NAD 3130 Amplifier-same fault 15 years apart!!

[Nick]

The thing that really kills speakers is DC offsets - decent directly-coupled amplifiers use a DC-servo to correct for any drift, and hopefully, a cutout (typically a relay) if the offset gets too large - a decent servo should keep the DC in the order of low 10s of mV... my Quad 405-2s have about 10mV DC at worst...

[Old Sparky]

Just to avoid any confusion, the model no. of the amplifier in question is 3130, not 313 as I erroneously put in the thread title!

[Mark Hennessy]

I've just amended the thread title for you

[Old Sparky]

Having fitted the new o/p transistors, I cautiously powered up the amplifier, and attempted to set up the DC offset for < +/- 30mV as per the service instructions, only to find that on both channels the pre-set pors (RV401 & 4020 needed to be set to max resistance. In addition, although I could get the offset down to zero as measured with a DMM across the speaker connections there was a problem in the LH channel in that the offset would drift to over 400mV. The cause turned out to be that both presets had gone high reasistance, in one case to about 13000 ohms!. I know this amp is maybe 30 years old, but didn't expect that! The original presets were open 'skeeton' types, but I'll be replacing them with enclosed Piher or similar components. The quiescent o/p collector current could, in both channels, be set to <30mA (<30mV across the 1 ohm resistor in each 2N3055's collector supply) so, with new, reliable pre-sets all should be OK.

[Old Sparky]

I fitted the new 1k presets today, and, having followed the setup procedure in the service manual, the DC offset iremains constant at ca. 1 - 2mV on both channels, whilst the idle current on the channel with the new o/p transistors is ca 27mA after about 5 minutes running time. That in the channel with the original o/p transistors is around 18mA after the same running time, all these measurements taken with no load as per the instructions, which also state that the idle current should be 30mA, though I can't see why a lower idle current should be a problem. I have since run the amplifier continuously, albeit at modest volume levels, for about 1.5 hours with no sign of anything untoward happening, and will continue to test run it over the next couple of days, but at the moment it appears to be a case of 'job done'
(I haven't yet fitted the temperature sensor to the heatsink, but may well do so)

[Mark Hennessy]

That sounds like progress

Providing there's no audible crossover distortion, I wouldn't worry about the quiescent current being a little bit low. Sure, if it was here, I'd measure the distortion, just to be sure, but as we are very sensitive to crossover distortion, a subjective test is probably good enough.

Very wise to let it soak, just in case. But I suspect that you're right about it being "job done"...

Out of interest, what transistors did you decide to use?

All the best,

Mark

[Refugee]

I would guess that it was those pots that let the bias go high enough to cause the transistors to fail.
Well done.

[Mark Hennessy]

The pre-sets are for DC offset, not quiescent current.

Looking at the schematic, the effect of that pre-set failing high is much-reduced by the surrounding circuitry - the pre-sets are in series with a 470 ohm resistor, and that combination is in parallel with 680 ohms. That feeds a transistor that is a current source which loads the collector of the input transistor (not even a long-tail pair!) and also is cross-coupled into the base of said input transistor. It's actually quite hard to be completely sure how the input stage is working because of the obscure arrangements they've made between these two devices. There's even bootstrapping (via R405) to raise the input impedance...

The quiescent current is set by fixed resistors, which is A Good Thing. Unfortunately, the omission of low value emitter resistors on the output pair is Not A Good Thing, and certainly won't help reliability.

Still, at least this stuff is fixable. Not so sure about modern amplifiers with SMT devices and ICs that go obsolete a decade after entering production

[Old Sparky]

Mark, I used the MJ15003 & 15004 you recommended. As to emitter resistors, I presume these could be added without a major circuit redesign. Something like 1 or 2 ohms 5 watt in series with each emitter at an (educated!)guess, though, as this amp. has been working for almost 30 years with only two failures in all that time, it seems to me that it's quite good 'as is'. I shall still soak test it for a while, and then monitor the DC offset again, just to be sure.

[Mark Hennessy]

Thanks for letting me know - I'm sure they'll stand the test of time...

For info, typical emitter resistors would be 0.1 to 0.47 ohms. In practice, 0.22 is seen most often. Even though these are ridiculously low values, they make a surprising difference to the quiescent current stability, which in turn affects crossover distortion (Douglas Self has written a lot about this - interested folk can find old editions of his book available as PDF downloads - highly recommended!). In addition, such resistors provide a point to monitor the output current for current-limiting purposes. And a point to measure voltage for setting the bias.

Anyway, as you say, it's done pretty well. Mentions of the emitter resistors (or lack thereof) is really a "point of interest" rather than anything more serious

Thanks again,

Mark