Hi all,
I've been out of action for an unexpectedly long time but have finally managed to get back to audio.
I got one of these locally advertised as being in working condition and it did work, after a fashion. the controls were obviously dirty and it seemed to need a couple of hours to warm up in order to give reasonable measurements.
I initially recapped the power supply ( as it was easy to pull out the card), the thing is otherwise built like a tank; lovely quality actually.
I am aware it's not the last word in distortion measurements but I like it's tactile nature and if nothing else, the notch filter and distortion residual output can be used with my soundcard based measurement system as a good quality preamplifier
The issue is that the residual output is noisy and shows mains related modulation.
I've opened it up again to re-cap all of it and have discovered that the carbon composition resistors it uses in presumably non critical locations all seem to have drifted high, out of tolerance.
I'm therefore rebuilding the boards one by one, starting with the easiest to get at, the Rejection Amplifier PCB no. 00333-66503. The circuit diagram is on page 67 of the service manual: https://www.davmar.org/pdf/HP334A.pdf
My first question is for some reassurance, I'm planning on replacing C13 & C15 (500uF 15V Sprague axials ) with 1000uF 50v jobs as the old ones are so huge that I have space and reason that doubling the size of the output coupling capacitor and the feedback capacitor won't deleteriously affect the function of the circuit. If this were an cap coupled audio amplifier I would not hesitate but test equipment is above my pay grade at this stage and there may be subtle factors or interactions at play that I am unaware of, any comments please?
The second question is about the tantalum capacitors (C3 and C6, same page of service manual) that filter and bypass the power rails, they test fine but I want to replace them as a precaution, they are 3.9uF 3v jobs. I've read on datasheets for three pin regulators that "an output cap of 1uF tantalum or at least 10uf electrolytic may be used". Is there some sort of convention whereby a tantalum cap should generally be replaced with an electrolytic of ten times the value in power supply decoupling?
With thanks,
I've been out of action for an unexpectedly long time but have finally managed to get back to audio.
I got one of these locally advertised as being in working condition and it did work, after a fashion. the controls were obviously dirty and it seemed to need a couple of hours to warm up in order to give reasonable measurements.
I initially recapped the power supply ( as it was easy to pull out the card), the thing is otherwise built like a tank; lovely quality actually.
I am aware it's not the last word in distortion measurements but I like it's tactile nature and if nothing else, the notch filter and distortion residual output can be used with my soundcard based measurement system as a good quality preamplifier
The issue is that the residual output is noisy and shows mains related modulation.
I've opened it up again to re-cap all of it and have discovered that the carbon composition resistors it uses in presumably non critical locations all seem to have drifted high, out of tolerance.
I'm therefore rebuilding the boards one by one, starting with the easiest to get at, the Rejection Amplifier PCB no. 00333-66503. The circuit diagram is on page 67 of the service manual: https://www.davmar.org/pdf/HP334A.pdf
My first question is for some reassurance, I'm planning on replacing C13 & C15 (500uF 15V Sprague axials ) with 1000uF 50v jobs as the old ones are so huge that I have space and reason that doubling the size of the output coupling capacitor and the feedback capacitor won't deleteriously affect the function of the circuit. If this were an cap coupled audio amplifier I would not hesitate but test equipment is above my pay grade at this stage and there may be subtle factors or interactions at play that I am unaware of, any comments please?
The second question is about the tantalum capacitors (C3 and C6, same page of service manual) that filter and bypass the power rails, they test fine but I want to replace them as a precaution, they are 3.9uF 3v jobs. I've read on datasheets for three pin regulators that "an output cap of 1uF tantalum or at least 10uf electrolytic may be used". Is there some sort of convention whereby a tantalum cap should generally be replaced with an electrolytic of ten times the value in power supply decoupling?
With thanks,
"... If this were an cap coupled audio amplifier I would not hesitate but test equipment is above my pay grade..."
Don't undersell yourself: look at the input stage of the schematic you referred to: it looks a lot like any amplifying stage from a mid-fi amp of the 70's. And the build "quality" is not much different either. Sorry, HP aficionados; back in the days, I was a Tek guy.
