Suitable Window Comparator Circuit for Power supply from 10VDC till 30vdc

Thread Starter

globalone

Joined May 27, 2013
22
Hello,

I am searching for a window comparator that can have :
- 1st REF: If main power supply drops less than 21 volts then output goes LOW
- 2nd REF: If main supply reaches higher than 30 volts dc then output goes LOW.
- Equal external hysteresis between both comparators with 100mv hys.
- low continous qiescent current at comparator supply V+ voltage (in micro seconds).
- Fast Response in nano seconds if possible .

If input voltage rangs between 10VDC till 30vdc or higher! , then which comparator i should use ... the 5v comparator or high voltage such as 36vdc comparator ?.

Should i use zener diode as a reference to the comparator and if you suggest to use 5vdc comparator , then the zener diode wont be the best choice for being a stable regulator . If i choose a normal regulator then it will add a very long start up delay and that wont work with my circuit , because i am building an oring circuit driving mosfets as back 2 back to switch between main power supply and auxiliary power supply .

This is the circuit that i have already made but i am not happy with the results .

- Hysteresis are not equal .
- "R_MON1" shows that the supply current of comparators is wasting 16mA all the time which is not good for a battery if it was an auxiliary supply!. Total wasted current 32mA for both comparators.
- D1 and D4 are not accurate due to the changing at the power supply since i have already mentioned that the input voltage might change from 10VDC till 30VDC or higher that is why i chose to use window comparator for validating input voltage .
- I am using 5vdc ultra fast comparator ~4ns (continuous 16mA).

This is the stable zener regulator that i could simulate with 5mV stability :


I am really confused and i hope i can find someone who will guide me or assist me to finish this circuit .

Thank you
 

crutschow

Joined Mar 14, 2008
23,487
An inexpensive TL431 adjustable reference can be used to provide a stable reference voltage.
low continous qiescent current at comparator supply V+ voltage (in micro seconds).
Do you mean microamps?
Fast Response in nano seconds if possible .
Why do yo need ns response?
That will add expense and complexity to the circuit.
If you are dealing with power supply voltages, their response time is usually in the millisecond time-frame, not nanosecond.
 

Thread Starter

globalone

Joined May 27, 2013
22
Hello crutschow,

I meant by ns "Nano Seconds" and I am sorry i meant by micro seconds as micro Amps (Continuous )

So If you suggest me to use TL431 as an adjustable reference that means i can only use this as references to comparators , but i still need a fast switching regulator that can supply both comparators . Would you suggest me to stay using LT1711 as a comparator (continuous 16milli Amps / ~4 nano seconds response time). 2 commparators would waste 32 mA of current as always as long as power supply is ON! and i dont want it to be like that since the power supply might be changed to a battery so i dont want to discharge the battery quickly.

i liked Lt1017 because of its bias supply current (100uA continuous) , but it is really slow.

Any suggestion ? .
 

Thread Starter

globalone

Joined May 27, 2013
22
I wanted to switch within nano seconds , because i want to turn off mosfets quickly just to avoid cross conduction between both power supplies . I can later attach my full schematic . I have been building an ORING Power Supplies with long researches.
 

crutschow

Joined Mar 14, 2008
23,487
I wanted to switch within nano seconds , because i want to turn off mosfets quickly just to avoid cross conduction between both power supplies . I can later attach my full schematic . I have been building an ORING Power Supplies with long researches.
Then you are trying to solve the wrong problem.
You don't need a fast comparator response time, you just need the drivers that control the two MOSFETs to be fast, (or generate a small non-overlap between the two signals) once the comparator has switched (which can be much slower).
 
Last edited:

Thread Starter

globalone

Joined May 27, 2013
22
Thanks for your suggestion .

Could you please provide me of a good comparator from Linear Technology so I can simulate it using LTSPICE .

I will attach my mosfet driver later with more questions too .

I appreciate your help.
 

