Op Amp as voltage regulator?

[Station X]

Update.

I measured the resistance between TP18 and ground suspecting that the 10uF cap, which is a tant, had gone circuit. This was not the case as I measured a very high resistance.

Then I looked around the board trying to find an undocumented resistor which might have gone open circuit, but found nothing.

At this point I realised I had reverse engineered the circuit incorrectly. The track to an unknown point goes to pin 3 of the Op Amp, not pin 6.

I traced the track visually and using an ohmmeter. It went to a pin on a 50 way connector which is connected to a ribbon cable bus serving all the PCBs in the SG. Documentation showed that the pin in question was allocated to an UNUSED interrupt line. This seemed unlikely! There was no connection from the ribbon cable pin back to the processor. Clearly this wire of the bus was being used for something else.

Measuring the resistance between ground and the plug of the ribbon cable showed about 400K. By disconnecting the ribbon cable from each PCB in turn I proved that this 400K was on the LF Synthesiser PCB.

Access to this PCB is very restricted without major dismantling, but I found that the ribbon cable wire in question went to an undocumented 4.7K resistor. The other end of this resistor is wired to one end of the track of a pre-set resistor R11. I suspect it's the right hand tag as shown on the attached extracts from the circuit. This tag is the wired away to (you've guessed it!) an Op Amp and a zener diode designated IC2. This is obviously the reference voltage source for the DACs.

I think I understand it all now except for the diode on the mother board.

[Mooly]

I'm sort of losing where you are up to with this now

Your first picture above shows one half of a dual opamp (TLO82?) connected as a unity gain follower. So whatever voltage is on pin 5 is buffered and appears exactly the same on pin 6

The next picture show another two opamps (another dual) in inverting mode. Looking at the right hand one... the gain is set by the ratio of the resistance "to the left of the wiper" lets call that Rf and the resistance "to the right" of the wiper", lets call that Rin. If you had two 5k fixed resistors then the gain is -(Rf/Rin) so a gain of minus 1. So applying 3 volts input would give -3 volts output. The opamp to the left is the same but with an offset depending on the voltage on pin 3.

The golden rule of the output doing what is necessary to keep the difference between the inputs at zero applies in all these cases.

Not sure if all that helps or not but it should enable you to check and confirm voltages and whether the opamps are behaving.

So which is the diode on the motherboard you are referring to ?

[Mooly]

Just refreshed the page and the pics seem in a different order

[Station X]

Ignore all the Op Amps except the one in picture 3.

The others are only included to show the connection of the reference voltage to several of the Op Amps on THIS PCB, ie the LF Synthesiser.

You will see that the reference line is connected to the right tag of R11. Connected to the same point is a 4.7K resistor which is connected via a PCB tracks and ribbon cable to pin 3 of the Op Amp on the motherboard.

To summarise it, pin 7 of the Op Amp of the LF synth PCB is connected to pin 3 of the Op Amp on the motherboard.

The diode on the motherboard is the one shown in post #17.

--------------

I edited one of the pictures which changed the order, so I reloaded them all in an attempt to clarify matters.

-----------------

I still have some way to go as there's obviously a problem with the LF synth PCB or its interconnections to the mother board.

Thanks for everyone's help so far.

[Mooly]

OK This sure is confusing...

So picture #3 (for me) is the one called "IC2". So on that you should have the desired reference voltage on pin 5 and an identical voltage on pins 6 and 7

Picture #2 called "R11" should take that reference voltage and invert it with the voltage appearing on pin 7 of that opamp. If the preset is not centred then the voltage on pin 7 will be either higher or lower (preset alters the gain) than the applied reference voltage.

[Mooly]

Your diode in post #17 shows a clamp that will prevent an applied input voltage to the opamp going higher than the 5 volt rail. So normally it would do nothing. It also serves to rapidly discharge the 10uf cap on power off by discharging it into the collapsing rails.

Here, I suspect its intended function is a clamp rather than rapid discharge.

[Station X]

I can see I'm going to have to put this all on one drawing. The Op Amp in the drawing "R11" is only there because R11 is connected to it. The right hand end of R11 is connected to the voltage reference line.

Also connected to the right hand end of R11 is a 4.7K resistor (not shown). This is connected to pin 3 of the Op amp on the motherboard, providing a reference source for the DACs.

[Mooly]

Originally Posted by Station X

I can see I'm going to have to put this all on one drawing.

OK, that might not be a bad idea actually and then we can pencil in the expected voltages.

[Alistair D]

The fact that you measured 400k on the bus cable kinda rules out a wiring fault and a short on the output of IC1b. The 330 ohm resistor should be easy to measure. That only leaves IC1 or IC2.

Looking at IC1, it is powered from +15 so the clamp diode is probably to protect the TLO81 in the event of a fault.

