Designer’s Guide Community

Cri Azzolini

Community Member

Offline

Posts: 48
University of Parma, Italy

Stability analysis on instrumentation amplifier
Jan 16^th, 2007, 6:54am

Hi all,

I would like to verify the stability of the instrumentation amplifier shown in figure below. Using the Spectre STB analysis, I verified the stability of the local loops corresponding to each op-amp. For instance, the voltage source V1 in figure is the probe for STB analysis on OP2 (OP1 and OP3 are considered feedback elements and load); afterwards I removed V1 and I placed V2 as a probe for STB analysis on OP3 loop (in this case OP1 and OP2 are the load).

If both simulations reported good stability margin, can I draw that the WHOLE instrumentation amplifier is stable? Should I perform more simulations?
Many thanks in advance and bye.
Cri

Back to top

IP Logged

MichaelTian

New Member

Offline

Posts: 3

Re: Stability analysis on instrumentation amplifie
Reply #1 - Jan 19^th, 2007, 12:14pm

Cri,

There are four loops in this circuit: the loop thru. OP1/upper R2, OP2/lower R2,
OP3/upper R4, OP3/lower R4. The first two are single loops, the last two are
differential loops.

What you can do is to put a probe on the upper R2 and lower R2 branch, to asses the
stability of the first two single loops. Then use the cmdm probe (available in analogLib)
and place that on the upper R4 and lower R4 branches to assess the differential mode
and common mode loopgain for the differential circuit.

If all loopgain shows good stability margin, you should have confidence on the design.
Also you can run pz analysis to double check the whole circuit.

Mike

Back to top

IP Logged

mg777

Senior Member

Offline

Posts: 131

Re: Stability analysis on instrumentation amplifie
Reply #2 - Jan 22^nd, 2007, 8:02am

Here's a related question I've often posed to myself - would an amplifier just below the threshold of oscillation exhibit a greater noise at its output? My reasoning is that the amplifier's 'tank circuit' would be repeatedly energized by thermal noise and break into little jiggles, which of course get damped out as the amplifier is still stable. Note that this noise wouldn't be white at all - in fact it'd have a sharp spectrum, but in a broadband circuit the overall fluctuations should increase.

Comments?

M.G.Rajan
www.eecalc.com

Back to top

IP Logged