such a modification needs more customers.

May be in the future the KissKonsole becomes an option.
My focus now is on the applications.
Checking phase difference or equalizing RX1/RX2 is an acceptable workaround.

Tks es 73 Jan PA0SIM


Op 1/29/2017 om 17:17 schreef Joe Martin:
> Hi Jan,
>
> Yes certainly, as Warren noted, the PowerSDR program architecture could be modified to accommodate nearly anything imaginable; but of course, sometimes such modifications to accommodate a particular application may not be optimum for the public version of PowerSDR.  I don’t know if this particular situation falls in that category or not but this kind of situation is what KissKonsole was made for; namely, allowing users to modify code, simpler than PowerSDR, to achieve what they wish.  Alternatively, I or somone could work with you to produce a special version of PowerSDR for your specific application if modifications to the public version of PowerSDR to accommodate this particular situation aren’t viewed by the custodians of the public version as being generally desirable.
>
> The truth is that doing such a software modification would not be quite as simple a change as you suggest because when the Diversity mode is active the IQ stream arriving in the software from DDC0 has been modified in firmware to contain interleaved IQ samples from both DDC0 and DDC1, not solely IQ samples from DDC0 as it does in normal Rx mode.  This would require some re-routing of data streams in the software to accommodate your desired scheme, not simply enabling Diversity mode.
>
> We discovered that the interleaving of IQ pairs had to be implemented during Diversity mode to ensure that software will always have correct, corresponding IQ pairs from each DDC at identical time instants to use together for the manual phase/amplitude controls.  Previously when we tried to use independent streams of IQ pairs during Diversity mode the correspondence (time relation) of IQ pairs from the two ADCs would occasionally “skip” one or more pair positions in one or both streams, for any of several reasons, which in turn causes a phase difference to be inadvertently introduced between Rx1 and Rx2 in the software.  But this is a technical nit that can be worked around by suitable changes in software and/or firmware in a specialized version of either/both for this specific application.
>
> Further discussion of the issue is probably warranted.  I’ll be happy to try to help you find a solution.
>
> 73, Joe K5SO
>
>
>> On Jan 29, 2017, at 1:30 AM, Jan Simons<jan.m.m.simons at ziggo.nl>  wrote:
>>
>>
>> Thank you Joe and all for the reply.
>>
>>
>> I think I understand now why the phase between RX1 and RX2 is not fixed.
>>
>>
>> Isn't it possible to also force both DDC receiver modules to use the same phase word in the C&C stream when the Stereo Diversity switch in enabled?
>>
>> In that way the RF IQ streams and the demodulated audio could be fixed in phase.
>>
>> I am not familiar with the architecture, so I don't know how complex that modification would be.
>>
>>
>> My problem using the IQ streams is that I have to program two receiver modules in my post processing.
>> That takes processing power and I have to make a GUI to tune to stations.
>>
>> Especially processing power is a concern. So I prefer demodulated audio.
>>
>>
>> As I understand VAC will only be an option if the signals of both receivers are send via a single stream.
>>
>>
>> Tks es 73 Jan PA0SIM
>>
>>
>>
>> Op 1/28/2017 om 20:39 schreef Joe Martin:
>>> Hi Jan,
>>>
>>> I find the description of your implementation of diversity using the ANAN-200D/PowerSDR interesting.  It is, of course, distinctly different from using the Diversity button option in PowerSDR.  Perhaps a few comments regarding the fundamental differences and a few words about how the Diversity feature is implemented in firmware will aid in evaluating the various possibilities that exist for phase-coherent diversity operations and perhaps shed some light on why the relative phase shift you observe between Rx1 and Rx2 occurs upon changing bands.
>>>
>>> A major difference between your configuration and what is done in the PowerSDR Diversity option is that you are comparing phases and amplitudes of two demodulated/audio IQ streams from two DDC receiver modules  whereas the PowerSDR Diversity feature  combines the two raw  rf IQ data streams, one from each receiver, to form a single raw rf IQ stream that is then processed by the software DSP routines to form a single panadapter display and single audio output signal.  PowerSDR Diversity provides the user the ability to digitally shift the phase and amplitude of an rf IQ stream from one DDC receiver module relative to the rf IQ stream from a reference receiver module to form the single IQ stream.
>>>
>>> The significance of this is that in your configuration selecting VFOsync on the PowerSDR front panel simply forces the frequency of Rx1 and Rx2 to be the same value and those values are used to form phase words for each receiver that are then sent separately to the two receiver modules in the firmware.  This is generally fine to achieve “pseudo-coherent” operation of two receiver modules, as they will be operating on the same frequency, but because the receiver modules are running independently and are not updated with precisely the same phase word at the same instant a phase shift between the two can occur when new frequencies/bands are commanded by the user, due to the fact that the phase words are delivered to the firmware in a serial stream of Command & Control (C&C) data.  This situation is avoided in the PowerSDR Diversity feature by forcing Rx1 and Rx2 to use precisely the same phase word from the C&C stream to update the frequencies of the two receivers simultaneously.
>>>
>>> If the firmware updates the receiver frequencies by using phase words from different locations in the C&C stream there will be a time delay (phase shift) introduced between the two receivers, even if the phase words themselves are identical in value because the receivers will be updated at slightly different instants in time due the fact that the arrival times for the two values from the serial C&C stream are different, as is the case in your configuration.  In the PowerSDR Diversity option, however, both receivers are forced in firmware to use the same phase word in the C&C stream, thus avoiding any such time shifts.
>>>
>>> Although I haven’t examined this in detail in PowerSDR and firmware with respect to band changes, I suspect that changing bands causes a greater shift in arrival times of the C&C values from the software than does a minor frequency change within the same band.  Also, of course, when bands are changed the RX filters being used change and that fact in itself can certainly introduce phase and amplitude changes in the relative signals of Rx1 and Rx2.
>>>
>>> I think there is an option in PowerSDR to send raw IQ values from a receiver to a VAC output so that may be useful to you, and operate with the PowerSDR Diversity option selected.  I’m not familiar with that possibility actually, but it would seem to me that using such a scheme would allow you to maintain a fixed phase relationship between Rx1 and Rx2 even upon band changes. I think you’d need to use two VAC connections though, maybe not.  Maybe someone who has used that raw IQ output capability can add a useful comment or two for you.
>>>
>>> Anyway, if I’ve missed something in my interpretation of your configuration I apologize.  I hope these comments will be helpful to you in any case.  Others may have better information for you.
>>>
>>> Good luck with your project!
>>>
>>> 73, Joe K5SO
>>>
>>>
>>>