How to simulate an OLG of the TMC amp

[dadod]

For same time I am trying to get a clear answer of the question. What is correct way to insert Middlebrook probe when simulate OLG of the TMC amp.
In most examples TMC resistor was connected direct to the output and not after Middlebrook probe. In this case OLG diagram looks very similar to CMC amp. My opinion is that if OLG simulation was done in this way you are not getting whole picture. How to explain distortion decrese at higher frequencies 10kHz or 20kHz. Phase behavior will not show whole stability problem.
If TMC resistor was connected inside global loop, after Middlebrook probe OLG diagram will be similar to TPC OLG digram(and this explain why TPC and TMC will show similar distortion figures.
Here is a table with simulated result for TPC, TMC with Middlebrook probe connected as I think it shoud be, TMC with common connected Middlebrook probe and last is CMC.
Fot each kind of a compensation there is didtortion in ppm and CLG per fequency.
CLG for the amp was set to 28dB.

Next are OLG digrams as follows:
TPC, TMC my way, TMC "standard" and CMC. The same amp was used for the simulation.

I appreciate comments.
dado

[Neutrality]

Well, this should be interesting.(I'm a little uncertain about this as well)

Also CLG. See LTSpice, Closed loop gain?

[dadod]

Originally Posted by Neutrality

Well, this should be interesting.(I'm a little uncertain about this as well)

Also CLG. See LTSpice, Closed loop gain?

CLG is simple. You set off Middlebrook probe(AC=0), and set AC=1 for signal generator.
dado

[Neutrality]

Originally Posted by dadod

CLG is simple. You set off Middlebrook probe(AC=0), and set AC=1 for signal generator.
dado

Well, I knew that, that is what I have always done, but as my topic explains it does not work that well for me. So I thought I might have done something wrong.

[obseedian]

I think both methods are correct, but they measure different things. The standard probe includes both the common mode and differential mode signals (between VAS output and OPS output, I'm not talking about the input signals here). Since the OPS gain is unity or slightly less, the common mode signal is much larger than the differential mode signal, hence you get a loop gain plot that looks almost identical to the CMC plot.

The alternative probe measures the loop gain as seen by the differential signal and here it is correctly showing a second order rolloff because TMC behaves like TPC to for the differential signal.

That's how I see it anyway, but maybe I'm wrong.

[dadod]

Originally Posted by Neutrality

Well, I knew that, that is what I have always done, but as my topic explains it does not work that well for me. So I thought I might have done something wrong.

Here is CLG of the TMC amp and I don't see a problem.
dado

[Neutrality]

Originally Posted by dadod

Here is CLG of the TMC amp and I don't see a problem.
dado

Hmm, when your CLG reaches unity gain at 4 MHz your Phase is -165/170 degrees, isnt that a problem?

[dadod]

Originally Posted by Neutrality

Hmm, when your CLG reaches unity gain at 4 MHz your Phase is -165/170 degrees, isnt that a problem?

No, phase margin at 0dB you see from OLG diagram not CLG.
dado

[dadod]

Originally Posted by obseedian

I think both methods are correct, but they measure different things. The standard probe includes both the common mode and differential mode signals (between VAS output and OPS output, I'm not talking about the input signals here). Since the OPS gain is unity or slightly less, the common mode signal is much larger than the differential mode signal, hence you get a loop gain plot that looks almost identical to the CMC plot.

The alternative probe measures the loop gain as seen by the differential signal and here it is correctly showing a second order rolloff because TMC behaves like TPC to for the differential signal.

That's how I see it anyway, but maybe I'm wrong.

I see what you mean, but mostly all here are using standard probe insertion and in this case you don get how much of NFB you are using at certain frequences.
I prefere the "alternative" probe insertion as I get more needed information from the plot.
dado

[Neutrality]

Originally Posted by dadod

No, phase margin at 0dB you see from OLG diagram not CLG.
dado

So phase on the CLG diagram could be ignored? It is the amplitude 0 dB point from the CLG diagram you then find on the OLG diagram and then you get your phase from that?

So if the CLG diagram shows an amplitude 0 dB point(unity gain) at 4 MHz you look for the phase at 4MHz on the OLG diagram?

Is that right?

Could anyone confirm this, if so that would be huge news for me, but good news nonetheless.