Testing CA741CT op amps?

[Chris Wilson]

I have two Farnell linear power supplies, an H60-25 and an H60-50.

One is working, the smaller one, the large one is missing a power supply board and has another issue.

There's a control board PCB that's exactly the same part number and revision as the one in the working H60-25. I decided to swap it over into the working PSU in order to test it. Using this other control board I get near zero volts out from the known good H60-25 PSU. I visually compared the two boards for obvious issues, then went in on the component level. I swapped out all electrolytics as a matter of course, for new ones. I removed and checked all diodes and Zener diodes and found an immediate issue. D1 had been changed to a 3.3 V Zener diode and it should be a normal rectifying diode. The other diodes and Zeners checked out fine under test. I also removed all the semiconductors bar the op amps. these too checked out fine with a Peak semiconductor tester. Resistors and pots gave similar readings and were the same value across both boards. I changed the Zener wrongly fitted at D1 for an IN4007 and tested again, but still no control of the supply. I am now thinking something's happened to an op amp, maybe as a result of someone fitting a Zener at D1, maybe coincidence. I am not good enough to work out what fitting a low wattage 3.3V Zener at D1 might do.... I wouldn't know how to start testing an op amp, in or out of circuit, and there are 3 off CA741CT used. I swapped one over between boards. The "good" board kept on working. The "bad" board remained inoperative. It was IC1 that I swapped. I managed to lift a trace doing this, so if there's a none substitution method I'd like to try testing the other two. Is this possible please? The schematic for the board is at:

http://www.chriswilson.tv/schematic.jpg

I'll try attaching it as well, but it may be too big.

I have a current limiting bench power supply, digital multi meters, a scope and a sig gen and frequency counter, but lack in depth know how! It should be easy with a working board next to a dead board, but I am don't want to do any more damage to the elderly PCB's taking op amps out needlessly. Thanks for any advice.

[Herald1360]

Hmmmm. If you can disconnect the PCB from anything it drives, leaving the +/- 15V dc supplies connected, try shorting the + input of a 741 to 0V, the -input to output then check that the output sits within a few mV of 0V. Open the -input to output short and the output should swing to one or other rail or thereabouts.

If that works OK, there's not much wrong with the opamp.

TBH only IC2 with a direct off board input looks capable of being damaged by other faults, the others probably have enough resistance in series with I/O to protect them.

[PETERg0rsq]

The very first test I make when testing OP-Amps (after checking supply rails) is to see if the OP-Amp is in a "balanced State".

An OP-Amp always wants to see its inputs equal in Voltage.
When the inputs are of equal voltage, then the output will be a voltage anywhere between +ve Rail, and -Ve Rail voltages.

If the inputs are at different voltages, then the output will be either fully ON or Fully OFF.

If the -ve input is more +ve than the +ve input, then the output will be the -ve rail voltage, and if the +ve input is more possitive than the -ve input, the output voltage will be the +ve supply rail.

Testing an op-amp is easy, if you can totaly isolate the input terminals from the output by lifting other components, by using the above information.

Peter

[Chris Wilson]

OK, thanks for the info, I wonder if measurements I have taken where this PCB slots into an 18 pin socket using PCB track "fingers" to carry the signals from and to the board. Some are radically different between the working board and the none working board.

I have altered the schematic to make the pin out numbering clearer. Please see the one attached below and the modded one still at http://www.chriswilson.tv/schematic.jpg

On the working board I see relative to the 0V rail on pin 5:

1: -15.5V
2: -39.7V
3: .017V
4: 6.27V
5: 0V (negative lead of meter here for all measurements)
6: 1.55V
7: 25.23V
8: 15.22V
9: 0.0V
10: 8.64V
11: Blank, not used
12: 6.28V
13: 0.0V
14: -.348V
15: 0.0V
16: Blank, not used
17: 8.63V
18: -0.347V


On the none working board :

1: -14.98V
2: 0.417V
3: 1.47V
4: 6.22V
5: 0V (Negative lead of meter here for all measurements).
6: -9.23V

All the rest of the voltages were pretty much identical to the good board.

Measurements done with the supply set to give 40V out in both sets of tests, which the good board held steady.

Voltage out from the PSU on the bad board was under 0.5V, with the settings left the same (40V out set).

So pin 2 was -39.7V on the good board, but 0.417V on the bad board, pin 3 was .017V on the good board, and 1.471V on the bad board. Pin 6 was 1.55V on the good board, but -9.23V on the bad board. Does this mean anything worthwhile?

I also measured ACROSS C3 (4.7 uF 63V) and on the good board saw 39V and on the bad board 1.4V.

[Chris Wilson]

Here's the overall schematic diagram showing interconnections

[cmjones01]

We should be able to tell quickly whether an op-amp is faulty by looking at the voltages on its pins. If you can tell us the voltages on pins 2, 3, 4, 6 and 7 of IC1, IC2 and IC3, then looking for the balance condition, as Peter says, is the best test.

Chris

[dseymo1]

It may not be very elegant, but if you have any doubt over the op-amps' condition, you could just replace them anyway. They're only 741s!

[Chris Wilson]

I have taken the route of swapping them over, luckily with no further damage to the PCB's. I have to assume they are all OK as the working board still works, and the failed board is still not working. As an aside, is there any reason not to hard wire in 8 pin DIL mount 741's instead of TO5 can types? They 741CT as fitted is quite dear and few seem to list them nowadays. Thanks for the help, not sure where to go now, I felt sure it must be an op amp problem.

If there is a middling voltage on the + input and a bit of resistance in the - input shorting - to output should give what is on the + on the output.

[Chris Wilson]

I think I am getting nearer.

On the working board, measuring voltages between the 0V middle rail and the cathode of D3 I see 1.3V

On the none working board I see -9.7V

On the anode of D3 I see 2.1V on the working board, and -8.6V on the none working board.

On D1 the none working board shows 0.0V on anode and cathode relative to the middle 0V rail. But on the working board I see .003V on the cathode, and .64V on the anode.

I have put some measurements from the NONE working board on this schematic below. Cheers.

[Chris Wilson]

The fix was found when i measured the voltages on all the pins on all the op amps and compared the voltages with those on the working board. I found IC1 was getting no signal to pin 2 and the reason? A microscopic crack in the pcb trace to the pin, needing a magnifying desk lamp to see. A wire bridge had the board working perfectly again. Thanks for all the help guys!

[threeseven]

Well done Chris, that was a tricky fault to find! I have a couple of H60-25 and H30-100's myself and they are excellent units.