HP5065A Odd Fault

[pamphonica]

My nice HP8065A is playing up.

I confess that it does not get run full time, but I run it up for a few days every few months and it has been fine for many years.  It is stored reasonably warm and dry.  It has an RVFR dated 1991, although a number of the circuit boards have caps with 1968 datecodes.  No serial number (scraped off), but the lid shows Serial No 916-00184 and RVFR Serial A07436 which matches the RVFR.

A few days ago I wanted to do the long-awaited caps change so I warmed it up for an hour and it started first time.  I noticed that it had developed an odd fault.  If I tapped it, even gently, the 2nd Harmonic leapt then returned to normal.  Anyway, I got it stable and left it on for a day.  I then powered down again to change all the caps that 5065AGuru (Corby) recommends, and heatsink that poor little mushroom rectifier on A2. I haven't done the 1.33K resistor in the RVFR yet.

Once done, another 1 hour warmup and it fired up fine so I left it on soak test overnight.  Come daylight the Continuous Operation (CO) lamp had gone out.  Damn!
But if I tapped the case a few times I could get the 2nd Harmonic to jump and if I pressed logic reset the CO lamp lit but extinguished after a few seconds.

It feels like I have something like a cracked resistor - does that make sense?  When it was working, everything was looking fine on the meter and on the various outputs.  I waggled boards gently to prove that the fault was not just bad pcb seating.

Are there certain resistors prone to cracking (maybe due to heat)?   I'll be doing a thorough check of the various circuit boards to see if anything measures oddly, but any hints would help.

Any help or advice welcome!

-Jeremy 

[pamphonica]

Update: some carefully calibrated tapping all over the instrument seems to locate the intermittent effect to the RVFR lump itself.  I'll remove it and check and if necessary replace those often-toasted resistors inside.
-Jeremy

[KE5FX]

The 1.3K resistor may be toasty but it's unlikely to be the source of the problem at hand.  You can rule it out by looking at PHOTO I instead of the 2nd harmonic.  If the photocurrent remains steady while tapping on the unit, then the problem is unlikely to be associated with the lamp.

Once the covers come off, the first order of business is to check the various SMC coaxial connectors on the bottom of the chassis to make sure they are tight.

It's a closed-loop system, so a good second troubleshooting step is to open the loop, and there's a switch on the control panel for just that purpose.  After you familiarize yourself with the block diagram, check out some of the documents published by Corby Dawson and others for some good test points to monitor with a scope. 

Another suggestion: switch back to closed-loop mode and see if you can tell whether the problem is in the OCXO versus the rest of the loop.  (In an updated unit at least, the OCXO is right next to the RVFR; you can't really bang on one without affecting the other.)  Compare the 5 MHz output to a known-good source with a scope, either in X/Y mode or simply by examining the two traces directly.  When the glitch occurs, does it ring down or recover immediately? 

If you can't tell anything with the scope, tune an HF SSB or CW receiver to 5 MHz and listen to the tone you get from the output jack when you tap on the unit.  It should remain rock-steady (at least from the perspective of your ears.) 

In either case, if the signal breaks up with a burst of fast static or jumps rapidly on the scope, the problem is likely in the OCXO or one of its power supplies.  If it comes across as a slow, subtle tone shift or artifact, the problem is more likely ahead of the integrator.  You could confirm this by pulling the integrator board.

Also, wiggle all of those SMC cables while monitoring the scope or receiver.  They can and do become flaky. 

[chuckb]

On my newest 5065A the 10811 coarse tuning piston cap was flakey. Sometimes it would lock, other times not. Tapping to 10811 or turning the coarse adjucstment screw isolated the problem tp the the variable cap. When I get a little free time I will open the 10811 osc again and clean the piston cap better.

[pamphonica]

Thanks for all the suggestions.

An evening spent checking everything has thrown up a few faults.

- Dry (or suspect) joints under the 1.5ohm resistors on A11.  Cleaned up and remade all nearby joints.
- Loose SMC connector under A7 module.  This connector is carefully isolated from the chassis so a bad connection leaves a flaky earth return. - tightened up.
- All SMB/SMC connectors checked.
- Also opened up the 3 lids of the optical unit and checked the 1.33K resistor inside.  It looks really good, no signs of more than moderate heating, all bands readable and measures 1.33K cold so I am happy to leave that there for now.  No fault on the 12ohm resistor soldered to the two tags on the other side of the board.  Carefully re-asembled (no sign of an "spacers" as detailed by Corby).
- Checked the 10811 OCXO trimmer shaft.  Perfect, moving freely and allowed easy setting of 5MHz.

