Masterpiece

[Stefanoo]

Significant update to MasterPiece, servo now has been simplified due to the reference point of the sources on both halves to GND through the same resistor.
Conceptually works and so does in simulation. I will have to make sure it works on the prototype as well as soon as I have that up and running.

Here is the schematic MasterPiece-R5

[Stefanoo]

And here is the noise simulation for it.

[Joachim Gerhard]

Looks good. When it works in simulation it should work in praxis too.

[Stefanoo]

Originally Posted by Joachim Gerhard

Looks good. When it works in simulation it should work in praxis too.

Thank you very much.

Once it is implemented I will post spectral density's measurements, frequency accuracy and distortion measurements.

[Stefanoo]

How much headroom should I be able to get at the output on the upper band?

I suppose this is current dependant, what do you consider a trade-off?

[FdW]

Originally Posted by Stefanoo

Significant update to MasterPiece, servo now has been simplified due to the reference point of the sources on both halves to GND through the same resistor.
Conceptually works and so does in simulation. I will have to make sure it works on the prototype as well as soon as I have that up and running.

Here is the schematic MasterPiece-R5

I like the fact that it now has a s single servo, but be aware If one side (of the diff amp) has an offset different from the other side it will set one side to -((offset(left)-offset(right))/2) and the other side will be set to +((offset(left)-offset(right))/2). To prevent this you need to be able to set the static-offset-difference to zero (volts difference between left and right side).

[Stefanoo]

I had that problem if you have a feedback resistor connected between the two halves you will get a common mode offset and it is nearly impossible at this gain to have a 0 differential and absolute offset by trimming the absolute.
And this can be proven on the simulation by unbalancing the two halves i.e. superimposing a different offset among the two sides.
So using the differential configuration carried on in the previous version, would slam the differential output two 0 and the absolute to a certain value that can be trimmed down to 0 acting on the house keeping current source's section.

In this arrangement however, simulation proves that not only the differential offset is 0 for whatever unbalancy you have, but so is the absolute.
If you think true, this is conceptually correct as on both sides the souces are now connected to the same potential generating a common mode voltage to correct will equally correct both of them and since the integrator is pointing at 0 thus the output voltage will be corrected to 0V.

I will post simulation results to support this statement.

[Stefanoo]

Here it is:

As you can see I have deliberately made the two side very different in terms of output offset however both sides are corrected down to 0.

In the first versions I would have gotten with this configuration a common mode offset.

In this arrangement, things change, take a look here.

Simulation seems to be set up properly thus I would hope that this is going to reflect the reality as well.

[FdW]

Originally Posted by Stefanoo

Here it is:

As you can see I have deliberately made the two side very different in terms of output offset however both sides are corrected down to 0.

In the first versions I would have gotten with this configuration a common mode offset.

In this arrangement, things change, take a look here.

Simulation seems to be set up properly thus I would hope that this is going to reflect the reality as well.

I'm not (yet) totally convinced, would you do the same simulation with both side a negative start offset (not one positive and the other negative). Already it looks better than I expected maybe I should have given it a bit more time (but I do not have that much of it (time) ).

[Stefanoo]

here you go....



Isn't it really beautiful? So far so good...see if what I have studied works in practice too...