[Skimmertalk] image rejection

[Pete Smith N4ZR]

What you know, you know right, Andy.  Other comments below.

73, Pete N4ZR
The World Contest Station Database, updated daily at www.conteststations.com
The Reverse Beacon Network at http://reversebeacon.net, blog at reversebeacon.blogspot.com,
spots at telnet.reversebeacon.net, port 7000 and
arcluster.reversebeacon.net, port 7000


On 6/3/2012 11:47 AM, Andreas Hofmann wrote:
> I am also having a hard time to fix mirror image.  In my case it occurs when stations are stronger than about 40db s/n. I had to take off my Skimmer off the air (RBN) as I do not want to send wrong spots (and even worse, not a new spot every minute).
>
> Can someone really elaborate about how the i/q balance thing is supposed to be done?  There are not many documented instructions.
>
> Here is what I understand:
> - It should be done with an antenna connected and a strong signal generated.  I used my normal station antenna for SDR and TX into dummy load.
> - The TX signal should be moving across the whole band doing dits.  Is this correct?
> - while doing this the I/Q balance must be on, with both check boxes checked.
> When I do this, I see both curves adjusting.
>
> Here is what I do not understand:
> - is this independent of a band? I think it might, but not sure.
yes - it is happening at audio frequencies
> - what has to happen to the two checkboxes in the i/Q balance screen when the adjustment is done?
correct balance must remain checked.
> - should I do this on a band that has stronger signals to begin with (i.e. 40 in the evening)?
Over-the-air signals can always be used for IQ correction, if they are 
loud enough (>25 dB SNR)
> -When I run the mentioned routine, what dB of s/n should I expect the main signal and the image signal to be?  I was able to see the 40db difference when I did this. However
> People mentioned that the signal was supposed to be turned down, so I am not sure if I should generate a signal stronger than 50db or not.  This is really the most unclear item for me.
Best not to overload the receiver, so 40 dB is plenty.
> When do I know I am done?
When the curve is matched to data points across the entire range.  At 
that point you should see another 30-40 dB of image rejection.  For all 
practical purposes, the images should be gone.