All you comments make sense:
Don't undersell yourself: look at the input stage of the schematic you referred to: it looks a lot like any amplifying stage from a mid-fi amp of the 70's. And the build "quality" is not much different either. Sorry, HP aficionados; back in the days, I was a Tek guy.
All you comments make sense:
- Yes you can, but I'd leave the original values alone because these affect the low freq response and may reveal more 1/f noise and make the reading a bit worse. No big deal either way.
- Yes you can, tan caps last a long time, but why not. I wouldn't worry about the stability of the regulators, the 2 40uF caps on the regulator board take care of that. The only regs that really require output caps, AFAIK, are the LT1083/4/5, and they were not invented yet.
Thanks Zung, the circuit indeed looked familiar or I would have chickened out... 😀
I've put the 1000ouF coupling and feedback caps in for now while I wait for some 470uF caps to arrive as I have none in stock of the right voltage.
I have replaced the tantalum 3.9uF caps with 4.7uF electrolytics, my question was is it accepted practice to replace tantalum caps with the same value or should it be higher, that was the only reason I mentioned the regulator datasheets as this was the only place I found reference to substituting electrolytic caps for tantalum rather than concerns for the stability of the regulator as such.
I also replaced all of the carbon resistors, they had drifted high by 10-15% across the board though some were strange values and so I had to parallel two resistors to make up the correct value.
I plugged the board back in and tested the instrument to make sure I hadn't buggered anything up before moving on to another board, all good.
The residual noise and low frequency signal modulation of the residual output is unchanged but the auto-nulling feature is much improved in warmup time and speed, though the quantum of the measurement of the distortion from my reference oscillator remains the same.
I'm going to tackle the LDR/lamps portion of the circuit next, it looked a bit grim under there with foam rotting and coating the bulbs and LDRs in sticky residue. I'm attempting to find a solvent to clean them but am wondering whether to replace the rotted foam and if so, what with?
I've put the 1000ouF coupling and feedback caps in for now while I wait for some 470uF caps to arrive as I have none in stock of the right voltage.
I have replaced the tantalum 3.9uF caps with 4.7uF electrolytics, my question was is it accepted practice to replace tantalum caps with the same value or should it be higher, that was the only reason I mentioned the regulator datasheets as this was the only place I found reference to substituting electrolytic caps for tantalum rather than concerns for the stability of the regulator as such.
I also replaced all of the carbon resistors, they had drifted high by 10-15% across the board though some were strange values and so I had to parallel two resistors to make up the correct value.
I plugged the board back in and tested the instrument to make sure I hadn't buggered anything up before moving on to another board, all good.
The residual noise and low frequency signal modulation of the residual output is unchanged but the auto-nulling feature is much improved in warmup time and speed, though the quantum of the measurement of the distortion from my reference oscillator remains the same.
I'm going to tackle the LDR/lamps portion of the circuit next, it looked a bit grim under there with foam rotting and coating the bulbs and LDRs in sticky residue. I'm attempting to find a solvent to clean them but am wondering whether to replace the rotted foam and if so, what with?
Attachments
Flashheart: It’s me, Flash! Flash by name, Flash by nature. Hurrah!
Blackadder: Where have you been?
Flashheart: Where haven’t I been! Woof!
🙂
I tend leave those 150D Series Sprague tantalums in place for the few HP and Tek restorations I've completed. They seem to hold up well. You may not want to replace tantalums in the direct signal path as they might have been chosen for a specific reason (like C4 on Board A2 1.8uF, 35VDC for example). I have found several bad 30D series aluminum electrolytics in the course of my restorations so I would recommend evaluating those first.
Thanks Drj, I have come to a similar conclusion about the tantalums but it's good to have it confirmed. they measure very well with no detectable leakage and good ESR. I'll leave them in on the other boards.
The larger value and higher voltage electrolytics seem to be fine but it's the low value and particularly low voltage ones seems to have suffered with their capacitances seeming rather high and showing electrical leakage in some cases.
