The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> A basic question about comparator
https://designers-guide.org/forum/YaBB.pl?num=1323041286

Message started by lunren on Dec 4^th, 2011, 3:28pm

Title: A basic question about comparator
Post by lunren on Dec 4^th, 2011, 3:28pm

Hi,

I recently saw several cases in which a miller compensated amp is used as comparator. I think there are some difference in terms of PSR, noise and speed between with and without miller compensation cap. But I don's think in these applications, noise, PSR is critical (supply monitor). Then in my opinion, the miller compensated amp is just making the circuit to response slowly, but this doesn't seem a advantage. I am wondering what's the advantage to use miller compensated amp as a comparator?

Title: Re: A basic question about comparator
Post by lunren on Dec 4^th, 2011, 3:44pm

I guess miller compensation cap makes the circuit more stable, meaning the output of the comparator won't switch unless the difference between the inputs of comparator is big (in magnitude) and long (in time) enough.
More comments are welcome.

Title: Re: A basic question about comparator
Post by Alexandar on Dec 5^th, 2011, 5:38am

I think the reason is that when you place in a negative feedback loop, it doesn't oscillate. So in that way you can calibrate your comparator.

Title: Re: A basic question about comparator
Post by buddypoor on Dec 5^th, 2011, 6:55am

lunren wrote on Dec 4^th, 2011, 3:28pm:

Lunren, I agree cause you are right.
You can use a frequency-compensated opamp as comparator - however, the compensation lowers the slew rate and, thus, the response time.
For stability reasons (tendency to oscillation) it is NOT necessary to use such a compensation scheme since there is no operational point within the linear part of the transfer characteristic.

Title: Re: A basic question about comparator
Post by raja.cedt on Dec 5^th, 2011, 7:19am

hello lunren,
up to my knowledge there is no need to use miller compensated op amp as first stage to comparator, because you are not going to keep comparator in a -ve feedback loop. have you seen any body did like this, if so please give me the reference?

as people explained in the earlier posts, Some times during offset cancellation comparator being kept in loop but it will be very very slow loop.

Thanks,
Raj.

Title: Re: A basic question about comparator
Post by loose-electron on Dec 5^th, 2011, 9:37am

A comparator requires no gain-phase compensation for phase margin stability.

It is not used in a feedback loop system.

A device designed as an op-amp, with compensation can be used as a comparator.

The compensation will slow down the response of the comparator.

A comparator put into a switched mode structure used for offset cancellation is generally not a linear feedback system. Rather, it is determining the amplifier offset characteristice for a fixed input (zero) and forcing that offset voltage to do something (cancel the offset) but that is not done in a feedback loop generally. One node (or differential pair input) is forced and held at a voltage.

Title: Re: A basic question about comparator
Post by raja.cedt on Dec 5^th, 2011, 1:52pm

hello looselectron,

i didn't understand your comment "offset cancellation is generally not a linear feedback system", let us say you want to implement input off storage then comparator will be connected in unity gain configuration.

Thanks,
raj.

Title: Re: A basic question about comparator
Post by loose-electron on Dec 5^th, 2011, 2:58pm

Its a swiitched capacitor structure where the two capacitively coupled input are tied together, and the feedback brings the capacitance to the equivalent offset voltage.

Thats not a linear function is it?

Title: Re: A basic question about comparator
Post by lunren on Dec 5^th, 2011, 6:21pm

raja.cedt wrote on Dec 5^th, 2011, 7:19am:

Sorry I don't have a reference, but I do see a lot comparators used in real time clock circuit have comparator (to detect supply drop) with miller compensation. I guess with miller compensation, if the drop below certain threshold level is small and short (like a glitch), it doesn't make sense to generate a power drop signal. Unless the drop is big, especially last long time, then it should generate power drop signal.

Title: Re: A basic question about comparator
Post by loose-electron on Dec 6^th, 2011, 2:22am

then you are intentionally slowing the response to essentially LPF what you are monitoring.

Title: Re: A basic question about comparator
Post by lunren on Dec 6^th, 2011, 10:49am

Yes, I think that's the reason here.

Title: Re: A basic question about comparator
Post by thechopper on Dec 6^th, 2011, 5:32pm

Agree with Jerry, specially for small input signa slopes, the comparator might tend to chat since the decision circuit has no hysteresis, therefore LPF is required to avoid potential chattering under these conditions, and this is provided by the Miller compensation.

Best
Tosei