Rotel RB-990BX fluctuating bias

[Fleime]

Hello,

First off, I am a electrical novice. I guess I know the basic stuff an average person does. I have actually learned most here on AK, but forgive me if I'm not using the correct terminology as of yet. I have a multi-meter.

After acquiring some Infinity Kappa 8.1, I was advised on the Infinity forum to get a decent amplifier. So I bought a second hand Rotel RB-990BX. The amplifier had been (somewhat) serviced by the seller (private audio repair man), after which he stated the amplifier was in top condition.

The amplifier has not been auditioned due to distance, and has been sent to me. After arrival I shortly hooked it up to check if it was working, which it did, however I did not test or audition them extensively (yet).

I let it run idle, no loads, to check the bias. I have adjusted the both channels to 7 mV (according to manual) using the bias trim pots (I think the're called). The pot on the left channel is much more sensitive than the right channel. However, the bias on both channels fluctuates up and down, with a total range of around 0.4 mV difference between min and max readings.

So, is the bias supposed to fluctuate? And if so, is the range acceptable?
If not, what could be the cause of the fluctuation? The trim pots?
Is there something else I can do to determine the cause?

Additional information, if relevant:

- the DC offset at the speaker terminals is 25 mV for the left channel, and -25 mV for the right channel. I have come to understand the polarity is not an issue;
- temperature at idle is 95 ℉ for the left channel, and 100 ℉ for the right channel. When in use the temperature gets to about 105 ℉ for the right channel.

Thanks in advance.

 

[Powertech]

Hello Fleime, The Rotel 990BX is a good solid amp in my opinion - I like the '90' series Rotels and use them myself. The bias may well fluctuate slightly with changes in temperature. 0.4mV drift measured across an emitter resistor is not really a great deal. In any case, the bias setting given for most AB class amplifiers is usually just above the output transistor switching threshold value and is not actually that critical. It is set to ensure avoidance of pure class B operation and to keep the output transistors within their linear operating range. I let my amps warm up for at least 30 minutes before measuring the bias.
A major increase above the stipulated value can have sonic benefits (I run my Rotel power amps at about 3 times the manufacturer's specified value to maintain class A operation at low signal levels). You do have to be very careful with amplifier running temperature if you raise these values significantly.
Your reported running temperatures are probably about right. Mine (even with increased quiescent running current) are running at about 35degC (95F), slightly higher if driven hard.
The DC offset is a little high, but not ridiculously so. - Not enough to damage loudspeakers. The negative value is of slight concern. The imbalances are probably due to gain drift in the power amplifier input stages (I haven't looked at the circuit diagram). If you can replace the transistors with matched pairs, that should reduce this. Anything between zero and +20mV is generally fine.
Sometimes, even pushing the input pair transistors together so the cases are touching and then gluing them to each other can help to match their temperature drift, and the DC offset (if the circuit layout allows this). I can't remember if these amps had single or double longtailed input pairs - I know the 970 did.

 

[Fleime]

Hello Powertech, thank you for your extensive reply, I appreciate it. Your answers do raise new questions, perhaps you, or someone else, can address these as well. I do however try to research most topics on my own first. I appreciate references as well.

I provided a circuit diagram I took from the internet. The coloring has not been done by me. For a official PDF I refer to e.g. hifiengine.

It is good to know the bias drift is not problematic, as well as the temperature. I also noticed that the amp took about half an hour to warm up to get towards a stable temperature. I have now also learned that there is a relationship between the bias and operation class.

Concerning the DC-offset, I do not exactly understand 'gain drift in the power amplifier input stages'. Can I determine this somehow? I also am unsure what you mean with matched pairs of transistors. Do you mean the same type, or same size, or both, or something else altogether?

I have measured/checked the transistors, diodes and capacitors on circuit board X-1176-01 (2/2), where power enters the system. Most measure OK, one transistor seems faulty (Q803). Current in this transistor is able to flow in both ways between the base and the emitter, and there is a voltage drop between the collector and the emitter.

I am unsure if any if these transistors on the above mentioned circuit board are 'input pair transistors'. I have not yet had time to research what 'single or double longtailed input pairs' are.

Finally, can you tell me why a negative DC offset value is off slight concern?

 

[Fleime]

Here are pictures of the left channel circuit board. I am no experts on circuit boards or soldering, but it looks like the board has endured some intens heat at one place. On the other side of this spot are weird transistors (white ones)? Some of the traces have come loose from the board,is this a problem, perhaps also in combination with the heat? Should something have been done about it? They do seem to conduct electricity still. Opinions are appreciated.