Pure Signal Issue

spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Pure Signal Issue

Postby spinomaly » Tue Dec 19, 2017 4:19 am

I have a 7000DLE running PSDR 3.4.5 and am having issues running PS. If PS is disabled then the transmitted two-tones look reasonable. If I uncheck the PS disable but don't enable PS-A I get very an output that does not look right. The output is independent of drive level. I can bring it all the way down to 1% and the result is the same. Below are figures of my setup and results. Any assistance would be appreciated.
Attachments
unnamed3.png
unnamed3.png (313.63 KiB) Viewed 21590 times
unnamed2.png
unnamed2.png (307.34 KiB) Viewed 21590 times
unnamed1.png
unnamed1.png (281.91 KiB) Viewed 21590 times
Dr. Loren Betts
W6VH
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Tue Dec 19, 2017 4:19 am

More figures.
Attachments
unnamed5.png
unnamed5.png (145.87 KiB) Viewed 21588 times
unnamed4.png
unnamed4.png (313.61 KiB) Viewed 21588 times
Dr. Loren Betts
W6VH
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal Issue

Postby w-u-2-o » Tue Dec 19, 2017 12:14 pm

Which firmware, Loren?
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Tue Dec 19, 2017 3:20 pm

PSDR version 3.4.5. Not sure about the FW in the 7000. In setup there is a reference to 2.1.
Dr. Loren Betts
W6VH
K9RX
Posts: 414
Joined: Fri Apr 14, 2017 3:47 pm

Re: Pure Signal Issue

Postby K9RX » Tue Dec 19, 2017 3:31 pm

just a precaution - it might not effect what you're doing but I noticed a couple things:
1) I know that if you have some (all?) of the noise abatement stuff on the panadapter will be effected on transmit. Mostly I notice that if NB2 is on that the xmit waveform gets bounced around instead of being steady ... so just to be cautious you might want to not have NB2, SNB on (nor NR/2) ...
2) I see you have an FT8 profile selected ... note with the latest version of software (starting I think with 3.4.5) the issues with incorrect removal of audio enhancements had been fixed ... previous to this the CFC could get left on ... also I believe the leveler. Now you don't need to select a separate profile for digital modes - just set up your SSB profile and leave it at that - the program will automatically turn off the EQ, CFC, COMP as needed ... with that said I don't know what you have defined in your FT8 profile - perhaps it is related to the issue at hand?

Gary
K9RX
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Tue Dec 19, 2017 3:37 pm

Thanks Gary,

The FT-8 profile is a modified one of a SSB profile I have been using. Only real difference is the equalizer is disabled.

Disabling all noise reduction did not appear to help.
Dr. Loren Betts
W6VH
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal Issue

Postby w-u-2-o » Tue Dec 19, 2017 4:45 pm

The various noise reducing/blanking options will not be in effect when monitoring transmitted signals. You should be using DIGU for digital modes in order to be assured that all TX audio processing is disabled (it is disabled automatically when DIGU is selected) but that is also not a factor in this problem.

I am very suspicious of whether or not the firmware is making timing on your particular Orion MkII board. Although the AmpView display shows good data, something is clearly going wrong on transmit.

There were three releases of Orion MkII firmware in quick succession: 2.0, 2.1 and the latest is 2.2. See here. Be sure to also look into the "previous" folder. These revisions came out in quick succession because there were reported problems with improper timing closure on certain serial number boards, and most of these reports were associated with PureSignal.

Two suggestions:

1. With PureSignal running, see if there is any substantive difference between DUP on and DUP off. DUP on is showing you a true, full duplex view of your own transmission as routed from the internal coupler to RX1. DUP off is showing you the calculated spectra of the data going out to the DAC. If they are effectively the same that implies that the problem is with receiving the PureSignal data or with processing it. If DUP off looks good, then that implies the problem is in the transmit side of the firmware somewhere.

2. In any case, I'd strongly suggest trying firmware versions 2.0 and 2.2 and see if that makes any difference.

Loren: are you comfortable using HPSDR Programmer to load new firmware? It's not hard. See here.