Photoresistors need calibrating:
Before:
V1 (30k) : 20.95K
V2 (15K) : 41.75K
V3 (15K) : 94.95K
V4 (150K) : 825.5K
V5 (150K) : 61.15K
One bulb is also noticeably dimmer than the other one.
(page 39 of the service manual above)
I swapped the bulbs around and the dimness is related to position rather than the lamp.
the lamp that serves the photoresistors V2-V5 is the dimmer one and all of the readings from that lamps are lower than they should be, I wonder if there are more issues than a bad photoresistor or two?
The figure in brackets above is the service manual expected and adjust to figure, the second figure are the actual measurements from my unit.
Before:
V1 (30k) : 20.95K
V2 (15K) : 41.75K
V3 (15K) : 94.95K
V4 (150K) : 825.5K
V5 (150K) : 61.15K
One bulb is also noticeably dimmer than the other one.
(page 39 of the service manual above)
I swapped the bulbs around and the dimness is related to position rather than the lamp.
the lamp that serves the photoresistors V2-V5 is the dimmer one and all of the readings from that lamps are lower than they should be, I wonder if there are more issues than a bad photoresistor or two?
The figure in brackets above is the service manual expected and adjust to figure, the second figure are the actual measurements from my unit.
Last edited:
So before getting much further into the photoresistor/bulb thing, I've decided to go back to basics and check all voltages first to establish a baseline.
The service manual gives the following checks to perform:
All voltages were either in spec or able to be adjusted to spec except for the bias of 2.65v at TP1 on the A3 (rejection amplifier) board. I don't think this has anything to do with the bulb issue but had completely re-capped this board and replaced all the carbon resistors.
The lowest I was able to get the voltage here was ~3V.
According to schematic:
the bias current through Q1 is set by R3 and R4 if I have understood correctly and therefore the easiest way to reduce the current is to increase R3. the R3 in my unit was actually 200ohms, rather than the 10K shown on the schematic above, I'm thinking of swapping it for something like a 560 ohms, does the idea sound reasonable or should i go for the schematic value of 10k?
The service manual gives the following checks to perform:
All voltages were either in spec or able to be adjusted to spec except for the bias of 2.65v at TP1 on the A3 (rejection amplifier) board. I don't think this has anything to do with the bulb issue but had completely re-capped this board and replaced all the carbon resistors.
The lowest I was able to get the voltage here was ~3V.
According to schematic:
the bias current through Q1 is set by R3 and R4 if I have understood correctly and therefore the easiest way to reduce the current is to increase R3. the R3 in my unit was actually 200ohms, rather than the 10K shown on the schematic above, I'm thinking of swapping it for something like a 560 ohms, does the idea sound reasonable or should i go for the schematic value of 10k?
It's worth a try but I'm wondering why R3 it is 200 ohms when it was spec'd to 10k. The backdated info (Appendix C, Change #8, page 85 of the pdf) states that A3R3 was 12k for some models. I am wondering if A3Q1 might have an issue if someone had to drop A3R3 from 10k or 12k to 200 ohms. I assume you verified that A3R3 in your unit is 200 ohms measured with that resistor isolated (at least one end of the resistor lifted from the board). Or maybe you were looking at A3R1 which is 240 ohms?
Yes, I was surprised to find a 200 ohm in that position, it was definitely a 200 ohm resistor though; red, black brown (gold). Definitely factory too, no one has been inside this unit before me.
The bias current was around 4.5 volts when first measured but Q1 was a 2n3904 from factory rather than a 2n3391 as per the parts list and I wonder if the higher bias current suits the replacement transistor?
In any event changing to 560 ohms made very little difference and putting a 10k in there allowed the bias to be adjusted to spec.
The other test voltages were all correct and with this now being correct I had another go at the photodiode/bulb calibration.
The figures are still way out, I think I have at least one faulty photodiode ( the one that reads 100s of KOhms instead of tens of kOhms) and the bulbs are possibly knackered.
