[time-nuts] 5065 Repair - Progress

[Ulf Kylenfall]

 5065A Repair - Progress...

As I wrote in an earlier mail, the restored 5065A
died (slowly), dropping the 2nd Harmonic reading
below 10 after which it lost lock.
I could not see any o/p signal coming from A7
so in a moment of confusion, I opened the Rb cavity
RX part to see if there was a loose connector inside.
Originally, I had written down the number of turns
used to remove the Rx, but I later realized that I 
had not been to careful when inserting the Rx again.
More about that below.
At first I took the 274Hz amplifier A7 apart, put it on the
lab bench and found the DC balance to be (badly/off scale)out of spec.
That caused me to replace one of the input transistors
that was indeed bad, and also the 709's with 741's after having
removed the frequency compensation networks.
A 1nF capacitor between the (+) and (-) input
of the 1st OPamp reduced the output noise considerably.
(Tested with 100 nF w/o any adverse effect on the output).Then I discovered that there was no 60 MHz SRD Rx pumping
power. 
This was caused by one of the JFETs in A3 gone bad.
Replacing it with a 2N4416 increased the o/p power but not enough.Replaced the final transistor Q9 with a 2N5109 and then the unit
produced +20dBm. Made a complete recal of A7, A3 and A8, replacing
A8R8 with a 10-turn pot. (Made the 274.0 Hz adjustment easier)
Having put back both units, the 5065 was able to lock but
with a reduced 2nd Harmonic reading. Obviously I had
missed inserting the Rx unit back into correct position.I then decided to try to tune the Rx cavity, which I did by
removing the RV ass:y, connecting it to extension cables 
and opening the Rx part, loosen the fastening and rotate
it and check for maximum AC. This was successful. The peak
was not extremely sharp, but the difference could be clearly
seen on an oscilloscope. (I also noticed that this procedure
should best be carried out using as little surrounding
light as possible.) (Are there no instructions how to do this"Professionally" out there?)

After reinsertion, the 2nd harmonic level was back to
what it was at first (~30), and the unit locked nicely.
(...Sigh of relief...)
Today I took it to my work, left it on during the better part
of the day, did fine frequency adjustment (Mag Field) against
the Hydogene Maser and the connected it to a Symmetricom 5120A
using the H-Maser as reference.
At first, the Allan Deviation exceeded HP's original specification
by almost a magnitude at 100 seconds, but as the time
 progressed, that reading went from 2E-11 at 10E-2 seconds
to 8E-13 at 100 seconds (minimum) going up to 2E-12 at 1H 26 minutes.Phase noise starts at -20dBC at 1E-3 Hz, -100 dBC at 1Hz offset,
then -150 dBC at 10E3 seconds (flatterns out there).
As far as I can tell, and if I dare to trust the Symmetricom instrument, the old 5065 seems to be working.
As I mentioned, I have also acquired a second unit, unfortunatelypartially scrapped. It hard to resist trying to fix that unit as wellbut then I need to look for a 10811 quarz unit for it.
Ulf Kylenfall    SM6GXV