Please express your opinion about this preamp

[iggz]

Hello people. Please post your thoughts about this input stage and filter section. This will be used in a 2.1 (hopefully high performance) pc speakers.

First thing: the input is differential to minimise pickups from the computer to the preamp. This is the reason the potentiometer/input is wired in such a way. Pot is linear (i prefer to use them). The output is also diferential so other AUX components (if needed) can use the same signal and the full benefits of differential mode signal transmission.

Second: The filtering section will be composed of a LP, BP and HP filters. The order of the filters is very low (2nd order sub, everything else is 1st order), but the units can handle it. The main concern: since the filter section is going to be single-ended, do you think this will be a problem?

Filter section will be attached in the next post.

Regards,

Igor

[iggz]

Filter section, see attachment.

Regards,

Igor

[iggz]

Forgot the file...For some reason I can't edit my previous post.

Regards,

Igor

[myhrrhleine]

The second Op-Amp will unbalance your balanced input.

[iggz]

Which schematic? Input or filter?

Regards,

Igor

[myhrrhleine]

Input

[iggz]

I have to dissagree on that. The second opa in the input section makes the circuit trully differential as it causes the input to see a completely simetrical load. If we take the inverting opa out and ground the resistor that was connected to that output, we get a differential circuit that represents a non symetrical loading to the source. In that case/aspect the input stage is not "trully differential".

I am mainly conserned about the problems that could arise from the single-ended filter section.

[iggz]

I am surprised that so few people are drawn by this thread? I am sure there are people that know the answer to my question. You are welcome to tell/describe your own experiences with such circuits. I am happy of every comment as long as it provides problems one can work with, and before I forget again - thank you for your thoughts, myhrrhleine.

[myhrrhleine]

Originally posted by iggz
I have to dissagree on that. The second opa in the input section makes the circuit trully differential as it causes the input to see a completely simetrical load. If we take the inverting opa out and ground the resistor that was connected to that output, we get a differential circuit that represents a non symetrical loading to the source. In that case/aspect the input stage is not "trully differential".

I am mainly conserned about the problems that could arise from the single-ended filter section.

Perhaps you are disagreeing on the degree of imbalance?
Common Mode Rejection (CMRR) depends on exact balancing.
Even 1% resistor error will seriously degrade balance (CMRR)

[iggz]

Originally posted by myhrrhleine



Perhaps you are disagreeing on the degree of imbalance?
Common Mode Rejection (CMRR) depends on exact balancing.
Even 1% resistor error will seriously degrade balance (CMRR)

If we forget the opa imperfections, the CMRR degrades by:

42dB if the resistors are 1% discreetes and a single one of them is 1% off.

Using a resistor network with absolute tolerance of even 10%, but resistor matching tolerance of 1% will result in 54dB CMRR.

Resistor matching tolerance in resistor networks is quite good and values of 0.05% are not difficult to obtain (altough the absolute tolerance of the network is say 1%), which gives us a CMRR of 80dB.

All this withouth acounting the performances of the opa and at DC, for the circuit in question.

I will try to get a network with 0.01% resistor matching for CMRR in the vicinity of 94dB. How close to 94? Probably not very close, but still enough. I havent choosen an opa yet and trimming the network for AC will be a problem with my equipment.

What about the filter section? I use smd tech., doublesided boards and 0805 resistor/capacitor package. The opa will be SO8, so you can get a rough idea how big the circuit will be. Wil it pickup much interference after the input section?

Another thing: don't take the absolute value of resistors too serious... 1k will not be 1k, i must choose a value for those. 100k is more a way to say "...don't load the outputs/signal too much..", 50ohm is there to say "...this must be low, calculate how low..." etc.