Bias adjustment

[Arvid Puschnig]

Hi,

some time ago I started repairing my old SONY amplifier TA-AV501R. It
took some time getting the spare parts, but now all DC voltages look
good. Since I had to change the power transistors, I now have to adjust
the bias voltage.

Procedure from service manual:
....
2. After the power is turned on, complete within 30 seconds to adjust
RV500 and RV550 so that the bias voltage becomes 1 mV.

I tried this, but the bias voltage increases more and more even after
the 30 seconds.
Would it be right to adjust the bias voltage so that it becomes 1mV 30
seconds after turning the power on?

Is it right, that the bias voltage increases more and more even after
this time?

Schematics:
http://www.sbox.tugraz.at/home/a/arvid/TA-AV501R/Power.jpg

@Tweetldee:
Thanks for your support so far!

arvid

 

[Asimov]

"Arvid Puschnig" bravely wrote to "All" (06 Dec 03 03:11:06)
--- on the heady topic of "Bias adjustment"

AP> From: "Arvid Puschnig" <arvid at sbox dot TUGraz dot at>

AP> Hi,

AP> some time ago I started repairing my old SONY amplifier TA-AV501R. It
AP> took some time getting the spare parts, but now all DC voltages look
AP> good. Since I had to change the power transistors, I now have to
AP> adjust the bias voltage.

AP> Procedure from service manual:
AP> ...
AP> 2. After the power is turned on, complete within 30 seconds to adjust
AP> RV500 and RV550 so that the bias voltage becomes 1 mV.

AP> I tried this, but the bias voltage increases more and more even after
AP> the 30 seconds.
AP> Would it be right to adjust the bias voltage so that it becomes 1mV 30
AP> seconds after turning the power on?

AP> Is it right, that the bias voltage increases more and more even after
AP> this time?

AP> Schematics:
AP> http://www.sbox.tugraz.at/home/a/arvid/TA-AV501R/Power.jpg

AP> @Tweetldee:
AP> Thanks for your support so far!

AP> arvid

I hope by bias voltage you mean the voltage across an emitter
resistor of an output power transistor and not the DC balance...?

1mv? That's an awfully small voltage, I usually see a higher mv...
Perhaps the emitter resistors are a really small value?

Start from zero and turn it up slowly until it becomes stable. It is
normal for the idle current to increase as the output transistors warm
up. However at some point it should stabilize and if it doesn't there's
some problem with the temperature feedback strategy.

Some amps stability factor design depends on the emitter resistors to
self limit the idle current and some use a diode, thermistor, or
transistor in direct contact with the heat sink for feedback. Maybe you
forgot to reposition some thermal feedback device when changing the
output power transistors?

.... Resistance Is Futile! (If < 1 ohm)

 

[Arvid Puschnig]

"Asimov" wrote:

AP> Schematics:
AP> http://www.sbox.tugraz.at/home/a/arvid/TA-AV501R/Power.jpg

I hope by bias voltage you mean the voltage across an emitter
resistor of an output power transistor and not the DC balance...?

Yes, as you can see on the schematics bias voltages are taken from R527
and R578.

1mv? That's an awfully small voltage, I usually see a higher mv...

That's why I had to borrow an 4.5 digit voltmeter...

Perhaps the emitter resistors are a really small value?

That's right, 0.22Ohm!

Start from zero and turn it up slowly until it becomes stable. It is
normal for the idle current to increase as the output transistors
warm
up. However at some point it should stabilize and if it doesn't
there's
some problem with the temperature feedback strategy.

I thought so too, but the service manual notes that this adjustment
must be completed within 30 seconds. Even after 60 seconds voltage
increases. Perhaps there's no temperature feedback?
It's very strange that this behaviour occurs on both channels, so this
would mean that there's a problem within the temperature feedback in
the left channel as well as in the right channel.

Some amps stability factor design depends on the emitter resistors to
self limit the idle current and some use a diode, thermistor, or
transistor in direct contact with the heat sink for feedback. Maybe
you
forgot to reposition some thermal feedback device when changing the
output power transistors?

I am sure there's no diode, thermistor or transistor in direct contact
with the heatsink or the power transistors.

... Resistance Is Futile! (If < 1 ohm)

Do you know, why this adjustment has to be finished within 30 seconds?
The only thing I can imagine, would be that the resistors get too hot,
if bias voltage would be too high, but I don't think that it will harm
them if I start with the lowest value and turn it up slowly until the
1mV get stable, even if it takes some minutes?

