Comparator using internal hysteresis

comparator internal hysteresis

Hi,

How to design a comparator with internal hystersis for POR? With one input from BGR and other from resistor divider. Assume BGR is negative input and resistor divider is positive input.

What value should you set for positive threshold? Do you have to worry about negative threshold for POR?

Thanks

comparator with internal hysteresis

The negative threshold would decide , when the POR triggers again if VDD starts to fall.
You could look into maxim site, one of theri application notes talk on it.
For example:
If VDD goes below VDD-10% POR should trigger again and generate the reset . this reset stays applied till VDD does not cross the positive threshold. Once the positive threshold is crossed, the reset still remains applied for certain duration say 150ms and then is removed. All is subject to condition that VDD does not go low again

Added after 7 minutes:

You can go through the following application note from MAxim.

Hi,

Right now my comparator uses push-pull type and I found out that glitches occurs at the output under certain corner and temperature during start-up when I ramp the supply voltage. I would like to eliminate this glitch so I was told to use hysteresis comparator. Will the hysteresis comparator do the job? My reset threshold is set at 1.3V and I am working under 1.8V supply voltage. My POR is for analog domain use only.

So under this constraint, I should set the positive threshold at 1.3V right? And the hysteresis comparator will automatically avoid glitch in the output during startup. Would it make sense to set the negative threshold at 1.3V as well?

Thanks

No, please donot do that. The basic aim of hystersis is lost
I am not clear about somthing, Your POR generates a rest pulse once VDD crrosses 1.3V. This pilse must stay for the time till the supply voltage has stabalised and if used for a microprocessor, the data needed to be loaded has been done also the clock has stabalized.
One would like that POR must reset the circuit if supply goes below the min. required i.e 1.62 if we use VDD-10%.
Are u doing a resistor division of VDD and the feeding it as one input to comparator or are u directly feeding VDD as input.
I would prefer positive threshold to be around 1.7V to 1.75V and negative threshold to be 1.62V to 1.6V.
The application note that i have uploaded would surely help u in deciding the positive and negative threshold.
If u want I could pass some other papers on POR designing

Yes please post more papers on POR design.

Basically my comparator takes input from a resistor divider and a bandgap voltage.
So my requirement is to reset until supply reaches 1.3V, after that the POR goes low. I will think about changing the reset threshold. So my other question is whether the hysteresis will eliminate glitches that I am getting during startup (refer below).

What u see is Vdd ramp (0-1.8V), the inputs to the comparator (resistor divider and bandgap) and POR output.

Thanks

In fact I could just stick a schmitt trigger into my current design such as the one below. However I am not sure how this circuit work. Can anyone provide an analytical explanation using equations and also how to size the transistors to achieve positive and negative threshold?

Sorry I don't have the Baker book.

Thanks

Dear ccw27,
I have posted some of the papers at following link


Added after 17 minutes:

For fig schmitt.jpg.
1. Concider Vout=VDD Vin =0
M1 and M2 are off M3 On.
As Vin increases M1 starts to turn On pulling the voltage at source of M3 down. When Vin= Vth2+Vs2 M2 turns on dropping output and turning M3 off.
At this point current through M1 and M3 are same.
Equate the 2 currents and get one of your switching points. Follow thw same with other case.
Vtrigger (output from high to llow)= V1
(W/L)1=[(Vdd -V1)/(V1-Vthn)]*(W/L)3
Vtrigger (output from low to high)= V2
(W/L)5=[(V2)/(VDD-V2-Vthp)]*(W/L)6

From bakers book (W/L)2 ≥ 5*(W/L)1

Added after 9 minutes:

For schmitt2.jpg
Assume Vout=vdd thus Vx=Low and Vin =high.
As Vin starts to decrease M2 starts to turn On and current flows through M2 to M3.
The pull up and pull down strength of M3 and M2 shall decide the Vin at whch output moves from hihg to low.
We could equate the currents in M2 and M3 and get the required W/L rations.
You could follow the same for other case
hope it helps

You can use an uncompensated OPAMP with positive feedback to work as a hysterisis comparator. I think that it will be less noisy than the current circuit.

Hi ambreesh,

Thanks for your help. Actually after simulating more carefully I found out the main cause of the spike was due to inverter chains after the comparator. So right now I am using schmitt2.jpg above and seems to be working, but I still need to do more simulations. Does anyone by anychance know which schmitt trigger is more popular for low voltage supply?

Thanks