Problem in Comparator circuit

[BABESH]

comparator 10v

Vipul,
If u suspect transistor, then it should not work even when it is operated with POT...

What if i/p voltage is brought well lower than ref voltage,Is o/p stable?

If that is the case and if u are finding swithing continuously, then i think u din't connect the feedback resistor 330K which is suggested by millwook ckt.

Is u r battery disconnected from load, after relay activation. in that case the terminal volatge may go up for the battery immediate after disconnecting load from battery, so this may shootup i/p voltage > ref voltage which inturn switchOFF relay.. an obvious comparator opeation

--babesh

-- babesh.

 

[Zenouir]

comparator schematic

Vipul,

If I may...I would like to make some observation and recommendation.

1. As I understand you would like to make a circuit that monitors your battery and you want this circuit to cut the battery off when it goes below 10V....right?

2. Your initial problem is, "your circuit is working fine without the transistor, but when transistor is connected everything got messed up".

I think Babesh and Millwood had pretty much answered this problem but I would like to elaborate on this. Referring to the latest diagram you uploaded, and assuming your battery has an initial full charge value of 12V. When the output of the comparator is HIGH (as it will be due to the voltage divider ratio of your resistors R1&R2 which will be around 5.7V@Non_inv input COMPARED to Inv input which is at your zener voltage), this will effectively turn the transistor ON with full force of so much base current.
Let's just say the Voh of the comparator is approximately equal to 12V,
then Ib = (12V - 0.7V)/470ohm
Ib = 24mA : this is far more than the saturation requirement!!!
BC547 only needs Ib=5mA so that Ic=100mA

Let's say we didn't break the transistor with this base current, the next
question will be : Will the transistor turn ON?
You never mentioned what relay you are using, this is critical since the coil
resistance will dictate the level of collector current. Assuming you are using
a 12V relay with a coil resistance of 200ohm and that somehow the
transistor is ON, you will approximately have around 60mA of collector
current!!!...You will be draining the battery and the "Domino Effect" will
happen, Vcc for LM358 goes down, zener voltage will not be maintained,
Voltage divider ratio goes down proportionally, transistor goes ON/OFF,
relay will cease working, then you shout " WHAT THE F**K !! ".

3. Obviously, there are design flaws...most of it, but the idea is there.
>>Resistor ratio is cumbersome : make R1=R2, life will be easier.
>>R3=10Kohm which limits the current going to zener may not be able to
maintain Vz. If Millwood was able to bench 1Kohm satisfactorily for
Vz=4.7V(1N750) then it's fine. But in case you have a problem with Vz,
R3 must be the culprit. It would be nice if you put a trimmer starting at 1K
going down to a few hundred ohms. If you check the datasheet for 1N750,
you will see that Izt is at 20mA.
>>LM358 as comparator is fine but I recommend a real COMPARATOR, say
LM311....and btw, you don't need a pull-up resistor for LM358 since it is not
an open collector.

>>I suggest you use a 7805 (5V regulator) to power up your comparator,
zener, and relay. This way, you will have a monitor circuit that is not point
by point reliant on the changing source which is the battery. Output of
comparator will be at 0V-5V. You will be able to control The Izmin more
accurately and you will be able to maintain the the relay coil voltage.

PLEASE SEE THE ATTACHED SCHEMATIC...
IF YOU HAVE FURTHER QUESTION ON THIS, FEEL FREE TO ASK

 

[vipul]

low voltage indicator on lm311

Hi Zenouir,

I will check the circuit and update you later...

Thank u very much for the reply...

-Vipul.

Added after 2 hours 6 minutes:

Zenouir,

I have checked circuit posted by u.
In this circuit relay does not trip even if the Battery voltage is 11.5 Volt or 6 V...

I have used Lm358 instead of LM311 because i do not have it presently...

So weather u have checked it or missed something in circuit u posted..

 

[Zenouir]

comparator with reference 5.6v

Vipul,

May I know specifically what relay are you using?can you send me the datasheet?

LM358 is fine but we may need to modify the circuit due to current requirement.
By the way, have you added the 7805 regulator circuit?

Added after 1 hours 10 minutes:

I mean, can you upload the datasheet

 

[Audioguru]

lm311 low battery indicator

The zener diode makes a lousy voltage reference. Its voltage drops as the 12V input voltage drops.

It was said before, Get rid of the 1k pullup resistor. It prevents the output of the opamp (it is not a comparator) from going low enough to turn off the transistor.

An LM10 is an opamp plus a very good adjustable voltage reference in an 8-pins case.

 

[millwood]

current indicator comparator circuit

The zener diode makes a lousy voltage reference. Its voltage drops as the 12V input voltage drops.

how precise do you want the reference to be? would a 0.01% reference provide superior results than a 1% / 10% reference here?

think about it.

