Schematic for Servo opamp.

[sonnya]

Very simpel but nice performance.

[capslock]

Hi Sonny,

this is going to have a huge offset voltage compared to an OP177. What is the point of building a discrete op amp for servo?

Regards,

Eric

[sonnya]

Yes you are right.....

only i am using 100k resistors and at the same time balancing the resistance seen by the inputs.
This will on other hand require film caps in the size of 2.2uF <-> 10uF like MCAP's.

The reason to make one was to make sure it was running in class A and not have cross over artifacts.. When it is active. Wich you most likely to have in any opamp because of the class AB .. mostly B the opamps are running in.

I do not want it ruin my amp by using a opamp with this problem.

I have made some trimming through a 1meg resistor and the trimmer is centeret. And i have no problem with drift because of the very little temp rise in the transistors.

Offset +/- 0 .. 1mV... I think this is good enough....

With 2.2uF and 100k my amp has a cutoff in the bottom at .5Hz@-3dB.

You can also change the values on some of resistors and have a small fast discrete opamp.

It is also very cheap to make... like 1 - 5 euro...

Sonny

[sajti]

" The reason to make one was to make sure it was running in class A and not have cross over artifacts.. When it is active. Wich you most likely to have in any opamp because of the class AB .. mostly B the opamps are running in. "

Hey man,

it's easy to avoid. Just put 3kohms resistor between the opamp output, and the negative rail. This generates 5mA (with 15V supply) bias for the opamp's output stage, and set it into class A. I use it in all of my poweramp.

Sajti

[sonnya]

Originally posted by sajti

it's easy to avoid. Just put 3kohms resistor between the opamp output, and the negative rail. This generates 5mA (with 15V supply) bias for the opamp's output stage, and set it into class A. I use it in all of my poweramp.

Sajti

And have a heavily changing bias level (with the changing output level) in the one BJT in the output turned on!

[capslock]

Is this really relevant? After all, the input voltage of a servo op amp will only change very, very slowly, and hence the currents in the output stage will change so slowly that it can hardly be called switching + the GBW of the op amp should be sufficient to deal with the output nonlinearity.


By the way, I was not questioning offset due to input bias currents. In my experience, it is difficult to select two discrete transistors for low offset and if you find them, their drift will be terrible.

Regards,

Eric

[sajti]

Originally posted by sonnya


And have a heavily changing bias level (with the changing output level) in the one BJT in the output turned on!

Only the NPN output transistor works. First time I try this solution with an old RIAA circuit, and it increase the sound!
Of course, heavy loading can set the output into class AB, and current regulator works better...

Sajti

[sonnya]

Originally posted by capslock
Is this really relevant? After all, the input voltage of a servo op amp will only change very, very slowly, and hence the currents in the output stage will change so slowly that it can hardly be called switching + the GBW of the op amp should be sufficient to deal with the output nonlinearity.


By the way, I was not questioning offset due to input bias currents. In my experience, it is difficult to select two discrete transistors for low offset and if you find them, their drift will be terrible.

Regards,

Eric

They will only drift terrible if the power dissipation is high, but in this case the thermal drifting is only due to the ambient temperature. I dont think you have calculated the dissipation in this opamp's diff. input pair! 216uA*~15V => 3.24mW in each BJT that is a temperature rise (with an thermal resistance 200degress/watt) in the silicon of .648degress above the ambient temperature....

Try recalculate the circuit .. i have not just thrown the circuit together blinded!

Originally posted by sajti


Only the NPN output transistor works. First time I try this solution with an old RIAA circuit, and it increase the sound!
Of course, heavy loading can set the output into class AB, and current regulator works better...

Sajti

Yes you are right Sajti... try this on a AD845, AD8610AR the do sound really sweet and detailed in htis configuration....

By the way ... this kind of way replying you two have done in this thread really spoils the fun of making circuit and posting them on this board so every one can make for them self with local easy to get parts! is this not the fun of DIY....

This is how it is intended to be used...

[sajti]

Nice circuit,
but I prefer to use inverting DC servo. I don't use any adjustment, because 1-2mV is no problem for power amplifier.
So, I use inverting, connect them to the input of the amplifier, and use RC filter at the output of the opamp to reduce the noise. Double filtering can result some ringing (low speed ringing - of course), and to avoid this I set the filter's frequency to 1:10, say 0.2Hz for the inverting opamp, and 2Hz for the following RC filter.

Sajti

[sonnya]

Originally posted by sajti
Nice circuit,
but I prefer to use inverting DC servo. I don't use any adjustment, because 1-2mV is no problem for power amplifier.
So, I use inverting, connect them to the input of the amplifier, and use RC filter at the output of the opamp to reduce the noise. Double filtering can result some ringing (low speed ringing - of course), and to avoid this I set the filter's frequency to 1:10, say 0.2Hz for the inverting opamp, and 2Hz for the following RC filter.

Sajti

The adjustment was added to lazy people (me) who haven't matched the BJT's.
I agree with you that 1 - 2 mV is sufficient enough.

I also agree on the inverting thing. but then you need to add another inverting stage.... You see i like my amps totally DC coupled. They are also noninverting. So i have to do the adjustment on the feedback point (in-).. But still after another thought, i can move the input filter to the output.... thanks for the comment.

The ringing... Don't you just move the ringing point? ... up to the area 20Hz - 200Hz?