HP5300B wrong DC rail voltages

[Uncle Bulgaria]

This has floated to the top of my project pile, after getting it from another Member a little while back. I'm a bit stuck with the repair, and with the testing results I'm getting.

The power supply made a loud whine, which the troubleshooting flowchart suggests is a problem. I have been following its instructions, so far without success.

I've attached the troubleshooting flowchart. The manual is too large to attach, but can be got from Keysight here, or a Dropbox link here.

I have a model with the TCXO (circuit board 05300-60020), serial prefix 1707. The manual shows boards up to 1704, but it's the best I have and they look very similar. I'm using the Parts Location diagram on page 8-5.

One coil (L2) was unsoldered at one end, and C18 was missing. I have replaced it with a measured good Vishay part. C23 had failed, so I have temporarily replaced it with the only one I had of the same value.

Connecting a Thurlby K1 (6V max. 3.5A) to C6 as per the flowchart resulted in a chirruping noise, low +/-5V, OK +/-17V and a small, but clean, square switching waveform at the Q17 collector.

The Thurlby decimal points were flashing as if it were hitting current limiting. I put an AVO8 in series for current measurement as the flowchart specifies a 2A supply. In circuit, it reads 0.8A and the HP3500B suddenly produces a consistent high-pitched squeal instead of the chirruping.

I'm not running it for more than a few seconds at a time, in case this is the 'whine' the flowchart is nervous about. However, with pins 25-50 linked, under these conditions the rails (nominally +/-5V, +/-17V, +3.5V) read:

+4.9V, -5.5V, +18.6V, -18.7V, +4.1V. This isn't bad for the specified tolerances, but there's no switching waveform.

I have checked the diodes on A2, as well as in the PSU (CR2, CR3, CR4, CR11, CR12).

Where do I go next, please, HP gurus?

[factory]

Intermittent overload on the +/-5V rails? I've just noticed the TXCO also uses those supply rails, the diagrams available are quite poor quality.

The previous thread on the 5300B counter ended with the +5V supply causing problems.
https://www.vintage-radio.net/forum/....php?p=1430948

David

[Uncle Bulgaria]

You know what, I think it's the very same 5300B!

[factory]

It sort of ended with possible transformer issue, but having seen the picture made me check where that TXCO was connected to and yes it is powered directly from the +/-5V rail, removing the lower half (input module section) doesn't disconnect the TXCO.

David

[Uncle Bulgaria]

I've checked with Mole42UK and it is the same 5300B, which gives us some more diagnostic information! Checking continuity on the transformers shows nothing amiss.

Taking off board A2 to remove the regulator from the equation, I have tested for continuity to ground from each of its 12 contacts. Pin 7, the +5V has a voltage drop biased both ways.

Removing the TXCO and lifting the anode of CR2 there should be no path to ground, if I've read the diagram correctly. Instead there's a slowly increasing resistance starting at about 1.8k. I'll try removing C24, a tantalum and testing it.

[factory]

If I read the book correctly, it mentioned not going over 6V for testing the PSU with the A2 board removed, as the over-voltage protection is part of that board.

As you've got the TXCO removed it might be worth testing that on it's own, with a separate current limited supply, if it does prove to be the problem, looking for a modern replacement TXCO shouldn't be too hard.

Going back to the standard crystal (non TXCO) might be more tricky, as it requires an obsolete MC10109P logic IC, would probably need to be creative with two modern ICs & some mod board.
Edit: the MC10109P seems to be available on ebay from several suppliers.

David

[Uncle Bulgaria]

The A2 board has a couple of links in so that Q17 doesn't switch without it - it's not possible to check the PSU without the regulatory A2 board in place.

I have found two resistors (R11 and R12) substituted for each other on A2. The 'parts location' page shows them incorrectly on the board - buzzing out the connections and comparing to the circuit diagram showed that R12 was electrically in the place that R11 should occupy in the circuit diagram.

Changing that so they match the circuit diagram drops the current to a steady 500mA rather than heading north of 2.6A, and there's a 6V switching waveform on Q17, but the +5V is still low at +2.5V. The manual section 4 in the 'theory of operation' says how the +5V dictates the others as a reference, defined by the transformer T2. I fear the transformer is at fault. The chirruping noise is coming from T2 - perhaps a harmonic resonance of failure. However, unless it's peculiarly intermittent, evidence against this would be that the other voltages are so close - surely they couldn't be if the +5V was short since that would break the whole secondary side and there'd be no output beyond an immediate short circuit.

The TXCO is out of circuit at the moment, so I don't think it's part of the problem. I think this is all contained between C6 and the interplay of the PSU and the A2 regulator board.

[factory]

Your right about A2 in the 5300 "B" version, the caution I read was for the "A" version where it mentions connecting pins E & F together with A2 removed and to not exceed 6V input.
And looking at my 5300 "A" version the A2 board is different to what you have, your "B" version A2 board should be 05300-60019, mine ends in 60003.