After re-assembly, warmed up for 45 minutes (it's already in a nice 21 degree ambient) and hit go.  Green light. Currently checking the output on an Agilent 53220A graphical display, locked to an SRI rubidium.  No glitches on the 2nd Harmonic unless I really thump the case.

I'll leave it on for a few days and see if it stays happy.  I'll report back.

[pamphonica]

The other potential problem I found was in the SMB connector on the end of the cable from the RVFR.  It goes into J1 on the main panel next to the D-type connector. The outer sleeve simply rotated freely.  On dis-assembly I found that the braid had been almost completely severed, due perhaps to incorrect rotation of the outer portion when tightening the connector.  I re-made the connector in-situ, replacing the tiny central pin with a new one.  Lovely and firm now.

[KE5FX]

If yours still has the spring-loaded mechanical contraption that extends the 10811's coarse frequency control to the front panel, you should remove it.  It's no problem to reach the trimmer with an alignment tool, and the extension hardware has absolutely no business near the delicate innards of the 10811.  Big HR failure at HP.

Speaking of HR failures, the lack of failsafe measures on the RVFR ovens was another one.  All three of mine have an Arduino that monitors thermistors installed at both ends of the RVFR.  Corby's suggestion of an extra thermal fuse is a good alternative but it only accounts for the lamp oven, and AFAIK both of them can overheat due to failures in A11. 

I keep meaning to document the Arduino hack one of these days, as it's great for remote parameter monitoring as well as RVFR protection.

[pamphonica]

150 mins in and the lamp went off again.  I fiddled with a few of the connectors underneath and got a healthy 2nd Harmonic again and it fired up OK.
After waggling everything gently underneath, I'm beginning to think that the fault lies somewhere amongst all the connectors after all, possibly under or inside the  A3 module.
There is an effect where the 2nd harmonic jumps off scale if I waggle one or two of the three SMBs there.  Looks like a poor connection.  I've just tried a tiny bit of De-Oxit on the SMB connectors under A3 and it's still happy but  I may also try a temporary SMB-SMB jumper to see if that helps.  Back to the manual to make sure I understand what is connected to what!
-Jeremy

[pamphonica]

24 hours in and still green light and all readings "nominal" as NASA so loves to say.  I'm going to give it a proper soak test over a week or two now.
I suspect SMB connectors, tarnished or dirty after all these years, even though they are gold plated.  The SMCs seem fine.

I just took it offline for 30mins to wire in a flat style 4A bridge rectifier onto the baseplate to replace the toasting diodes on the A15 board.  Now back online, no problems and ready for a good soaking.  Don't you love the fact that these devices have no fan, so can just sit quietly and get on with being wonderful!

[pamphonica]

Due to space constraints, I have to leave the 5065A on its side for the soak test.
I have put spacers underneath to allow airflow into the vents.
Is this position OK for the unit?  It seems very happy so far. 
-Jeremy

[KE5FX]

Sure, it doesn't care.

[5065AGuru]

Yes the 5065A is happy on its side for extended operation (months)!

Cheers,

Corby

[pamphonica]

Thanks for the re-assurance.  Still happily sitting vertical,  everything normal and stable.  I just checked the temperature of the replacement bridge rectifier.  It's running just warm, the same as the chassis temperature.  Those poor little mushroom diodes were running at 50-60 degrees C.  Definitely a recommended mod.  The wires are simply desoldered from the last 4 pins of the edge connector and get dressed neatly to meet the new diode bridge.
-Jeremy

[chuckb]

On my newest 5065A the 10811 coarse tuning piston cap was flakey. Sometimes it would lock, other times not. Tapping to 10811 or turning the coarse adjucstment screw isolated the problem tp the the variable cap. When I get a little free time I will open the 10811 osc again and clean the piston cap better.

Just FYI
The problem was a loose nut holding the piston cap to the pcb. Works fine now.

[pamphonica]

Still working fine, although the 1MHz and 100KHz outputs stopped at one point.  I only spotted that as I was trawling round the meter positions seeing if all was well.  That was just the 1MHz board needing re-seating.

But after all this and a bit of honest thought, I realise that I really don't need such a wonderful beast so I have offered it via EEVBlog Sales initially.  Hopefully it will go to a good new home...  I still have lots of other references and no real need of such a lab-grade instrument.  And I need more shelf space for new bits and pieces!

Thanks all, I think this thread has served its purpose.

-Jeremy