How can we increase the speed of a CMOS hysteris comparator ?

[vbhupendra]

How can we increase the speed of a CMOS hysteris comparator ?

 

[Syukri]

Re: Comparator Speed

Comparator response can be improved by adding hysterisis equal to or greater than amount of largest expected noise amplitude.

Ways to introduce hyterisis is by external method and internal method..

Maybe the speed is better with internel method coz it does not require external feedback.

 

[vbhupendra]

Re: Comparator Speed

Comparator response can be improved by adding hysterisis equal to or greater than amount of largest expected noise amplitude.

is hysterisis has impact on speed ?

 

[jcpu]

Re: Comparator Speed

Syukri

Comparator response can be improved by adding hysterisis

please provide some reference material.

thanks in advance,

 

[sunjiao3]

Re: Comparator Speed

Maybe the speed is better with internel method coz it does not require external feedback.

Why? Why extrernal feedback may cause the drop of speed?

 

[suhas_shiv]

Comparator Speed

As far as I know hysteresis helps in reducing the effect of noise. To improve speed, the current must be increased.

 

[sunjiao3]

Re: Comparator Speed

However, as the tail current increases, the width of the hyesteresis window increase.

 

[kumard35b]

Re: Comparator Speed

you can increase the gain of your input stage as much as possible high enough ,so that you can achieve your speed in the forthcoming stages.

 

[jwfan]

Re: Comparator Speed

Comparator response can be improved by adding hysterisis equal to or greater than amount of largest expected noise amplitude.

is hysterisis has impact on speed ?

Yes
The hysterisis means there is positive feedback, which will increase the speed

 

[sengyee88]

Re: Comparator Speed

A few ways to improve rise/fall time of comparator:
1. Increase comparator gain. ie gm, Rload
2. Increase current.
3. Positive feedback.
4. Add buffer (inverters)

The hysterisis means there is positive feedback, which will increase the speed

Positive feedback doesn't means having hysterisis.

 

[jwfan]

Re: Comparator Speed

Positive feedback doesn't means having hysterisis.

But having hysterisis means have positive feedback

 

[ch1k0]

Re: Comparator Speed

What kind of comparator ? Continuous or clocked ?

If continuous, what you really really want is to have fast time constants, so use a cascade of low-gain, fast stages (say 3 or 4) to obtain the fastest possible response.

If clocked, make the latch devices ultra-short and keep capacitances small, so that the regeneration is fast. Your offset will be bad, but it's possible to design around it.

Hysteresys helps reject noise, in continuous comparators. It's a problem in flash ADCs, so it depends on what you want the comparator for. If there's hysteresys, then you have positive feedback, but you won't always have better speed because of hysteresys in your comparator...

 

[melc]

Re: Comparator Speed

How looks a "clocked" comparator ? It's the first time I heard this terminology.
A comparator is a comparator. If has other tricks to increase the output switching speed and minimise jitter is not anymore a simple comparator but a comparator followed by a digital logic circuit...

 

[sengyee88]

Re: Comparator Speed

You may refer to the book Analog Integrated Circuit Design, David Johns & Ken Martin regarding detailed clocked/latch comparator.

 

[aborigini]

Comparator Speed

I designed the comparator with hysteresis internal, but I found the gain of the comparator is low. Is it supposed to be that way?