"Arfa Daily" wrote in message
...

"Eeyore" wrote in message
...


Arfa Daily wrote:

If all else fails, the temperature that the output
transistors run at on the rebuilt channel, compared to those on the
other
channel, is usually a pretty good guide as to how evenly the two
channels
are running with respect to one another. This should be checked with the
amp
just idling.

And the load not connected.

Graham


I've never really thought about whether or not a load should be connected
when making that assessment. I suppose it might depend on how the bias
arrangement works, or if there are any other adjustments such as offset,
but in theory, surely if it's a standard complimentary output stage, with
the output DC coupled to the midpoint, and with the amp just idling with
no input, there should be no current in the load ? If there's any kind of
asymmetry, leading to a DC offset at the midpoint, then yes, I can see
some value in not having a load connected, but also, I would suggest that
it would indicate a further problem that needed resolving ?

Interested in your thinking here, Graham.

Arfa


Even a slight DC offset would affect the voltage measured across an emitter
resistor, if a load is connected. I put a pencil to it and found that with a
50 millivolt offset (a bit high but certainly not uncommon) and a 6 ohm DC
resistance representing the speaker, this would mean about 18 millivolts
additional across a .22 ohm emitter resistor. Reason enough to adjust the
offset first, I would say. Unfortunately, many amps have no offset
adjustment as such. I will try to look up the 2240PE bias adjustment at work
and e-mail Dave directly. I've seen some NAD's which require measurement
across a 1 ohm resistor which is made available by desoldering two adjacent
foil pads. Once the adjustment is done, the pads are resoldered.

Mark Z.