Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
Amplitude and phase I/Q imbalance corrections not applied
|
I use Quisk only for reception. I recently noticed with Quisk 4.2.41, python 3.13 running in an up to date Archlinux that the image rejection was very poor (no more than 20dB).
I followed the well explained procedure through Config/Config/RX Phase... I can put the image of the frequency under test down at least 60dB. The phase and amplitude corrections are well stored in each corresponding cells. I checked this by readind the quisk_init.json file. But when I restart Quisk everything is exactly as before the correction was done : same poor image rejection on band reception, same image level of the frequency under test in the correction procedure.
?
I use the imbalance correction procedure from the beginning of using Quisk. I am sure that until Quisk 4.2.37 the corrections of the I/Q imbalance were perfectly applied and the image rejection was very good (at least 60dB).
The hardware is the same : a home made SDR RX board forwading I/Q streams to a Xonar SE sound card.
?
To day I removed all the Quisk stuff then rebuilt it from source and reset it to my hardware but without solving this problem.
Here is the logfile for the last restart:
*** Quisk started on linux at Mon Feb 24 21:06:25 2025
WxPython 4.2.2 gtk3 (phoenix) wxWidgets 3.2.6 Start of wdsp Found private wdsp Wdsp version 118 Library wdsp is active Open channel 1 ?
I attach my custom, very simple hardware file : quisk_hardware_xonar.py. It makes Quisk works from the beginning.
?
I don't know what I am missing or doing wrong now, so any help and/or explanation will be greatly appreciated. |
|
Many thanks Jim for your quick answer.
?
I installed Quisk 4.2.37 in a separate directory. To get it work I had to installed the package of telnetlib for python 3.13
I did the corrections for 15 m. The values are in good limits, they are in quisk_initi.jason-4.2.37. It results of an image rejection of at least 35 dB : the strongest signal I received over noise.
?
I checked back with Quisk 4.2.41 for 15 m. No change.
?
I had to stopped my rig. When I fired it again on the 15 m band with Quisk 4.2.37 the good image rejection had gone away. I did a new correction. The values were as bad as with Quisk 4.2.41. They are in quisk_init.jason-4.2.37-bis.
Before sending the first message I tested with an other sound card in my PC but with the same behaviour. I am wondering if I have a random problem with my SDR Rx board.
?
Anyway Jim keep enjoying skiing
?
Pascal
F6FJF
? |
|
Hello Jim
?
This morning I checked thoroughly my rig as I did two weeks ago before investigating in Quisk and I also found everything good.
?
I did the corrections for 15 m with Quisk 4.2.37 : perfect result. I switched to Quisk 4.2.41 and with the same quisk_init.json the imbalance was quasi perfectly corrected, just some little tweaking.
I did many stop and start both of my rig and Quisk 4.2.41? Always I got at least 60dB image rejection in the correction procedure and at least 50dB when listening the band.
After a rather long break I restarted my rig with Quisk 4.2.41 and all is still perfect.
?
I don't know what was going wrong during two weeks, what makes the problem disappear. But I am quite sure that it is not due to Quisk.
So, if you aggree, as it is not a Quisk related issue I will delete this topic.
?
Before that I want to tell you how much I enjoy using Quisk since the beginning a little more than one year ago.
Congratulations for this software.
?
Pascal
F6FJF |
|
What hardware/audio interface are you using with Quisk? On Wed, Feb 26, 2025 at 10:28?AM Pascal F6FJF via <pelletier.betp=[email protected]> wrote:
|
|
Never mind, I just found the "Xonar SE sound card" remark. It might be time to test that card to see if it has the subsample delay problem. On Wed, Feb 26, 2025 at 10:29?PM Chuck Ritola via <cjritola=[email protected]> wrote:
|
to navigate to use esc to dismiss