HP 5065A Rubidium Standard Repair Log

[trobbins]

Just noting that A3 in hp5065a_part2.pdf for -6078 for series 1420.

[bnz]

Yep, I've seen that. But there seem to be substantial differences:

[KE5FX]

Wow, 18 adjustments in this module alone. 

Far be it from me to call out HP for criticism, but... that just ain't right.

[chuckb]

I'm in the middle of aligning a new series 2644A A3 module right now. It's a project!
Corby posted the -2644A schematics and manual pages on the time-nuts forum a few years ago.
Attached is the schematic.

[chuckb]

The time-nut link for the Series 2644A A3 module, thanks Corby!

https://febo.com/pipermail/time-nuts_lists.febo.com/2020-September/101338.html

[Whaleback]

Corby - thanks for that. I checked the EFC+ and EFC- leads and they were correct. The problem was the position of the phase splitting pot.
I had it set for beautiful split, at inverted phase. After winding it around to the correct polarity (at the end stop), UNIT ACHIEVES LOCK!

I have three things remaining really to sort out:


1) Neg 20 rail is at -19.7 when positive rail is exactly at +20. This only when A7 installed. When I disconnect A7 power, neg rail returns to -20 exactly. I have a concern something is still amiss in there.
I expect the supply rails should track perfectly under good operating conditions, is that presumption incorrect?

2) Phase splitting pot range winds out. I can't quite get to good splitting even after tweaking all of the drive and match pots/caps on A3. I'd really like to re-center the phasing range on A8 from the edge of the dial - is this an aging effect maybe from dried out caps on A8? Has anyone looked into this?

3) The unit came with a 10811-60109 OCXO. When the unit is set to open-loop, the BNC control signal reads around 300 mV when tuned to on-frequency. Control drops to zero when I disconnect the 10811 EFC- SMB. When I swap out the 10811 for a 00105-6013, control is 0 V as expected. It appears I have an issue with the EFC- pin on the 10811? Has anyone seen this happen with units with a retrofit 10811?

Gents - thanks for all the help getting this far!

[Whaleback]

So I've been thinking about how I can contribute to this community. One thing I thought might be a nice to have could be a remote monitoring device for collecting "meter" data off the instrument and/or also tracking other testpoints.

I'm thinking about designing a whole A17 replacement card with a LTC2499 24-bit 16-channel delta sigma ADC. ADC data is I2C, could also put in a 10k thermistor input and uC controlled latching relay to monitor A12 and cut heater power on fault. Or could use 67F100/102/105 instead of L version so no current in wiring unless fault. Anyone interested in this. I'd do the design in Altium and make it available at PCBway or OSH Park?

I'm also considering a redesign on the PSU module around LT silent switchers and TPS7A4700/TPS7A3300 high PSRR low noise linear regs.

[5065AGuru]

The -20V supply is only regulated by two selected zeners in series. I would not worry about the absolute value. It only runs the A7 and is not critical.

The phase adjust pot can end up at one end. As long as your splitting is OK that's fine.
If splitting is a bit off try reducing (going CCW) on one of the two pots on the top of A3 (can't remember which one) until you can get proper splitting.

If that does not work you need to apply an external 137Hz Through a 10Meg resistor to the input of A7.
Monitor the output of the A7 where it comes into the A8 card connector. adjust the frequency of your generator for a peak amplitude. You are measuring the actual frequency if the A7 tuned circuit. Note it.

Then return A7 to normal and adjust the A8 frequency to match. I have found that this will allow better adjustment of the splitting

The control signal is biased by the new arrangement used with the 10811, If you look at the old versus new loop gain procedure you will find the old one checks the control voltage on both sides of resonance. The new one only checks one side as the other side will not be balanced. This is not important in closed loop as the loop gain will cover both sides adequately.

You might make sure the resistor arrangement was converted to the new circuit for the 10811, It has one resistor mounted on the standoff by the fine Quartz adjust vice the two in the old configuration.

Cheers,

Corby

[KE5FX]

So I've been thinking about how I can contribute to this community. One thing I thought might be a nice to have could be a remote monitoring device for collecting "meter" data off the instrument and/or also tracking other testpoints.

I'm thinking about designing a whole A17 replacement card with a LTC2499 24-bit 16-channel delta sigma ADC. ADC data is I2C, could also put in a 10k thermistor input and uC controlled latching relay to monitor A12 and cut heater power on fault. Or could use 67F100/102/105 instead of L version so no current in wiring unless fault. Anyone interested in this. I'd do the design in Altium and make it available at PCBway or OSH Park?

I'm also considering a redesign on the PSU module around LT silent switchers and TPS7A4700/TPS7A3300 high PSRR low noise linear regs.

Remote monitoring is a good idea.  I went down that road a while back -- having used an Arduino to provide thermal cutout support, it was easy enough to have it monitor photo current and 2H -- but didn't go any farther than that.  Obviously its built-in ADC isn't in the LTC2499's league, but arguably you don't need that kind of performance unless you just want to go overboard.  NTTAWWT.

[bnz]

...
So a word of caution to everyone buying a used 5065A, you might want to check the caps or RIFAlitis.

Good advice I think these schould be promoted in the list of capacitors to check in the 5065A.

[Uky]

Food for thoughts:

Since my 5065 A15 regulator board was badly corroded, I designed an improved regulator board.

Using the "hair-dryer" test method, it outperformed HP's original design in terms of temperature drift.

It is a drop-in replacement but it requires the original transformer/capacitor/inductor
AC supply to be replaced with a switch-mode version AND filtering bringing the input voltage
down to 24VDC.

Check out KO4BB's 5065 folder.