Quad 405 modules - output voltage varies!

[AKlein]

Hi All,

Dont know if anyone can help.

I have made a clone 405m2, which consists of a pair of modules upgraded as per Net-Audio mods, Dual mono power supply, with a new tranformer & case.

After switching power on, I measured the output voltage from the speaker terminals which reads as follows:-

1) After switch on, the voltage immediatley reads 0v.
2) After a moment, each chanel starts to increase in voltage, by 0.01v which is also a negative voltage.
3) This voltage increases steadily, untill it reaches about -2.5v, after which it starts to vary wildly.
4) After switch off, the voltage is now going all over the place, positve, negative and at around 12v to 25v!

Anyone come across this problem before?, this is my only transistor amp and would like to get it up and running.

Regards

Adam

[Ouroboros]

This is the dc output offset I guess?

The input op-amp on the 405 also functions as the dc-servo stage. It would seem that there's something wrong with this part of the circuit. Are you using original 405 boards that have been modified, and if so, was the dc output ok before you did the mods?

[AKlein]

I bought the boards off ebay and although the seller stated they worked, I can't be for sure. They are the mark 2 boards.I have replaced the op-amps to OPA627, replaced the zener's etc, capacitors plus to output transistors & drivers.

[orbanp]

Hello Everyone,

Interesting Adam mentions this!

I was just playing with the simulation of the Quad amplifier (I might want to build a clone of it sometime). One interesting phenomena I noticed was that the output of the amp would sometimes latch up, it would go out to one of the power supply lines.
First I thought it was simulation error, resulting from initial conditions (the analysis was by using the .op DC analysis). When doing transient analysis (.tran ... startup) then the output would settle properly to 0V.
One explanation is that the DC servo feedback provides the whole output voltage to the input of the opamp. The opamp runs from +/-15V, the rest of the amp runs from a lot higher voltage. Potentially the inverting input of the opamp could go onto higher potential than the +/-15V supply voltage. Different opamps behave differently when the input is outside the operating voltage. Jfet input opamps (like the TL072) would phase reverse, and hence the lockup.
In short, the latch-up is a valid solution to the DC operating state of the circuit! (BTW, the 5532 would also phase reverse in simulation if the input is driven outside the supply voltage range.)

The solution in the simulation was that I reduced the DC feedback with a voltage divider that even if the output is at full rail voltage, the DC servo voltage at the opamp input would remain between +/-15V. This solved the startup problem, and it hardly effected the zero bias of the amp. I will certainly implement that divider into my circuit if I build the amp!

Please note that this is all from simulation and not from actual circuit measurements! YMMV!

Regards, Peter

[Ouroboros]

I know that if you change the input op-amp on the 405, you sometimes have to allow for the fact that many op-amps draw more current than the old LM301 that was originally used. The op-amp supplies are simply fed through series resistors and clamped with Zener diodes. A reduction in the value of the dc feed resistors to provide a higher operating current may be required. (Make sure the Zeners and the resistors wont be over-dissipating though!)

[AKlein]

Thanks for your reply,

But it's a little beyond me to be honest.

Last night I noticed a very silly mistake, I hadn't earthed the inputs!

So, anyway once connected, one channel stayed at -0.02v continuously.

The other channel however was the same as before. Switched on and the output dc measured -0.02v for around a couple of minutes. After which it started to vary widely, measured between 0.00v then -15v, then -12.36v (varies continuously)

Maybe the op amp has gone. I've read elsewhere that the OPA627's may be too 'fast' for a 405?

[orbanp]

Hi Adam,

Check your solderings, you still might have a cold solder joint, that could be a reason for the output floating.
I do not think that an the OPA627 is "too fast", it is internally compensated (stable) for unity gain. There is also a low-pass filter after the opamp stage in the circuit. If the circuit were unstable, it would oscillate, you would be able to see it on a scope, and the final transistors would be quite warm.

Good luck, Peter

[AKlein]

Thanks Peter,

I'll check all joints this evening and let you know tomorrow.

There is definately an odd smell from the board, not quite a burning smell - and it seems to come in wafts.

Sorry, no scope i'm afraid. Only just armed myself with a DMM!

[bosedtobosed]

sounds like its oscillating have you shorted the i\p?

[Lumba Ogir]

bosedtobosed,

sounds like its oscillating

I agree.