73,

Scott
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Tue Dec 19, 2017 6:29 pm

Hi Scott,

Thanks for the info on DIGI. Disabling DUP appears to show a good response. I will try updating the FW in the 7000.
Attachments
Untitled.png
Untitled.png (871.5 KiB) Viewed 21547 times
Dr. Loren Betts
W6VH
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Tue Dec 19, 2017 6:36 pm

I am noticing something strange on receive. Looking at CW signals you will see that there are additional side tones when PS is enabled whereas they do not exist when PS is disabled.
Attachments
Untitled3.png
Untitled3.png (1.59 MiB) Viewed 21546 times
Untitled2.png
Untitled2.png (1.59 MiB) Viewed 21546 times
Dr. Loren Betts
W6VH
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal Issue

Postby w-u-2-o » Tue Dec 19, 2017 6:52 pm

I'm also noticing that rolling noise baseline. Try some different firmware!
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Tue Dec 19, 2017 7:04 pm

I'm pretty sure the rolling is some type of external signal. It is prominent in the mornings. I will try loading FW tonight.
Dr. Loren Betts
W6VH
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Tue Dec 19, 2017 9:34 pm

Loading FW 2.0 or 2.2 did not solve the problem. It doesn't appear it is a coupler issue because it would have shown bad spectral content when PS was disabled. It also doesn't appear to be a final driver issue for the same reason.

Is anyone else using PS with the 7000?
Dr. Loren Betts
W6VH
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Tue Dec 19, 2017 9:39 pm

I also tried 3.4.6 from my developers branch and it also shows the issue.
Dr. Loren Betts
W6VH
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal Issue

Postby w-u-2-o » Tue Dec 19, 2017 10:06 pm

PureSignal definitely works for others on the 7000. It would not be a software issue otherwise all radios would likely be affected.

Does this happen into a dummy load?

Desperation shot: do you have another computer you can use?
Joe-W4WT
Posts: 167
Joined: Sun Apr 09, 2017 5:27 pm
Location: Cumming, GA

Re: Pure Signal Issue

Postby Joe-W4WT » Tue Dec 19, 2017 10:22 pm

I had exactly the same problem with exactly the same panadapter view with 2.0, 2.1, and 2.2. I dropped back to 1.9 on my 8000 board and all is working well so it appears to be a firmware issue with my 8000, at least.

73,

Joe W4WT
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal Issue

Postby w-u-2-o » Tue Dec 19, 2017 10:28 pm

Wow, that's going a ways back, Joe! Worth a shot, though, Loren...

eta: I'm at 2.1 on the 8000 here and all is well. I've been avoiding going to 2.2 since if it isn't broken...
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Wed Dec 20, 2017 12:17 am

Hi Scott,

I found out the cause of the issue although I do not understand the mechanism. I had the sample rate in the Primary Audio tab set to 38,400. Setting it to 19,200 appears to solve the issue. I wonder if it is a display issue since the amp view looked reasonable.

Also, should this be used in digital modes?
Dr. Loren Betts
W6VH
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Wed Dec 20, 2017 12:18 am

I will try version 1.9 as well.
Dr. Loren Betts
W6VH
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal Issue

Postby w-u-2-o » Wed Dec 20, 2017 2:58 am

Ah ha! That's the problem! PureSignal is not supported at sample rates above 192KHz. This is combination of two things. First, although on paper it should work, 100MBit/s Ethernet has trouble keeping up with the bandwidth required when sample rates are at 384KHz. During transmissions you are sending an IF data stream to the radio, and receiving 2 IF data streams back, for a total of 3 IF data streams, plus audio and control/status data streams. This is also why stitched receivers are disabled when PureSignal is enabled (although I can see a software/firmware method to get around this limitation, it's just not implemented). Second, the PureSignal algorithm is optimized for 192KHz.

So go back to firmware Rev. 2.1 or 2.2, keep things at 192KHz, and be happy! :)

73!

Scott
spinomaly
Posts: 33
Joined: Fri Oct 20, 2017 5:44 pm

Re: Pure Signal Issue

Postby spinomaly » Wed Dec 20, 2017 3:35 am

Thank you for the explanation. What are the two data streams coming back? IQ data?
Dr. Loren Betts
W6VH
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal Issue

Postby w-u-2-o » Wed Dec 20, 2017 4:11 am

Hi Loren,

See Page 19 of this presentation:

http://www.n3sh.org/Tech%20Section/PureSignal_final_selection.pdf

The presentation is a bit dated, but should provide a good introduction to the basics. In order to obtain the best possible time aligned samples, the outgoing data stream is picked off just before the DAC and sent back interleaved with the feedback data stream.