Thanks so far,
arvid

 

[Asimov]

"Arvid Puschnig" bravely wrote to "All" (07 Dec 03 01:39:12)
--- on the heady topic of "Re: Bias adjustment"

AP> From: "Arvid Puschnig" <arvid at sbox dot TUGraz dot at>

1mv? That's an awfully small voltage, I usually see a higher mv...

AP> That's why I had to borrow an 4.5 digit voltmeter...

Perhaps the emitter resistors are a really small value?

AP> That's right, 0.22Ohm!

0.22 ohms is actually a relatively large value, I've seen 0.1 ohms and
smaller used. 1mv across 0.22 ohms means an idle current of 4.55mA which
is actually a rather small idle current as many amps use 20 to 60mA.

Start from zero and turn it up slowly until it becomes stable. It is
normal for the idle current to increase as the output transistors
warm up.

AP> I thought so too, but the service manual notes that this adjustment
AP> must be completed within 30 seconds. Even after 60 seconds voltage
AP> increases. Perhaps there's no temperature feedback?

With 0.22 ohms the idle current will probably not runaway. I think the
30 second constraint is to take advantage of the transistors being at
room temperature. If the adjustment were made when warm then the idle
current wouldn't increase according to the expected temperature graph.
This could result in too little idle current when hot and possibly
exhibiting more distortion than the specifications. That's the only
reason I can think of.

Personally I would adjust idle current according to distortion at the
33% power level. The 1/3 power level results in the most heating of the
output transistors and at that point dissipation is 25% of full output
power for each device. This is the greatest thermal stress point. i.e. a
100 watts rms amp run at 1/3 output causes 25 watts dissipation in each
transistor. However at 100 watts output the dissipation reduces to about
15 watts per device.


AP> Do you know, why this adjustment has to be finished within 30 seconds?
AP> The only thing I can imagine, would be that the resistors get too hot,

Indeed the resistors were selected to dissipate much more than 4.5
"micro" watts!!! I'll bet they are rated to dissipate 3 to 5 watts maybe
even 7 watts. I don't think the resistors would get hot enough from the
idle current to make a difference.


AP> if bias voltage would be too high, but I don't think that it will harm
AP> them if I start with the lowest value and turn it up slowly until the
AP> 1mV get stable, even if it takes some minutes?

Well, the technique of adjusting it to 1mv within 30 seconds seems okay
if the design calls for it. Like I said I'd get the amp preconditioned
at 33% power and then adjust the current for least distortion but it's
probably more complicated because of the extra analyzing equipment
required. Seems to me the designer did the leg work to make it simpler
for the technician. That's a rare thing these days and was more common
in the heyday of discrete component stereo.

.... The current limits placed are based on resistance

 

[Arvid Puschnig]

"Asimov" wrote:

Well, the technique of adjusting it to 1mv within 30 seconds seems
okay
if the design calls for it. Like I said I'd get the amp
preconditioned
at 33% power and then adjust the current for least distortion but
it's
probably more complicated because of the extra analyzing equipment
required. Seems to me the designer did the leg work to make it
simpler
for the technician. That's a rare thing these days and was more
common
in the heyday of discrete component stereo.

Thank you very much.
I'll try the simple way to adjust bias voltage to 1mV within 30 seconds
so that voltages for both channels increase in the same way. There's
just one question left: Since voltages increase very fast within those
30 seconds, should I try to "reach" the 1mV exactly after those 30
seconds or would it be better to "reach" them earlier? (10/15/20
seconds?)
I will use 2 voltmeters to perform this (one per channel).

arvid

 

[Jerry Greenberg]

The volume control must be set to zero. The bias current in most of
these types of amps is about 60 to 80 ma or so. Set the voltage
accross the 0.22 ohm emitter resistors to about 15 to 20 mv. After
setting this, check the DC offset is as close to 0 Volts as possible
at the speakers, with no speakers connected.

I normaly check the power output in to a dummy load, and use a
distortion analyser accross the dummy load to look at the distortion
through the bandwidth of the output stage at 1 Watt, 10 Watts, and
about 20 Watts. The amp is fed with a low distortion audio sweep
generator to test. This way, I can also set the bias for optimum,
using as low a current I can get away with, and still have as low
distortion as possible, and proper output power.

It would be best for you to obtain the information from Sony. If you
were to purchase the origional service manual, it should normaly be in
there.

