stb analysis of a fully diff amp

You would use a diffstbprobe when you have a differential loop (i.e. two feedback paths). What you're doing is measuring the loop gain at that point - so it's usually fairly obvious where you place it - it needs to be a point where there is a single path (in each of the two halves) - if you have another path which is bypassing the route through this component, you're not analysing the loop in the right place.

The diffstbprobe can be controlled from the stb analysis form - on the form you can choose whether it analyses the stability of the differential signal or the common-mode signal in that loop.

If you have a separate common-mode loop, you'd place an iprobe in that loop too, and then use the stb analysis to separately analyse the stability of this second loop.

Regards,

Andrew.

Thanks Andrew for your prompt response. I have tried diffstbprobe and it works fine. I have also placed the iprobe in the CMFB loop and checked the stability. Though, I am trying to find a way of checking the open loop gain-phase response of the amplifier from the same testbench. Any ideas?

Kind Regards,

Costas

Costas,

Can't you just plot the loop gain from the stb analysis (it's on the Direct Plot->Main Form)? That's generally what you really want, rather than trying to bias the circuit to actually function correctly open loop (given that I assume you're not going to actually use it open loop). 

Or maybe I've misunderstood your question...

Regards,

Andrew

Hi Andrew,

Yes, I can plot the loop gain from the main form, but what I also want to do is to plot the open loop gain of the amplifier individually . This may not be feasible from the same testbench, since the amplifier is already connected in a feedback loop. Unless, I will remove the RC feedback and conect it as a unity-gain amplifier (but in this case I have modified my testbench).

I hope my question is clear.

I know that this is a more technical question but anyway thanks.

Costas

Costas,

If you really want to simulate it open loop, then really you need a different testbench. Normally the loop gain is similar to the open loop gain (although that of course depends on your feedback loop, which from what you said is changing the loop response).

You could use the spt1switch/spt2switch etc components from analogLib which allow you to configure connectivity differently for different analyses - although not sure that's quite what you want either.

Or you could use the same testbench and have two different config views (created with the hierarchy editor), and use the "Bind to open" capability to omit certain components in different simulations - you'd have two ADE states (or two ADE XL tests) to do this.

Regards,

Andrew.