I did mange to get V1 to the proscribed 30kOhms but the rest are still way off and don't seem to respond to the same degree as V1 to being moved towards the bulb.
Not sure where to get these parts...
The bias current was around 4.5 volts when first measured but Q1 was a 2n3904 from factory rather than a 2n3391 as per the parts list and I wonder if the higher bias current suits the replacement transistor?
In any event changing to 560 ohms made very little difference and putting a 10k in there allowed the bias to be adjusted to spec.
The other test voltages were all correct and with this now being correct I had another go at the photodiode/bulb calibration.
The figures are still way out, I think I have at least one faulty photodiode ( the one that reads 100s of KOhms instead of tens of kOhms) and the bulbs are possibly knackered.
I did mange to get V1 to the proscribed 30kOhms but the rest are still way off and don't seem to respond to the same degree as V1 to being moved towards the bulb.
Not sure where to get these parts...
The A6V1-V5 are type 5 CdS photocells with peak response at 550 nm.
The NSL-5160 might be a candidate for the CL905HL (A6V1, V2, V3). The NSL-5160 is available here: https://us.rs-online.com/product/luna-optoelectronics/nsl-5160/70136785/?keyword=NSL-5160. The NSL-5162, available both at Digikey and Mouser, is the non-hermetically sealed version of the NSL-5160. I have some NSL-5160 I bought many years ago but I haven't tried them in my 333A. In fact, my 333A is un-restored and is a project waiting in my "I'll get around to it some day" que.
And the NSL-5170 might be a candidate for the CL905HN (A6V4, V5) but I haven't found a source for the NSL-5170. Maybe there's an equivalent available in a non-hermetically sealed device. But devices with a 467 Megohm dark resistance seem to be nonexistent these days.
Additional info:
https://us.rs-online.com/m/d/e1e0263e6b49f734997bf8faba1a0488.pdf
https://www.technicalaudio.com/pdf/Clairex_Corp/Clairex_107_CL-900_Series_TO-18_case.pdf
https://www.diyaudio.com/community/...s-photoresistors-nsl-7540-and-nsl7550.131289/
The NSL-5160 might be a candidate for the CL905HL (A6V1, V2, V3). The NSL-5160 is available here: https://us.rs-online.com/product/luna-optoelectronics/nsl-5160/70136785/?keyword=NSL-5160. The NSL-5162, available both at Digikey and Mouser, is the non-hermetically sealed version of the NSL-5160. I have some NSL-5160 I bought many years ago but I haven't tried them in my 333A. In fact, my 333A is un-restored and is a project waiting in my "I'll get around to it some day" que.
And the NSL-5170 might be a candidate for the CL905HN (A6V4, V5) but I haven't found a source for the NSL-5170. Maybe there's an equivalent available in a non-hermetically sealed device. But devices with a 467 Megohm dark resistance seem to be nonexistent these days.
Additional info:
https://us.rs-online.com/m/d/e1e0263e6b49f734997bf8faba1a0488.pdf
https://www.technicalaudio.com/pdf/Clairex_Corp/Clairex_107_CL-900_Series_TO-18_case.pdf
https://www.diyaudio.com/community/...s-photoresistors-nsl-7540-and-nsl7550.131289/
Thank you so much for taking the time to look into this, it is unfortunately of course the higher value photocells that appear to be buggered.
In my desperation I have decided to go through all of the rest of the boards first and see if that helps. I can't get the 8.2V at A2, for example. Lowest I can get is 8.5V which is out of spec, if only by a little.
Once I've eliminated everything else, I'll start on the bulbs and photocells, I'll order those you linked and few claiming to be originals from eBay and see what arrives.
Thanks for your help by the way, it's a hobby and so taking the amount of time it needs to change all the caps and CC resistors doesn't feel entirely wasted... 😀
In my desperation I have decided to go through all of the rest of the boards first and see if that helps. I can't get the 8.2V at A2, for example. Lowest I can get is 8.5V which is out of spec, if only by a little.
Once I've eliminated everything else, I'll start on the bulbs and photocells, I'll order those you linked and few claiming to be originals from eBay and see what arrives.
