-
Hello,
Inspired by everyone on this forum I've managed to get my hands on a reasonably priced 3456A sold as "powers up" ... apart from the fact that it's as big as a small caravan everything seems to be working with the exception of the AC (and AC+DC) ranges which both just show an overflow condition.
Running through the service manual it looks like the AC Converter is at fault (A40) and specifically Q2 (a dual channel JFET) and/or U6 (an op amp) ... neither of them seem to be working properly, but I am probably pushing my capability in really trying to understand this properly.
So Q2 both sides seem to be switched off (pins 1 and 4 are at nearly -15v), so both inputs to U6 are at the same potential (-14.8v) shouldn't that put the output mid rail ... i.e. 0v?
If I follow the manual and short Q2 pin 6 to U6 pin 6 and ground Q2 pin 3 then I'm supposed to see nearly 0v at pin 6 ... but I get in excess of 10v.
Q3 pin 3 is at +8.6v ... which I'm really confused about ... Q5 and Q9 should be on ... so shouldn't that mean Q2 pin 3 should be pretty much 0v (plus a diode drop) ... or is Q2 a bit different? Or that that indicative of a failure situation? Or do I just not understand?
I think I've been staring at this too long now.
Anyway, any help would be appreciated, but my main question is whether I can replace those devices with...
(a) for Q2 ... could I just use two JFET's? Or a different dual channel version ... those things look hard to find and are very expensive (at least to get to the UK)?
(b) for U6 ... similar question can I use something more readily available?
Many thanks for any help,
Lee. -
Having nearly -15 V at both OP inputs points to a problem with the JFETs (or possibly the OP shortening towards -15 V). I don't know the type of OP used, but it is well possible that -14.8 V is outside the common mode range and thus the OP has any right to refuse working.
At least for a start one could replace Q2 with 2 reasonably matched JFETs and suitable (likely relatively small), as dual JFETs are hard to get and expensive. There is already some offset adjustment, so no need for a perfect matching. Anyway the offset would need adjustment.
I am not sure how fast U6 is, but it might be a good idea to have a socket for U6 anyway, so one could test possible replacements.
There are some cross ref. tables to show what OP is behind those HP specific numbers.
However there is also the option replace U6 with an more modern JFET based OP and just replace Q2 with bridges and remove R17/R18. Modern FET OPs are easier to get than the 2 JFETs to make the old workaround circuit from times, when there were no good (low noise) JFET OPs. Something like an OPA134 could be a possible replacement JFET OP. It could also work for just U6. -
Already checked power supply
? Stupid advice, I know 
. But most (like 90% or so) of the 3456As I fixed only had bad PSU caps.
HP3456A are built like a tank and have lots of protection on the input, thereby parts deep inside die rarely.
If the Opamp is output is low, it also might be because it has bad positive supply. -
Kleinstein you are a star -- thank you!
Your comments helped me to better understand what was going on ... I had completely the wrong idea before.
Anyway, I was convinced that the Q2 gate voltage was the real problem here so I decided to follow your advice (kind of) ...based on your comments my assumption now is that the dual jfet is basically a voltage follower to buffer the inputs to the op amp (U6) presumably because it's not low impedance enough, therefore if I just remove the dual jfet and connect the two gates directly to the op amp (removing R17/R18 as you suggest) then I should at least be able to test the theory ... if that works then I can either look for new jfets or do as you suggest and replace the op amp with something better.
Success!
Bypassing the Q2 it all seems to be working! Therefore I'm very sure that Q2 was the problem, and the op amp seems to be fine, so I'm now a bit torn between putting a socket in and trying some different op amps or just trying to find some jfets .... any advice and suggested components here would be really welcome (I will order a couple of OPA134's anyway.)
I also seem to have some intermittent behaviour on the ACV+DCV range where it seems to lose the input signal ... I half think it's the relay, but it's not consistent enough for me to properly test at the moment ... more time and patience needed.
MadTux -- thanks for the steer, but yes I already checked the power supply ... all fine.
Thanks for all the help.
Lee.
Edit: just found the description of the circuit in the manual ... doh. rtfm. -
There is still the possibility that the JFETs are still OK, but have only drifted too much so that the offset adjustment is off too much.
So it might be worth to test the JFET pair. It might still work.
The extra JFET pair and likely BJT based OP are likely used because of low noise. So it would be about reasonably fast low noise OPs with not too much input capacitance. So JFET based good audio OPs (e.g. OPA1641 (or similar OPA140) or the OPA134) are candidates. Chances are the OPA134 is already lower noise than the dual FET. I think the input is not super low noise anyway, due to the 1 M resistor at the input.
-
So I put a socket in today and tried a couple of different op-amps (including the OPA134) to no avail ... in each case I just got over "overload", it looks like I was getting oscillation on the feedback pin.
Eventually I put back the original (still without the dual fet, and with bridges) and it's back to working ... it actually seems pretty good!
The relay for the AC+DC range is definitely not working, and they also seem difficult to get in the UK, so I may have to get creative on that one.
I actually think I may have damaged the dual-fet while I was trying to diagnose the problem ... thinking back, right at the start I did get good AC readings on ACV, but didn't get anything reliable on ACV+DCV ... after a bit of diagnostics I was just getting overflow, so I suspect I damaged it then.
So again I'm a bit stuck ... do I just leave it without the FET's? What's the implication? It seems to be reading as well as my Keithley 2000? Or do I try to find some FET's to put in there ... if so, where do I start!
(And if anyone knows a source for the relays in the UK, or a sensible alternative then please let me know (0490-0683)
Edit: How about these? http://uk.farnell.com/cynergy3/s2-05e/relay-reed-spst-no-350vdc-1a-tht/dp/2480006 They should fit reasonably well, with only one lead needing to be messed with.
Thanks,
Lee. -
For the choice of JFETs one can start with the operating current: R17/R18 set the current to about 200 µA, as the voltage over the resistors should be around 16 V. For the JFETs there is a typical current for which the temperature drift is minimal. From the current one could find suitable JFETs. My first guess would be something like U401, but 2 separate JFETs could work too, as DC drift is not that critical. New fets would need a new offset adjustment, especially with 2 not so well matched ones.
If the OP is too fast or slow, this could lead to oscillation. An extra cap (e.g. 10-100 pF range) between pins 6 and 2 should be OK to slow down the OP. It might help to know which OP is the original. At least to know the GBW. It could be a rather unusual OP (rather fast or slow), as at the time moderate speed OPs like the LF356 where available at the time.
Another possible trouble source could be the offset adjustment - though this should be with all OPs, not just the new ones.
The downside of using the likely BJT based OP would be additional current noise of the, that could be significant in the 1 V (and smaller if there) range, because the input is high impedance in this case.
-
If the OP is too fast or slow, this could lead to oscillation. An extra cap (e.g. 10-100 pF range) between pins 6 and 2 should be OK to slow down the OP. It might help to know which OP is the original. At least to know the GBW. It could be a rather unusual OP (rather fast or slow), as at the time moderate speed OPs like the LF356 where available at the time.
Seems to be a variant of the HA-2625. GBW is 100 MHz!
HP were aware of LF35X OPs - the circuit in question actually feeds an LF357 so HP must have had a reason for the dual FET buffer/high speed OP.