Jerry Greenberg
http://www.zoom-one.com

--





"Arvid Puschnig" <arvid at sbox dot TUGraz dot at> wrote in message news:<3fd13a5a$0$18044$3b214f66@aconews.univie.ac.at>...

Hi,

some time ago I started repairing my old SONY amplifier TA-AV501R. It
took some time getting the spare parts, but now all DC voltages look
good. Since I had to change the power transistors, I now have to adjust
the bias voltage.

Procedure from service manual:
...
2. After the power is turned on, complete within 30 seconds to adjust
RV500 and RV550 so that the bias voltage becomes 1 mV.

I tried this, but the bias voltage increases more and more even after
the 30 seconds.
Would it be right to adjust the bias voltage so that it becomes 1mV 30
seconds after turning the power on?

Is it right, that the bias voltage increases more and more even after
this time?

Schematics:
http://www.sbox.tugraz.at/home/a/arvid/TA-AV501R/Power.jpg

@Tweetldee:
Thanks for your support so far!

arvid

 

[Arvid Puschnig]

"Asimov" wrote:

Well, the technique of adjusting it to 1mv within 30 seconds seems
okay
if the design calls for it. Like I said I'd get the amp
preconditioned
at 33% power and then adjust the current for least distortion but
it's
probably more complicated because of the extra analyzing equipment
required. Seems to me the designer did the leg work to make it
simpler
for the technician. That's a rare thing these days and was more
common
in the heyday of discrete component stereo.

Now I set the bias so that voltages get 1mV within 30 seconds for each
channel.
Voltage for right channel increases a bit faster so 1mV is reached
after 10 seconds, for left channel it takes about 30 seconds. When
transistors are warm (after a few minutes), bias gets stable at about
6mV and I set voltages to the same value, because of the different
speed of increment. So voltages become 1mV within 30 seconds (10/30
seconds) when transistors are "cold" and voltages are the same when
transistors are warm.
Do you think this would be right or doesn't it matter if voltages have
different values for each channel when transistors are warm?

arvid

 

[Asimov]

"Arvid Puschnig" bravely wrote to "All" (10 Dec 03 14:40:02)
--- on the heady topic of "Re: Bias adjustment"

AP> From: "Arvid Puschnig" <arvid at sbox dot TUGraz dot at>

AP> "Asimov" wrote:

Well, the technique of adjusting it to 1mv within 30 seconds seems
okay
if the design calls for it. Like I said I'd get the amp
AP> preconditioned
at 33% power and then adjust the current for least distortion but
it's
probably more complicated because of the extra analyzing equipment
required. Seems to me the designer did the leg work to make it
AP> simpler
for the technician. That's a rare thing these days and was more
AP> common
in the heyday of discrete component stereo.

AP> Now I set the bias so that voltages get 1mV within 30 seconds for each
AP> channel.
AP> Voltage for right channel increases a bit faster so 1mV is reached
AP> after 10 seconds, for left channel it takes about 30 seconds. When
AP> transistors are warm (after a few minutes), bias gets stable at about
AP> 6mV and I set voltages to the same value, because of the different
AP> speed of increment. So voltages become 1mV within 30 seconds (10/30
AP> seconds) when transistors are "cold" and voltages are the same when
AP> transistors are warm.
AP> Do you think this would be right or doesn't it matter if voltages have
AP> different values for each channel when transistors are warm?

AP> arvid

Your results seem quite acceptable. 6mV across 0.22 ohms is about 30mA
which is a typical value for that type of amplifier circuit. Be
reassured that a little difference is to be expected and quite normal.
Consider that even no two speakers are completely alike.

Keeping in mind that crossover distortion is most objectionable at very
low listening levels, then it is quite appropriate to equalize the idle
current after the temperature stabilizes.

Some designs use a higher idle current so that the amplifier operates in
Class-A mode all the time at low listening levels and then in Class-B at
concert levels. Hence the design name of Class-AB.

.... I worked hard to attach the electrodes to it.

 

[Arvid Puschnig]

Hi,

I want to thank all of you for your time and your help.
My amplifier is now working again and even without the usage of special
equipment for setting the bias voltages it sounds quite good. Now it's
time to start the 24h test...

thx,
arvid

 

[Arvid Puschnig]

Hi,

I want to thank all of you for your time and your help.
My amplifier is now working again and even without the usage of special
equipment for setting the bias voltages it sounds quite good. Now it's
time to start the 24h test...

thx,
arvid