Remember that the ADC sample rate is 122.88MHz, and typically 14 bits wide, but that is transformed by the DDC to a data stream that consists of a 32 bit word, 16 bits of I and 16 bits of Q, at the first IF sample rate (typically 192KHz).



73!

Scott
Joe-W4WT
Posts: 167
Joined: Sun Apr 09, 2017 5:27 pm
Location: Cumming, GA

Re: Pure Signal Issue

Postby Joe-W4WT » Sun Dec 24, 2017 1:37 am

Hi Scott,

After you mentioned you are at 2.1 with your 8000, I decided to give it another try to see if I maybe I had a bad firmware load but I still have the same problem as shown above. Don't know why that would be other than, of course, I have a "hybrid" system with the 8000 board and the 200D PA/Filter board. Obviously I have a different PuresSgnal return path and maybe that affects things. Version 1.9 works perfectly so I'll stay there and try whatever is next to see if it might work.

73,

Joe W4WT
K9RX
Posts: 414
Joined: Fri Apr 14, 2017 3:47 pm

Re: Pure Signal Issue

Postby K9RX » Sun Dec 24, 2017 4:32 pm

Scott,

You said in March: "The various noise reducing/blanking options will not be in effect when monitoring transmitted signals. "...

I didn't see that previously (i have been through this thread before) but wanted to comment this is not correct. NB/NB2 will effect monitoring TX signals. The TX waveform will jump around on the screen until these are turned off. I had submitted a list of things I've found - this was in that list. I see its still not 'fixed' in .7 ... I had asked that these be turned off when TX'ing.

Or maybe I'm the only one seeing it in which case I'd have to wonder if it is a combination of settings because it is most definitely 'real'.

Gary
K9RX
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal Issue

Postby w-u-2-o » Sun Dec 24, 2017 4:42 pm

Gary,

You may well be correct, however I will say that no combination of noise blanking or noise reduction modes makes the waveform "jump around" on display during TX.

I would suggest that further discussion of this issue is worth it's own, separate topic, if you'd like to start one.

73,

Scott
K9RX
Posts: 414
Joined: Fri Apr 14, 2017 3:47 pm

Re: Pure Signal Issue

Postby K9RX » Sun Dec 24, 2017 8:32 pm

Scott,

Not sure what you're saying ... "you may well be correct" and then followed by "no combination... makes the display jump around on TX"?

My post is there because it DOES indeed make it jump around! Are you saying yours doesn't?

When I say 'jump around' what I'm saying is it appears similar to when PS is "searching" for a solution and not there yet - i.e. the shoulders 'jump up' - appear at a higher signal level.

[I can start another thread based on your response - if your response is "I've this moment tried this using NB2 and it isn't acting strangely at all for me" then I'll go back and check to see what else might be adding to the anomaly as it will then be. ]

Gary
User avatar
w-u-2-o
Posts: 5540
Joined: Fri Mar 10, 2017 1:47 pm

Re: Pure Signal Issue

Postby w-u-2-o » Mon Dec 25, 2017 12:01 am

Mine does not jump around regardless of NB or NR state during transmit.
K9RX
Posts: 414
Joined: Fri Apr 14, 2017 3:47 pm

Re: Pure Signal Issue

Postby K9RX » Mon Dec 25, 2017 2:37 pm

Scott,

I have to say this is the one thing I find very frustrating about PowerSDR ... the interdependencies can be so high ... I am and have been seeing the jumps, as far as I remember, since day one - just like other issues I've described that you were not able to duplicate - they ARE real ... and in all (most?) cases I've found the connections and solutions. The most recent thanks to Rob being the use of QUICK REC/PLAY. But it is a bit tiring to have to do so.

I'll see what else can be causing it. I will start a new thread under PowerSDR General Use.

Gary

Return to “PureSignal Operations (PowerSDR & Thetis)”