DOA repair

[saint gillis]

Hello I'm repairing a preamp from a brand that scratches the numbers on the components...
There are some discrete op amps inside, that I've reverse engineered, and fortunately they didn't scratch everything conscientiously and it seems they use 2N3904 and 2N3906 transistors (I'm not sure for every transistor),
but if my schematic is correct I've considered they were all EBC,
Of course there might be mistakes, it was a tough work, feel free to comment

 

[JohnRoberts]

I still have 2n3904 and 2n3906 in my back lab... they were generic low power NPN/PNP devices back in the 70s.

JR

 

[moamps]

The base of Q12 should be DC biased. It may need to be connected to the LED's cathode.

 

[saint gillis]

Well done, I've corrected it

 

[saint gillis]

also I've inverted + & - inputs, corrected too

 

[saint gillis]

The preamp input stage (there might be mistakes)

 

[saint gillis]

There's a problem with U5. I've repaired it, and after a while it failed again.
I measure a huge offset at it's output : -9V (but sometimes -6V, it changes)

U3b seems to do it's servo job (it outputs a V+ voltage), but not enough to make the correction on U5.

U5 has -0,8V at its IN- and 0,14V at its IN+

 

[saint gillis]

Maybe I should match Q1 and Q2 ?

 

[saint gillis]

Any DOA specialist?

 

[saint gillis]

I ve tried to replace U5 with a NE5534 powered in +/-15V
The output offset is 83mV when I remove the servo circuit, and -5mV withe the servo DC circuit.

So there is something wrong with the DOA U5 ...

 

[saint gillis]

The yellow led should be biased around 0V and is biased around -8V , and I don't manage to find what is wrong..

 

[saint gillis]

Well it's not easy to repair a DOA, all the transistors and diodes are new, I've matched Q1-Q2 and Q11-Q12, checked for continuity and shorts all over the board and there's still something wrong.

I've put the DOA in LTSPICE, opened the neg feedback loop (removed 5K1 feedback resistor on the input stage schematic on U5) , and it seems there is something wrong around the led in the DOA.. but I can't find what..

 

[saint gillis]

Daaamn, it was the 100p silver mica capacitor that was broken, one day to find out !
I'm talking to myself but I don't care, it might help people in the future. And it was the opportunity to correct few mistakes in the DOA schematic.
It works !

 

[ReDolby]

Thank you for your work and for sharing!

I was working on a pair of these for weeks last year; a task not for the faint of heart. I did not find this thread in time and also reverse engineered the DOA, your version revealed a couple of mistakes in mine. Some of the DOAs had none of the transistor values scratched off and you are right- they are all 2N2904/2N2906. The problem with one unit was the 2n2 C1 near U1 - servo would sometimes take a very long time to null DC, sometimes not able to fully null. This took 99% of the time spent on both units to find.

The other unit had a damaged meter movement (it was dropped). Also found cold and missing (!) solder joints, specifically the the interconnecting wires (solder then trim sometimes does not leave a solid joint, thermal cycling probably does not help).

Both units back in service for close to a year, no problems. Note that Hoyt meters are still in business and were able to supply an exact replacement. They are made to order and are not inexpensive, but good quality and Hoyt were a pleasure to work with.

I have been a keen observer of this forum for years, learning tons, many thanks to all. This was such a tough one, it was time to share my experience in hope that it helps someone else.

 

[JohnRoberts]

That input long tail pair does not look like it is getting much quiescent bias current.

[edit- I'm just speculating so bread board it up and see what smokes.../edit]

JR

 

[Dylan W]

The preamp input stage (there might be mistakes)

Trying to understand what U1 is doing... it's servoing U2's input back to the same place?

 

[saint gillis]

Trying to understand what U1 is doing... it's servoing U2's input back to the same place?

I guess so, to keep it at clean 0VDC and avoid scratches at the first half of the gain switch

 

[abbey road d enfer]

That input long tail pair does not look like it is getting much quiescent bias current.

[edit- I'm just speculating so bread board it up and see what smokes.../edit]

JR

Well, I haven't breadboarded it, but simulation hints at about 500uA each, which is on the high side compared to most DOA's that run them at 50-100uA, which suggests it's optimized for low source impedance.
OTOH, the full schemo seems to indicate an OSI around 10kohms, which would put it in the same ballpark as many other DOA's.
Either there is something wrong in the schemo, or I've lost my marbles.
Now, if R6 was 560k instead of 56, it would put everything in order, with about 50-60uA for each input device.

BTW, the power arrangement is quite quirky. Q13 & 14 constitute a bipolar almost constant current source, so under strain or even failure, current would be limited to a safe value (about 150mA).
I wonder what are the effects of this arrangement on performance...

 

[Philbo King]

Hello I'm repairing a preamp from a brand that scratches the numbers on the components...
There are some discrete op amps inside, that I've reverse engineered, and fortunately they didn't scratch everything conscientiously and it seems they use 2N3904 and 2N3906 transistors (I'm not sure for every transistor),
but if my schematic is correct I've considered they were all EBC,
Of course there might be mistakes, it was a tough work, feel free to comment

It looks like the voltage regulator transistors are drawn weird.
The arrow on the emitters face the wrong way (the arrow should always point toward negative). Also I would normally expect to see the emitters facing the voltage source for series-pass regulator transistors.

 

[abbey road d enfer]

It looks like the voltage regulator transistors are drawn weird.
The arrow on the emitters face the wrong way (the arrow should always point toward negative).

The arrow indicates the current direction from positive to negative. Seems correct to me.

Also I would normally expect to see the emitters facing the voltage source

That's exactly what they do, but...

for series-pass regulator transistors.

It's not a voltage regulator arrangement, its' a CCS (constant current source), for which the collectors are the outputs.
Actually, the arrangement is not used as a current regulator, just as an overcurrent limiter.