Al

[Station X]

Here it is on one drawing. I'll try to identify the diode. It looks like a zener.

I'll have to reinstall the mother board in order to test the SG and that means I'll lose access to the track side of the MB. I hope the fault is on the LF Synth board.

It would be useful to have a break out 50 way ribbon cable to gain test access to the bus. I'll have to have a look in the attic.

[Mooly]

So lets see what we have... and what do you actually measure ?

as drawn you have what appears to be a plus 2.45v reference in IC2. That 5 volt rail on the 330Ω (R1 point A) is plus 5 volts, not minus 5 ?

So what voltage do you have on pin 7 of IC1b ?

The output feeding into the 4k7 and then the diode doesn't look quite right tbh unless the diode and 4k7 is the clamp for the lower opamp input (TP18). If so then the output of IC1b would feed off somewhere else too.

To summarise... as drawn the top IC is self contained and generates a buffered voltage at its output identical to the voltage on IC2.

The top IC has no influence on the lower as drawn. The diode and 4k7 could be a clamp but its unusual to clamp to an IC output... but if it really is as drawn then it would work as such if the source impedance feeding TP18 was relatively high.

The 100Ω and 1k look a little odd as drawn out for a couple of reasons. Might be worth just having another gander at that

(I appreciate its a hugely complex item and can imagine access is well nigh impossible)

[Station X]

The feed to the 330R resistor is from the +5V rail.

The diode has the markings
HP2
7D4

I've attached a couple of pictures of the mother board so that anyone who wants to have a go at reverse engineering the circuitry around the Op Amp can do so. I won't give any guidance so as to avoid leading you along the wrong track. No pun intended!

I won't be in a position to take any voltage readings until I've reinstalled the mother board which is a major job.

Thanks.

[Radio Wrangler]

Hi Graham, that whole area seems to be a bodgefest. One gotcha with zener diodes: always take a peek at the data sheet for the part being used, some of them only have their voltage specified at quite large currents can be 20mA on lowish voltage ones. Some less common diodes would be OK at 5mA.

With opamps, beware of output current ratings.

It loks like Racal took several versions of fiddling in this area rather than doing it properly from the start.

David

[Mooly]

Hmmm...

I wonder if the mods were to stop instability around the opamp. Many opamps oscillate at around 10 to 100MHz when loaded capacitively (even a few tens to hundreds of pf) which could be from PCB traces or the input capacitance of the devices it feeds.

If you suspect the opamp then its as quick to replace it as wonder. The TL071, TL081 are available anywhere. The plated vias mean removing it could be tricky. I would cut it out from the top and then remove the lead fragments finally running a piece of stainless wire through the hole to clean it up.

[Herald1360]

PTHs can be cleaned out neatly using a solderpump with a bit of rubber (Hellermann) sleeve over the nozzle and sticking out a couple of mm. This both makes a better seal and reduces the effects of Newton's third law on the pad

Hold the board vertical (third hand or whatever) and apply soldering iron and primed pump to opposite sides of the via. Melt the solder and wait a second then trigger the pump. Job's a good'un.

[Station X]

Thanks for all the advice on changing ICs. To be quite frank with everyone, I lack the vision and manual dexterity to attempt such a job. If changing an IC is necessary I'll get someone else to do the job, as I have no wish to turn a partly working SG into an expensive door stop.

One of the disadvantages of having the voltage reference for the DACs, on the motherboard, fed from another PCB is that it makes it impossible to test the mother board in isolation. That's what I had been trying to do, blissfully unaware that the circuit had been modified, which accounts for the strange voltage readings around the Op Amp.

I have now reassembled the SG and brought TP18 out on a flying lead. The reference voltage is rock steady at 2.448 Volts (spec 2.45 Volts), so no problem there.

However the SG still has the same intermittent fault, sometimes outputting the wrong frequency. V ref does not change when the fault is present. By following the fault finding flow charts and using the built in diagnostics I have proved to my satisfaction that the fault is either with the DACs or the synthesiser board. I do hope it's the former.

I am unable to progress this any further until I obtain some break out ribbon cables. I will then be able to operate the tester whilst having access to the test points on the mother board. Diagnostics are available which cause the DACS to generate a staircase waveform which can be monitored with a 'scope. I'll also be able to compare bus voltage under fault and no fault conditions.

So thanks for all your help everyone. I now know a lot more about Op Amps.

I guess the moral of this story is to assume nothing and make sure you have the correct documentation.

[Mooly]

Well, good luck with it Graham. Its certainly a nice piece of kit and worth saving if possible. The DAC's are Ferranti ZN429's I think (haven't got the manual on this PC). An oldie, I remember making an audio reverb with those back in the early 80's.