 

[Audioguru]

cheking lm358

I used a 5.6V zener in series with a 1.7V red LED and a current-limiting resistor as a low battery indicator.

When the battery was 9V then the zener voltage was 5.6V and the LED was bright.
When the battery was 7V the zener voltage was 5.0V and the LED was dimmed.
When the battery was 6V the zener voltage was not measurable because it had no current.

 

[millwood]

relay comparator circuit

I used a 5.6V zener in series with a 1.7V red LED and a current-limiting resistor as a low battery indicator.

When the battery was 9V then the zener voltage was 5.6V and the LED was bright.

so when the battery goes from 13.5v to 10v, how much voltage change do you see on that zener?

When the battery was 7V the zener voltage was 5.0V and the LED was dimmed.
When the battery was 6V the zener voltage was not measurable because it had no current.

not "measurable"? if so, you should throw out your meter and get a new one.

 

[Audioguru]

voltage divider comparator

The zener diode regulated the voltage very poorly at low currents. And the "low currents" were not very low.

But a voltage regulator IC regulates the voltage perfectly from no current to its max allowed current.

Added after 4 minutes:

The zener diode regulated the voltage very poorly at low currents. And the "low currents" were not very low. Datasheets for zener diodes show how poorly they regulate voltage at low currents.

But a voltage regulator IC regulates the voltage perfectly from no current to its max allowed current.

 

[tdy]

the comparator loading is too heavy

 

[millwood]

Added after 4 minutes:

The zener diode regulated the voltage very poorly at low currents. And the "low currents" were not very low. Datasheets for zener diodes show how poorly they regulate voltage at low currents.

OK. how poorly is "very poorly"? when will you run into this problem in this application? does it matter at all for this application? and why does such a "low current" situation even matter here at all?

a typical 6.2v zener for example goes from 6.36v to 6.46v when the current goes from 8ma to 26ma (that's equivalent to a voltage of 9v to 15v applied to the zener and a 330ohm resistor in serial).

for a regulation better than 2%. you don't think that's good enough for this application?

But a voltage regulator IC regulates the voltage perfectly from no current to its max allowed current.

a voltage regulator IC has its own problems as well.

 

[Audioguru]

There is a Lab Power Supply project on another website.
The designer poorly operated its 5.6V voltage reference zener diode at only 1.2mA and did not spec a part number for it.
Many people built the project with zener diodes that are rated at 44mA and their zener diode voltage was too low and their voltage regulation was poor.

I recommended increasing the current in the zener diode to 5.6mA and spec'd a zener diode rated at 5mA which worked much better.

 

[millwood]

but none of that explains why a zener would be a "loosy" reference here. it is a battery monitor that trips at 10v.

so you don't need a precision reference that works from 1v to 50v. nor do you need the reference to be very precise.

all you need is for the zener to provide a reasonably good reference maybe around 9v - 15v. and a zener or even a diode can do that just fine.

now, tempco compensation could be a big problem for a zener / diode in some applications (automotive for example) and that may necessitate a reference chip.

 

[Zenouir]

Vipul,

I simulated the circuit and placed several current/voltage meter everywhere.
R3 and LED1 represents your target circuit.

Battery at 10V relay K1 and LED1 is OFF.
Battery at 9V relay K1 and LED1 is ON.

check it out

 

[Audioguru]

I simulated the circuit

Since it has a voltage regulator then its zener diode can be a resistor.

 

[millwood]

check it out

most zeners provide the best regulation within a specified current range. so you should make sure that the zener works in that current range. you should also need to make sure that if the voltage goes outside of an expected range, the zener isn't damaged. thus people usually use a resistor higher than 10ohm in front of a zener.

also, most opamp s/ comparators need some "room" between its input and rail voltage. 0.1v here is a little bit too tight.

once you have used a 3-terminal regulator, adding a zener doesn't add much value so I would replace it with a resitor / capacitor to further improve noise resistance.

I would also configure the opamp into a schmitt trigger so that the relay doesn't get on / off when the voltage is on the margin.

But overall, you are much closer to a workable product now.

 

[Audioguru]

The simulated circuit uses an LM358 opamp with both of its input voltages at its power supply voltage where the inputs don't work.

With a 5V supply the max allowed common-mode input voltage is only +3.5V. When the input voltage is higher than +3.6V then the PNP darlington input transistors turn off.

Why doesn't the SIM program know that?

The 5.1V zener diode is fed from a 5V supply so it might not work or it might burn out. You can calculate the 5% tolerances.

 

[millwood]

Why doesn't the SIM program know that?

because a sim program is NOT supposed to know that.

a sim program basically takes a model and run it through spice. so if a model doesn't model certain behaviors of a given device, the sim program is NOT supposed to know that.

and very rarely a model models every single behavior of a given device. so it is up to the person running the sim to know if the spice program / model is producing weird behaviors.