Thanks for your help by the way, it's a hobby and so taking the amount of time it needs to change all the caps and CC resistors doesn't feel entirely wasted... 😀
Glad to be of some help. Following your work is inspiring me to pull my 333A off the shelf and have a go at restoring it. But first I need to wrap up a HP 8601A RF signal generator restoration that is almost done.
It looks like Digikey has 1869 bulbs available through one of their marketplace vendors at a reasonable price. https://www.digikey.com/en/products/detail/wec/1869/14769702#mktPlaceViewSection. Digikey P/N 289-1227-ND. They are not the "D" version though but they might work. The D version I found are expensive.
It looks like Digikey has 1869 bulbs available through one of their marketplace vendors at a reasonable price. https://www.digikey.com/en/products/detail/wec/1869/14769702#mktPlaceViewSection. Digikey P/N 289-1227-ND. They are not the "D" version though but they might work. The D version I found are expensive.
After replacing all the resistors/caps and setting the voltages, I have the new resistances above in red. This gave me some hope that not all the photocells are bad and so i had another go moving the photocells around in the cover.
I was able to adjust V1 to 30k as per the service manuals but the others, i.e. all around DS2 are not able to be adjusted to spec. DS2 is also still noticeably dimmer than DS1, even after the refresh. Measuring the voltages going into them , the DS1 has 16v on pin 8 as per the service manual spec whereas DS2 is reading 18.5v on pin 6.
I reasoned that the higher voltage was due to a lower current draw by the bulb and swapped them around, no effect. The dimmer bulb stays as DS2.
Given that the the photocells need more light to show a lower resistance and and 3 out of four (V3,V4,V5 but not V2) all seem to indicate they need to be closer to the bulb, I think the fault may be related to the circuit driving DS2.
Tracing it back to the lamp driver transistors, I have correct voltages on the bases and emitters of Qs6&Q14 but 2 volts higher on the emitter Q14, which is the channel with the dimmer bulb.
Worth pulling the transistors and testing for gain? It would mean disassembling the swinging board mechanism again... 😀
Last edited:
Pulled Q14 and it's analogue Q6 and tested them with the DCA.
All the results were broadly similar except for the gain vs Ic graph, the transistor from the leg with dimmer bulb is the lower, less flat set of traces.
Bad transistor?
What sub for a 2n3391, C945?
All the results were broadly similar except for the gain vs Ic graph, the transistor from the leg with dimmer bulb is the lower, less flat set of traces.
Bad transistor?
What sub for a 2n3391, C945?
Swapped the transistors around and the problem did not follow on the respective leg, perhaps it is a bulb issue after all.
Ok... not sure what happened but as I was poking around to confirm voltages the issue somehow seemed to correct itself... The transistors remaining in their transposed positions I started probing the voltages of the components upstream of them. All voltages appeared as per the schematic and then I noticed the bulbs were now equally bright.
It seems there must have been some sort of loose connection or bad solder joint that is now good and I was able to calibrate the all photocells to resistance spec apart from V5. That one showed a low resistance (~60k) even when moved as far away as possible. It, of course, being one of the impossible obtain ones I had a flash of inspiration and coloured the front lens in with a black marker until it balanced. It still seemed to respond to light level change like the other one and so i put it back together. The instrument seems to work now, much better than before and with a much reduced warm up time.
I'm going to make some noise measurements and compare the noise floor to before next, just need to build a shielded 600 ohm load.
It seems there must have been some sort of loose connection or bad solder joint that is now good and I was able to calibrate the all photocells to resistance spec apart from V5. That one showed a low resistance (~60k) even when moved as far away as possible. It, of course, being one of the impossible obtain ones I had a flash of inspiration and coloured the front lens in with a black marker until it balanced. It still seemed to respond to light level change like the other one and so i put it back together. The instrument seems to work now, much better than before and with a much reduced warm up time.
I'm going to make some noise measurements and compare the noise floor to before next, just need to build a shielded 600 ohm load.