Thread Starter

globalone

Joined May 27, 2013
22
I still need to solve hysteresis to be equal between comparators . 120mV of Hysteresis would be nice to avoid noise .

How to add equal hysteresis to my circuit ?.
 

crutschow

Joined Mar 14, 2008
23,487
Could you please provide me of a good comparator from Linear Technology so I can simulate it using LTSPICE .
The LT1017 should be okay if you want low power.
How to add equal hysteresis to my circuit ?
You need the same resistor value between the output and the plus input, as well as the same value resistance from the plus input to the rest of the circuit, for both the comparators.
 

crutschow

Joined Mar 14, 2008
23,487
Below is the LTspice simulation of a window-comparator circuit with non-overlapping outputs that I previously designed, to illustrate what I mean about having the OR'd comparator outputs drive both MOSFETs.
The non-overlap between outputs A and B is about 2.5μs, as determined by the C1R9 and C2R10 time-constants.
The non-inverting gates in parallel at the output are to help drive the high capacitance of a MOSFET gate.

upload_2017-10-2_18-4-57.png
 

Attachments

Thread Starter

globalone

Joined May 27, 2013
22
What a wonderful circuit . Thanks for sharing your circuit .

I am sorry for the delay . I've been a little busy , but i simulated TL431 with LTSPICE and i didn't like its behavior so i have replaced it with "LT1009" . I have also noticed LT1017 needs about 200 uA quiescent bias current and it is really slow and it seems if you add a pull up resistor , it fixes the situation , but it still not stable at all . For example if the comparator LT1017 is at LOW output then i see switching spikes while the power supply or VDD is turning off or ON . I replaced LT1017 with LT1716 since it has a self quiescent current ~42 micro Amps and it switches within 4 micro seconds and it is quite stable at its output stage if the output is at LOW . It does not need any pull up resistor .

I would like to mention that the main power supply might reach 31 Volts DC or 30 Volts DC . The good thing is that LT1009 and LT1716 can handle such high voltage so i am not worried , but i am so much worried about your selected NOR GATES "CD4001" . This NOR GATE can handle up to 20 Volts DC VDD.

Regarding "A2" and "A3" , I still have not found any part number to accept up to 36VDC NOR LOGIC GATE , AND LOGIC GATE to switch on/off LEDS through NPN transistors. What do you suggest ?.

With the below circuit , i can adjust comparators for low-reference 21VDC and high-reference 30VDC , but without hysteresis . How can i calculate a hysteresis resistor of about 120mW ?.

 

Attachments

Thread Starter

globalone

Joined May 27, 2013
22
Please see my full project simulation file with LTSPICE . i still need to update comparators and replace them with LT1716 and add the reference part number LT1009.

I could create a dead time between power supplies using XOR GATES and give the priority to the main power supply even if both are ONLINE. I can discuss the problems later with you if you don't mind :) .
 

Attachments

Thread Starter

globalone

Joined May 27, 2013
22
BTW you will see that i have built my own fast mosfet driver using push pull and level shifting with little continuous 15 micro Amps due to voltage divider R4 and R14 and speed up switching through Q3 which will force turn off quickly . I am sure D12 is to clamp the voltage and protect the mosfet gate , but it might break down due to high power dissipation that is coming from "R38" controlled by "Q17" which speed up turning On .

I have implemented Reverse Voltage Protection by installing a comparator to hold the channel OFF unless no reverse voltage detected!.
 

crutschow

Joined Mar 14, 2008
23,487
I would like to mention that the main power supply might reach 31 Volts DC or 30 Volts DC . The good thing is that LT1009 and LT1716 can handle such high voltage so i am not worried , but i am so much worried about your selected NOR GATES "CD4001" . This NOR GATE can handle up to 20 Volts DC VDD.

