The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> hysteresis in a comparator
https://designers-guide.org/forum/YaBB.pl?num=1233720551

Message started by arunm on Feb 3^rd, 2009, 8:09pm

Title: hysteresis in a comparator
Post by arunm on Feb 3^rd, 2009, 8:09pm

I am designing a continuous-time comparator. The typical signal to it is a ramp. The ramp slope can be really small. I want to make sure that the output doesn't toggle near the comparison point. But I am not sure how much hysteresis to put in.
Disturbances from power supply and other coupling can be quantified, but what about noise? Whatever hysteresis I put in, if the input signal slope is really small, then low frequency flicker noise can make the output toggle right?
I would like to know if this is really a practical concern, and if it is, how do we go about solving this?

Title: Re: hysteresis in a comparator
Post by viks_iisc on Feb 4^th, 2009, 2:28am

Why don't you use a preamp before the comparator. It will solve maximum of your problems. Well it depends on the exact application but usually preamplifiers come handy in these problems. :)

Title: Re: hysteresis in a comparator
Post by loose-electron on Feb 10^th, 2009, 8:45pm

Agreed - you should get the signal amplified up to the point where offsets, flicker/thermal noise are no longer a concern. Suggest that you define the limits and characteristics of your input signal and then go from there.

Also, any comparator has limited gain, its not infinite, so you will be dealing with a linear response if you are trying to make a decision with a small input.

Title: Re: hysteresis in a comparator
Post by thechopper on Feb 15^th, 2009, 6:24pm

Hi,

I also agree with the pre-amp idea. Nevertheless, here are a couple of additional thoughts:

1) Quantify the incoming noise: assuming it is gaussian noise, either flicker or white noise, the pk-pk noise will be given by 6σ. Therefore you could quantify the hysteresis of your comparator based on this number as a minimum required.
2) In case the incoming noise is larger than what you can afford for the comparator hysteresis and for any reason you cannot use a preamp, if it is flicker noise you could chop the amplifier (and you should clock your comparator too) so that you remove the flicker component from your signal band (in addition you would be removing the comparator offset component too).
3) You could add an analog delay at the comparator output: for example integrating a current into a cap will allow filtering some of the noise driven toggling your comparator might experience. This delay might or might not be an option depending on your application
.

Hope this also helps
Tosei

Title: Re: hysteresis in a comparator
Post by arunm on Feb 15^th, 2009, 8:42pm

Thanks a lot for your replies.