My results were like yours. The final cal that I did a U and then S was done in 128seg. All of my fials for the 2 channels were around .00001-3 for all freq. There is still a rise in the scope center at 48khz of about 30db. Is this normal? I know the value of your time so I will continue to play and read your web page. My next challange is to figure out the B=;C= and D cal routines.
> I am not having success eliminating the image spur.
> I am looking for directions on how to do this on the
> Linrad site. In the Linrad program I have been trying to:
> C= cal FREQ response
> at next menu
> A= Balance spurs
> I select any number of segments to 256 and move my SIG
> gen until it is all green. After I U=update correction
> in ram and then S=save correction in ram; I do this 2
> times and X back to Regular CW. I don't see any difference
> in performance or screen appearance.
1) Place your signal generator at 2.510 MHz
2) Go to the mode, A for weak cw etc.
3) press X, C and A for Balance spurs
4) Look at the last text line. It gives phase and
amplitude of the image spur.
For an uncalibrated system, the line could looks like
Ch(0) A=0.012490 ph=-1.173634 Ch(1) A=0.008494 ph=-0.457400
This is actually what I see on my system here when running
it uncalibrated. As you can see, the amplitude is about 0.01
which is 1% or -40 dB.
If you a see much smaller amplitude, your system is already
calibrated. Check what the calibration is doing by pressing Z.
(If you exit without pressing S you will not destroy the
5) Wait for one section to become green, do not
change the generator (default segments=32)
6) Press U, then return to the cal screen, do not
change the generator.
When you do this you calibrate with a single segment which
means that you apply the same correction for all frequencies.
Since the signal generator is still at the same frequency,
the correction should be perfectly fine for it and the
last text line should look something like this:
Ch(0) A=0.000029 ph=-0.657651 Ch(1) A=0.000042 ph=-1.544115
The spur is now suppressed by more than 70 dB. This is with
fft1 bandwidth 200 Hz. With a fast computer you can select fft1
bandwidth 20 Hz and have 90 dB suppression.
If you now movce the frequency slowly, you will see that
suppression is good at one frequency only. The one that gave
the calibration coefficients applied for all frequencies.
By doing exactly as you described, you should get good
suppression on all the frequencies that were used to
shift segments green. It should be enough with 64 segments.
If the above does not clarify, I would like to see screen
dumps but off list.
> >>What to do depends on what you want. If you want to listen at 2.5 MHz<<
> I will be using Linrad to do VHF and above contesting
> in the US. I have all of your hardware through 70 MHz and
> am patiently waiting for the 144 MHz converter to become
OK. They should have arrived here last week.
> I will be obtaining at least one more set of boxes after
> I get this receiver working. The quality of the commercial
> RX2500 and converters is fantastic. I am a manufacturer
> in the US and am amazed at how these units could have been
> produced at this low cost.
I have done my best effort to keep all costs as low as
possible and there is no profit in this project. Sometimes
I have to wait when something more profitable pops up at
Leif / SM5BSZ