Regarding "A2" and "A3" , I still have not found any part number to accept up to 36VDC NOR LOGIC GATE , AND LOGIC GATE to switch on/off LEDS through NPN transistors. What do you suggest ?.
I suggest reducing the voltage with an LM317 regulator to perhaps 10-15V to power the electronics.
That way you won't have to worry about varying or excess voltages for component operation.
That will also stabilize the hysteresis value.
 
Last edited:

Thread Starter

globalone

Joined May 27, 2013
22
What are the capacitor's values across the window comparator circuit and where i should put them ?.
should i limit the current or just add a safety resistor biasing comparators something like "RMON1" , "RMON2". I have added them just to monitor the current through the simulator program . I have also noticed that by adding a capacitor at the bias input comparator does not affect anything unless i add a series 1k resistor near the capacitor. Very strange if i add a capcitor near any wire then the capcitor sucks Very High Current of 100 Amps charged. Is this real ? or just the simulation . I had to add a resistor with the cap. (RC) to limit the very high current (High Inrush Current).

Can you please explain to me why I see a really high inrush current if i add capacitors and the more i add capacitor to the circuit then the inrush current goes high and high 300Amps or 1KAmps . I am not sure if this is real!.
 

crutschow

Joined Mar 14, 2008
23,487
an you please explain to me why I see a really high inrush current if i add capacitors and the more i add capacitor to the circuit then the inrush current goes high and high 300Amps or 1KAmps . I am not sure if this is real!.
It's not totally real.
The simulation has an ideal supply and wires with zero impedance and the supply has no current limit.
In the real circuit there is some resistance which limits the inrush current.
Normally that inrush in not a problem, unless the source is sensitive to momentary high current draw.
Most aren't.
 

Thread Starter

globalone

Joined May 27, 2013
22
I have simulated LM317 , but I couldnt find what is its self quiescent current LM317 .. I have also found that the adjustable resistors are wasting current through the voltage divider in few milli amps . I decided to use a step down regulator that has a self quiescent current in few micro amps with enable , disable pin ..

I need to turn off regulator at Auxiliary Supply when the main supply is ON since i am giving a priority to the Main Supply even if both supplies are on .

For Example ; The Auxiliary Power Supply can be a 24vdc battery .. If main power supply is ON and the load is switched to the main power supply , but the battery is always ON and waiting for the main supply to turn off . In this situation , the regulator will keep dissipating power as long as the battery is still Alive even if the load is not switched to the battery (AUX SUPPLY). That is why i need a pin Enable at the regulator to force it to sleep when the load is switched to the main power supply . The control should come from Non-Overlap Logic Gate .. so i send a signal from main channel to the regulator of aux supply .

I came across LT8607 [http://www.linear.com/product/LT8607]. This step down regulator has ultra low quiescent current of [2.5 micro Amps] . I am not sure if i should worry about its soft start as it takes about 450 micro seconds to reach 12 vdc output regulated voltage .

I dont have LT8607.asy so I simulated LT8609 instead since it has same features , but different load .
I think 750mA of output current is more than enough since the comparators and all channel logic gates have a total quiescent current in micro Amps .
Do you prefer any better than what i already chose from below Linear Technology URL?
http://www.linear.com/parametric/Micropower_Buck_Regulators#!cols_1038,1036,2167,1107,1367,1040,1033,1032!s_1040,0!gtd_!t2_0

Is there any suggestion how to use the Enable Pin in the regulator to force it to sleep or wake up ! ..

 

Thread Starter

globalone

Joined May 27, 2013
22
Thats what i could do for now ..

Prioritized Non Overlap Control Logic :
Over Voltage & Under Voltage




Iam assuimg that both power supplies are ON and Active , the main power supply takes the load and U16 compares UnderVoltage and Over Voltage at AUX SUPPLY then output "1" then U17 assures that 2 inputs are 1 then allow the load of AUX to pass.

A very important question ...
Should i get rid of LT1009 (U3) if i have a step down regulator or i must use them both ?.
 
Top