It took a while but I promised I would report on this after I had completed it,
for the archive.
This radio came to me with the 5.0 to 5.5 digital VFO completely out of lock
and no amount of adjustment to the VCO or tuning encoder would lock it up.
The VCXO was working, and moving as it should to fill in the gaps in the 10 KHz
resolution of the main PLL, so I knew that data was arriving from the CPU. They
share a data link.
I could see the 5 MHz output change with the VCO coil, and the signal was
plenty strong, but it would not move in frequency with the VFO knob.
I first changed the PLL chip as I had thought it was ignoring data from the CPU
due to an internal failure. That went well enough but still no joy.
I then made some measurements on the feedback loop IF and found an open 10 Ohm
resistor in the Drain of the loop IF dual gate MOSFET. I replaced that and the
transistor.
I then proceeded to work on alignment since the thing would only lock and tune
on a small portion of the band.
I treated the loop IF amp like an IF amp from a TV set and stagger tuned the
coils as follows:
1) I connected an oscilloscope to the test point at the 10 KOhm test pint. You
cannot use a DVM because this moves around a lot and is too noisy for a DMM.
2) I peaked L3 for the center of the band. I peaked L2 for the low end and L1
for the high end. I made these choices because this is where the coils ended up
most centered.
3) In order to ensure easy lock of the VCO during this process, I moved the VCO
to near each frequency at which I was working, and looked for lock condition as
that is the only way I could be certain there was the correct frequency signal
passing through the IF for the coil I was adjusting.
4) At the completion of this sequence, I was able to get the thing to lock from
about 5.100 to 5.400 MHz, nothing better no matter where I tried to compromise
on VCO frequency slug setting.
5) I pulled out my working Omni 5 and measured signals, in particular, I
measured the signal into and out of Q4, the amplifier that buffers the VCO
signal into the VCXO mixer to produce the difference frequency as an IF. The
Omni 5 showed about 3 to 1 voltage gain in this stage, the Omni 6 only showed
unity gain. I replaced this FET with one I had in a junker board and the gain
went up and now the thing locks at all frequencies.
Notes:
1) Don’t adjust anything unless you need to. I tried to pull the VCXO to get it
closer to the specified 5.00000 MHz and ended up with a condition in which the
VCXO would tune almost the right amount but the wrong direction, giving an
effect a lot like one would see if a bit were stuck on the BCD encoder only
worse - kind of like a hundred whistles blowing by you and you’re hearing the
Doppler shift of each one as it passed by when you tuned across a carrier.
2) Finding a process for the Omni 5, using Barry’s reference, I was able to get
this back to working, but only after I reset the coil L6 back to where it was,
and then I painstakingly twiddled until the VCO went just 10 Hz when told to do
so, not more, not less, or I would hear steps in the tuning.
3) As mentioned above, a DVM is not your friend on one of these.
4) A working radio as a reference was really handy!
5) For reference, the VCO tuning voltage goes from about 1.3 volts on the low
end of a band to about 7 volts on the high end. Once repaired, the thing is
pretty forgiving in that regard as I was able to get it to lock at both ends
with the VCO tuning voltage setting shifted 1 volt either way (by adjusting the
VCO coil).
This is a really neat design, especially considering how old it is. The soft
steering of the LO crystals in the HFO does not introduce significant phase
noise and, while the 5.0 to 5.0 MHz VFO has plenty of phase noise, it is
reduced in bandwidth by the ingenious division of the VCO frequency by 40. That
is why this thing works so well. On the Omni 5, even the tiny noise from the
PLL acting on the crystals was totally absent (as was frequency accuracy over
time).
As many unkind things as I have said about this radio’s receive audio, it is
otherwise still a really nice toy due to its excellent design work.
And then they went and topped it with the Orion! Hope there’s something new and
even better coming along…
73
Gary
W0DVN
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|