Electronics > Repair
HP 3478A Multimeter - Parts for overvoltage protection circuit
floobydust:
Isn't the IDSS important? I see it as lost current as the output stage is after the CCS, so it would be corrected for?
Weakness in these circuits is they wrongly assume there will never be a -ve overvoltage seen, only +ve which CR201 blocks fine.
Kleinstein:
Idss is not important for a JFETs used as diode. The current that is lost is the gate leakage current. Even here it is only the change in the current (e.g. with temperature) that really matters. Any constant part would be included in the calibration.
R_on may be a littler as a small part of this will be in series to the diodes. Still the JFET as diode will only see a low current (e.g. 1 mA range) from the 240 K resistors.
Of cause there is also protecton against a negative voltage. In this case Q205 will start conducting and Q201-Q204 will more or less turn off, so that most of the voltage will drop at the BJTs and not get to the MOSFET for the regulation.
What may be a problem are fast transients from positive to high negative, as the transistors may not be that fast to turn off. Also CR201 may not be that fast blocking.
11mrjones:
Thanks for all the great advice everyone. So I'm going to:
1. inquire about orig 1855-0298 (Sphere Research, if available). per pqass
2. order a couple 2N4117 and 2N4119 (Digikey). per Flooby and Wallace
3. Try BC jcn of BC547 (I actually have some of these) and look for some low-leakage diodes on DIgik to try for kicks per Kleinstein.
Background fyi:
The 3478A was bought on the bay and seemed to be working well. A while late I noticed for the 3K ohm range, at approx 2.5-3.0K ohm load the meter read low by about 10% and was not able to successfully finish cal on the 3K ohm range.
On the 3.0K ohm range for load resistor = 0, 1, 2, 2.2, 2.5, 2.8, 3.0 K, I measured 1, 1, 1, 0.99, 0.95, 0.93, 0.85 mA, respectively, for the ohm current source (the meter read the resistor value low approx by the same %). All the other ohm ranges worked perfectly, with the current source constant over 0 to fullscale.
A shorting clip applied between Q201 emitter and CR201 anode restored the ohm current source to 1mA across the whole 3K ohm range and proper working order and successful cal. I'm guessing one of the transistors is not completely bad but is borderline, but I'm no expert :-// .
In the meantime I replaced the 4 Rifa caps, the 4 big electrolytic caps, and battery (temporary battery in place during R/R so as not to loose cal data). So I should be good to go shortly.
It will be a while till I get some parts but I'll let all know how it all ends up. Again big thanks for everyone's help, such a great community here.
Victor
Wallace Gasiewicz:
I had a very similar experience with a 3456. The high ohm readings were bad. There was a big low leakage diode in the 3456 over voltage circuit and that diode was bad. The Service Manual was very helpful.
I think someone had connected this unit to lots of volts while in resistance mode.
There were other FETS that I replaced and HP used FETS at several places as clamping diodes. I also replaced the Pre Charge FET., Used 2N4117A. Now I have two 3456 that agree with one another and have been calibrated to a volt standard that I have had professionally calibrated. (I bought it from Sphere)
pqass:
@11mrjones FYI:
You can extract the calibration data from the 3478A via a GPIB adapter; either commercial or even an Arduino UNO-based (AR488 project).
See here and follow the links.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version