Creek Evolution 2 bias current,hm

[misfits73]

I'm trying to adjust the bias current and according to the procedure i got from Creek it should be 1.5mV voltage drop across the emitter resistor (same reply another chap got for his Evo 1 in another thread).3mV in total across the output pair.

I decided to go up to 22mV per pair,the sibilance is gone and the sound is fuller (these were my issues).

So basically i don't know what to do,listen to the factory or trust my ears?Can somebody advise?

I also got the schematics if that will help

Cheers

 

[JonSnell Electronic]

1.5mV across the emitter resistor sounds a mistake as the output transistors are still in crossover mode.
Depending on the output transistor type and resistor value, try to get about 15 - 20mA current to flow. That will remove crossover distortion and it will still run cool.

 

[misfits73]

"You can increase the bias current if you want, but it will start to get hotter. Note that there is no over-temperature cut-out. Without a distortion measurement it is hard to quantify or compare."

Idle current adjustment:
Use a multimeter to measure the voltage of the idle current from the R-MID to junction
of R522, R516 and Q510 emitter. Adjust R532 to 1.5mV. +/- 5%

Use a multimeter to measure the voltage of the idle current from the L-MID to junction
of R624, R616 and Q610 emitter. Adjust R632 to measure 1.5mV. +/-5%

Something's not right here

 

[Mooly]

Just over 100 milliamps would be the correct figure for that topology and component values. On 90 volt rails (total) that is at least 9 watts per channel dissipation.

That may be more than the heatsinking capacity can tolerate as a constant dissipation. Manufacturers often under bias designs... I guess because of that constraint.

 

[misfits73]

Just over 100 milliamps would be the correct figure for that topology and component values. On 90 volt rails (total) that is at least 9 watts per channel dissipation.

That may be more than the heatsinking capacity can tolerate as a constant dissipation. Manufacturers often under bias designs... I guess because of that constraint.

100mA between the emitters of Q510 and Q515 you mean?

 

[nigelwright7557]

1.5mV across the emitter resistor sounds a mistake as the output transistors are still in crossover mode.
Depending on the output transistor type and resistor value, try to get about 15 - 20mA current to flow. That will remove crossover distortion and it will still run cool.

I tend to set bias current using sig gen and a scope.
I apply a signal and monitor output.
I turn up bias until crossover distortion just goes.

You can get odd things happening if the dc offset isn't zero and a speaker is connected. Instead of current going through both output resistors it partly goes to the speaker. My method gets around that.

 

[Mark Tillotson]

100mA between the emitters of Q510 and Q515 you mean?

100mA flowing through everything in fact since current is conserved round a loop circuit. V = IR gives you the voltage difference that produces for the emitter resistor(s).

 

[misfits73]

I tend to set bias current using sig gen and a scope.
I apply a signal and monitor output.
I turn up bias until crossover distortion just goes.

You can get odd things happening if the dc offset isn't zero and a speaker is connected. Instead of current going through both output resistors it partly goes to the speaker. My method gets around that.

That's what i've been advised but i don't have a scope.DC offset is around 0mV, there is a trim pot for that.

 

[misfits73]

100mA flowing through everything in fact since current is conserved round a loop circuit. V = IR gives you the voltage difference that produces for the emitter resistor(s).

At the moment is set at 50mA (11mV) per output pair.
The heat dissipation is 1.1W if my calculations are correct:
I=11mV/0.22Ω=50mA

P=22mV^2 / 0.44Ω=1.1W (i believe)

Still the heatsink is not hot enough

 

[nigelwright7557]

At the moment is set at 50mA (11mV) per output pair.
The heat dissipation is 1.1W if my calculations are correct:
I=11mV/0.22Ω=50mA

P=22mV^2 / 0.44Ω=1.1W (i believe)

Still the heatsink is not hot enough

Its the dissipation in the transistors that is important not the resistors.
i.e. 40 volt rails and 50mA gives 2 watts into each output transistor if a single pair.

 

[Mooly]

100mA between the emitters of Q510 and Q515 you mean?

100ma flowing in each emitter resistor. That would mean a volt drop seen across each of 22 millivolts.

At the moment is set at 50mA (11mV) per output pair.
The heat dissipation is 1.1W if my calculations are correct:
I=11mV/0.22Ω=50mA

P=22mV^2 / 0.44Ω=1.1W (i believe)

Still the heatsink is not hot enough

Heat disipation in the reistors is minimal. W = I squared*R which is (0.1*0.1)*0.22 giving 2.2 milliwatts.

or:

W = V squared/R giving (0.022*0.022)/0.22 giving 2.2 milliwatts

or:

W = I*V which is 0.1*0.022 giving 2.2 milliwatts

Each output transistor sees 45 volts between C and E and so with 0.1A flowing we have a dissipation of 45*0.1 giving 4.5 watts per device.

Four output transistors in total means we see at least 18 watts dissipation. That is quite a lot of heat.

 

[Ian Finch]

Almost all bipolar output stage biasing is specified as the voltage across Re or both Re's as the case may be. This is the bias current by proxy and the way commercial amplifiers are set up at manufacture and servicing. Look at any proper service manual, noting this applies to output stages with multiple pairs too. It's quick, simple and reliable for the factory setting but if you want to tinker with it, don't forget that warranties are voided and risks of thermal drift and runaway increase exponentially.

No, our ears are not the best judges when trying to overbias a class AB amplifier. Set it to spec, perhaps a tad more or less or buy a class A amp.

Mooly mentioned 100 mA bias as being appropriate current for a bipolar output stage. I think this would be more appropriate for old style (2N3055 etc.) bipolars and mosfets, myself. Today's audio BJTs are usually run with only about 35mA bias and seen to do quite well with it. That at least makes a good starting point and reduces the standing current and heat markedly. The voltage drop over a single 0.22R resistor would thus be 16mV

I suspect the 1.5 mV spec. is an error - its low enough to be useless as any decent kind of output transistor bias current.

 

[misfits73]

So why the manufacturer suggests 1.5mV instead of 16mV?That is a lot of a difference.Unless the amplifier is intended to work in warm countries so he is worried about damaged components from overheat in the long run but in England the room temperature rarely goes above 20C.I thought he had scoped the amplifier for optimum sound quality before leaving the factory.

 

[Mooly]

You will probably never find out 'why'.

It might be worth you turning the bias down in stages. You have listened to it at 100ma and found it OK. Try 75ma, then 50, then 25. You have only yourself to please so be honest and if you genuinely can't hear a difference between say 50 and 100 then aim for the lower value.

In practice I would expect all audible distortion to vanish at anything over a couple of milliamps tbh.

If the amp has been repaired and transistors substituted then I would want to check it on a scope to make sure it was stable. Maybe something else is going on beside just a low bias setting causing audible problems.

 

[jalejos]

Does anyone know the reference of the optical encoder of the volume control?

 

[nigelwright7557]

In practice I would expect all audible distortion to vanish at anything over a couple of milliamps tbh.

I design my own SS amps so there is no bias set up given for it.
I apply a sine wave and turn bias right down.
I monitor output with a scope.
I turn up bias slowly until cross over distortion just goes and leave it there.

I have some strange results in some amps where there is zero cross over distortion even with no bias current ! I think one was a lateral mosfet amp. So with this I just set it to 100mA which is the recommended bias current.