I wonder how it ever worked with resistors swapped.
There is +2.5V the other side of CR2 from the +5V rail.

I've re-read both threads and it seemed to be intermittent at the end of the other one, not sure what to suggest checking next, maybe the solder joints on the connector for the A2 board. I remembered someone having trouble with the little plastic cased transistors being intermittent in another HP item a few years back on here, can't tell from the picture if it uses those on the A2 board, my "A" version certainly does (picture in the other thread).

I might have to check my "A" version with the module removed, to see if it chirps with no load, won't be able to check till next weekend though, need to dispose of two RIFAs at the same time (2028A s/n prefix).

David

[factory]

Well I've set my 5300A up to run on +6V, it also chirps & gives strange low voltage readings on the output test points.
But increasing the input voltage a bit, it stops chirping at around +6.8V (354mA current draw measured with DMM) and gives more sensible readings on the output.


This makes me wonder if some of the resistors on the regulator board have drifted with age.

David

[Uncle Bulgaria]

That is interesting! I went one further and put 22V across C6, as if it were being powered from the bridge rectifier in normal use. Current draw is (with TXCO removed) 160mA, +5V is spot on, -5V about -5.2V, the 17s are OK and the +3.5 is around 3.7V. It's also silent.

All this rather makes me wonder why the flowchart asks one to check the rails with 6V applied, as it doesn't seem to give the switcher enough juice to function!

Yes, my A2 board is 05300-60019. Now I need the right fuse and a suitable replacement for C23...

Do you have an idea how I could test the TXCO while I have it out of circuit? Say plug 5V on it and see if it oscillates?

[factory]

The TXCO needs a dual supply, +5V & -5V, not sure why as they don't seem to cover the TXCO in the circuit operation section, probably because it's a bought-in module. And yes check for correct output.
Edit : could it be a negative output?

150mA 6.3x32mm slow fuses seem easy to find, the Sprague 150D tant capacitor 0.68uF at 35V might be a little trickier to find, none in the UK on ebay, I could check my parts boards to see if I have any.

I checked the resistors on my A2 board, they were all OK, the two carbon comp ones are right on the limit for the 5% tolerance (both have drifted up).

David

[Uncle Bulgaria]

Jumping the +/-5V supplies from the board (no regulators to hand) and reading off the remaining pin got rock steady, evenly spaced oscillations on the oscilloscope. The period was ~36µs which suggests kHz rather than the 9.9999984MHz on the lid... However it was steady and nothing smoked so I'm tempted to infer measurement error. I'll put it back in circuit and see if the whine returns.

[Uncle Bulgaria]

Well, supplied from 22V across C6 with the 5302A attached, the display lights up and there is no whining from the power supply, just a quiet but very insistent, very high-pitched noise rather like a flashgun capacitor at the end of its charge. It might lessen with the case on...

Any suggestions of tests to see if it's functioning properly? The display changes its units as the knobs are operated, so the display appears to be OK.

I was going to get a film capacitor to replace the tantalum.

[factory]

I gave my 5300A a quick check at the weekend after removing the RIFAs, I used the 1MHz oscillator output on another counter to check the ref was OK (need something like an off-air standard, or GPS disciplined oscillator for properly checking cal).
And a sine/square oscillator, or function generator for checking the various functions, something that go up to 50MHz would be useful too.

The 209A I used to check all digits were counting/displaying correctly*, seems to get picked up on the AM radio, that can tell me if the frequency is correct too.

*this is quite important for older number tube based counters, as the decoder/driver ICs or transistors can fail, giving incorrect or missing numbers.



David

[Uncle Bulgaria]

I've put the confetti back in the box - setting up to check the accuracy just now, the 68.1V zener diode CR10 started smoking.

[factory]

That's not good news, I seem to remember CR10 was originally a 0.4W part and got replaced with a regular 68V 5W part in the previous thread, has it shorted again?
Looks like were back to trying to find the cause of the intermittent overload.

David

[Uncle Bulgaria]

Yes, you're right. The zener's shorted. With one leg lifted I get ~25 ohms between primary and secondary of T2, which rises as if charging a capacitor. I guess I'll have to take T2 out and test the primary/secondary isolation.

[factory]

Check the resistance with the A2 board unplugged first, before removing T2.

David

[Uncle Bulgaria]

I'm sorry, I should have specified - this is with A2 removed. I've worked out the 1-8 numbering of the T2 terminals. Between 1-2 is continuity as expected, and between 3-8. With CR10 out of circuit and A2 removed, I get the resistance results above between primary and secondary of T2.

[Uncle Bulgaria]

I've removed T2, and get no continuity between primary 1-2 and the other secondaries 3-8. I can imagine this could change under load. What would be a good test method for checking it under more likely load conditions? A high-frequency input to the primary from a signal generator?