continuous time comparator

Hello everyone! My question is general, and i want any ideas, no matter how basic they are, or links to papers or whatever you have to offer me I am designing a continuous time comparator (one input looks a threshold, the other looks a slow varying signal) , so i have no external clock (not a discreet time comparison) but i am continuously comparing. My only solution was just an amplifier with too much gain. But that is not stable over temperature, which is a problem. So any idea is welcome. I care more about over temperature stability, than the fast response of the comparator output.

Any clever topology or anything please let me know

thanx

PS i also do not care about hysteresis. Even a comparator with zero hysteresis is ok, thx to the topology of the whole design

Thanx again

dear geozog86,

Regarding analog comparision, I think that any solution you choose, must be based on saturation, in order to provide boolean output.
And, inevitably the equivalent circuit model will decay on comparator topology.

+++

so what's wrong with just a generic voltage comparator chip, like a lm311 or lm393? What kind of accuracy do you need?

It's also no clear, why an OP shouldn't be "stable over temperature". Not generating clear logic levels near the threshold would be the only problem, which can be avoided by introducing a small hysteresis. If the switching speed of an OP (e.g several us) is sufficient, there should be nothing against it.

@andre: ok, i agree it will be based on saturation, but you say "will decay on comparator topology". So my question is what is the comparator topology in general? i have been using successfully a voltage controlled voltage source with saturated output.... so is the "topology" just an amplifier? nothing clever can be done on that?

@mtwieg unfortunately i'm working with low voltage (maybe i should have put the thread on IC design) cause i'm having a threshold of 500mV and I reach it in 17usec, so i'm rising slowly (29.4*10^3 V/sec) ....So it'll all be transistor parameters limited

@Fvm: Don't laugh, but with OP you mean....?? a gain stage?? What i want is the rising slope of the comparator when it changes state to cross the threshold at which my memory that follows change state, at the same time, no matter what temperature i'm operating. I mean, i want to change my memory for example, every cross+50nsec. I dont care when the comparator will start changing or if it will change faster for higher temps than in low (as is happening). But that the point of changing state in my following memory will be the same all the time. Maybe i'm asking too much from my circuit. Plus i don't want hysteresis cause I have a memory following so it's kinda another memory before my memory, isn't it? Hysteresis is implemented by some kind of memory action (delaying, but memorizing the previous state even though i have "changed"). I wanna decouple the two actions of comparing and keeping the state let's say. And simply want the comparing action (like the usual ADC comparator keeps the state in a latch....i do have a memory block following.....) Hope i'm making myself understood, sorry if i'm making things worse

geozog86 said:

@mtwieg unfortunately i'm working with low voltage (maybe i should have put the thread on IC design) cause i'm having a threshold of 500mV and I reach it in 17usec, so i'm rising slowly (29.4*10^3 V/sec) ....So it'll all be transistor parameters limited

Click to expand...

That's not slow or low enough voltage for transistor parameters to dominate. What kind of overdrive voltage do you want to respond to?

@Fvm: Don't laugh, but with OP you mean....?? a gain stage?? What i want is the rising slope of the comparator when it changes state to cross the threshold at which my memory that follows change state, at the same time, no matter what temperature i'm operating. I mean, i want to change my memory for example, every cross+50nsec. I dont care when the comparator will start changing or if it will change faster for higher temps than in low (as is happening). But that the point of changing state in my following memory will be the same all the time. Maybe i'm asking too much from my circuit.

Click to expand...

You're not making much sense here... do you want the propagation delay to be very short, or do you want it to just be consistent over temperature? I'm not sure what kind of "memory" you're talking about or how it factors into this.

maybe i should have put the thread on IC design

Click to expand...

Yes, obviously. This also implies, that you should clarify the available IC technology. Otherwise, most suggestions will be void. I also understand that "temperature stability" mostly refers to dynamical behaviour, so in contrast to my understanding of your previous statement about fast response, precise dynamical behaviour is most critical. I guess, you should show example circuits, that you have already tried to illustrate your intentions.

I assume, that someone among the edaboard members engaged with IC design can possibly help.

geozog86 said:

...my question is what is the comparator topology in general? i have been using successfully a voltage controlled voltage source with saturated output.... so is the "topology" just an amplifier? nothing clever can be done on that?

Click to expand...

Hi,

Coud you provide the circuit topology you´re using ?

Now it´s clear why you´re worried about temperature interference.
Depending on based circuit, gain really may be not so high, but I think temperature range involved must be too wide to change gain enought to change circuit behaviour.

+++

Guys, thank you all, but i'm opening a thread on IC to make things more clear, cause this one was not a good choice

Thread closed. See here https://www.edaboard.com/threads/204716/