Low battery level circuit

[kuksul08]

Using a 6AA battery pack (7.2V nominal, 8.4V fresh off charger, <6.0V discharged), does this circuit look like it will work?

All I am after is for an LED to light up constantly when voltage drops out of its normal operating level, so at about 6V. I have found several online, but this one seems the simplest and most applicable to these low voltages. If not, do you have a better suggestion? If so, any suggestions for a particular TL431 component? Digikey has a lot of different tolerances, cathode currents, adjustable/programmable, etc.

Thank you!

http://homepages.paradise.net.nz/bhabbott/lvw.html

 

[kuksul08]

bump.

 

[xul]

The problem is, with a 2v LED and whatever circuit configuration you decide on, you will have no LED indication at above 6v and below 2v unless the circuit is powered by other than the battery itself, so you will have false positive [the below 2v event] readings sometimes.

 

[kuksul08]

I see. So with say, a 6AA battery pack, it should never drop below 2V even if the cells drop to .5V/ea or whatever under load.

 

[xul]

Some other thoughts: I'd make R3 variable so you can tweak the TL431 turn-off voltage to high precision. Also, the LED current might be a max of 4v/2200 ohms = 2 mA so they won't be very bright.


Can you post the datasheet for the 431?

 

[kuksul08]

Some other thoughts: I'd make R3 variable so you can tweak the TL431 turn-off voltage to high precision. Also, the LED current might be a max of 4v/2200 ohms = 2 mA so they won't be very bright.


Can you post the datasheet for the 431?

I would definitely use a pot to tune in the exact resistance, and then probably switch over to a high quality resistor, +/- 1% once I knew the correct value. I'll probably be using a red 5mm LED, so 15mA would be the target.

I could post the datasheet. I'm not sure exactly which model to use though.
Here's one: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=296-10183-1-ND , http://www.ti.com/lit/ds/symlink/tl431a.pdf

What relevance is the cathode current?

 

[xul]

1
I'll probably be using a red 5mm LED, so 15mA would be the target.

I could post the datasheet. I'm not sure exactly which model to use though.
Here's one: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=296-10183-1-ND , http://www.ti.com/lit/ds/symlink/tl431a.pdf

2
What relevance is the cathode current?

1
To get 15 mA @ ~4v across the two LEDs with this circuit R1 should be 130 ohms instead of 2200 ohms.

2
According to the datasheet it is the recommended current through the device, cathode to anode, and is 1 to 100 mA.

The circuit shown above pulls slightly more than a max of 2 mA; is this close enough to an open circuit for your application?

How can this circuit work when the battery is above 6v? In that case there'll be 6v across the two LEDs which will probably overdrive them. I must still have jet lag from my recent trip east.

 

[kuksul08]

1. I see, so the LED current once it shorts is determined by R1? I wonder why the guy who made it up used such a large R value.
2. 2mA is totally fine. The purpose is a low battery LED for a bike light which will be drawing upwards of 2800mA. I really don't know anything more than the description included in the link above, hence my post here.

 

[xul]

You could do this function with a single-supply micropower comparator/op-amp with some hysteresis put in for this slowly moving voltage, but there may be another way.


Ideally a relay that pulls in and drops out at 6.0 v would work. This would have a rated coil voltage of 12 v. It'd be SPDT and the LED would be powered from the NC contact.


In the real world, relays pull in at ~2/3 of rated voltage and drop out at ~1/3. Aiming for (7.2-6)/2 = 6.6 v with a rated coil voltage of 13.2 v this relay would pull in at 8.8 v and drop out at 4.4 v, a range that is too wide for you.


Let's say you can find a 3 v relay that pulls in at 2.0 v and drops out at 1.0 v. Put the relay coil in series with your programmable Zener set to 6.6 - 1.5 = 5.1 v. Then the circuit would pull in at 7.1 v and drop out at 5.1 v and this band can be tweaked up or down by adjusting the Zener.
If this isn't workable we go the comparator route.

 

[Norm]

I've used this successfully in the past, LED Battery Voltage Indicator Kit
Norm

 

[kuksul08]

xul- Now I'm thoroughly confused. My electronics knowledge is pretty limited. I'm trying to decipher everything you wrote, but it looks like good info.

Norm- Thanks for the link. I would like to DIY the circuit for fun and for cost reasons. That one you linked looks pretty simple though, I'm sure I could figure out how to make one.

 

[xul]

I've used this successfully in the past, LED Battery Voltage Indicator Kit
Norm

If you can reverse engineer the schematic of this other voltage ranges can be easily put in. The main thing is, what are the numbers on that chip?

 

[Norm]

I've used this successfully in the past, LED Battery Voltage Indicator Kit
Norm

This seems to be the schematic for it.

 

[xul]

This seems to be the schematic for it.

So they're using the LM358 opamp as a comparator. The 10k and 4.7M gives some hysteresis and the 10 uF prevents the whole thing from oscillating. Ideally this cap should be tantalum rather than aluminum electrolytic but I guess it works anyway. Usually they parallel this cap with a 0.1 ceramic for the high frequency response.

OK, this circuit can be modified to cover other voltages and the advantage is that we know it works in production.
The simpler circuits do not have a high likelihood of working as designed. That's the price paid for simplicity.

 

[kuksul08]

So they're using the LM358 opamp as a comparator. The 10k and 4.7M gives some hysteresis and the 10 uF prevents the whole thing from oscillating. Ideally this cap should be tantalum rather than aluminum electrolytic but I guess it works anyway. Usually they parallel this cap with a 0.1 ceramic cap for the high frequency response.

OK, this circuit can be modified to cover other voltages and the advantage is that we know it works in production.
The simpler circuits do not have a high likelihood of working as designed. That's the price paid for simplicity.

I think I understand. So a comparator simply compares the reference voltage to the source voltage based on R values and will turn on/off accordingly?

I know what hysteresis is from AC motors, where there is inherent loss in the iron core. Is it the same concept.. a type of nonlinear loss in the middle of its operation so that it won't toggle repeatedly (blinking LED)?

The circuit looks easy enough. How could I modify it to only use one LED (red illuminates only when voltage drops below threshold)? I assume I could simply remove one of the opamps, but I'm not sure what purpose the resistors and capacitors serve or which ones could be removed. I enjoy building these circuits because I learn how general components work each time. My last circuit used a decade counter and 555 timer - sweet little chip.

 

[xul]

In this case hysteresis snaps the opamp output from high to low or vice versa without dithering regardless of how slowly the input voltage is changing.
A Schmitt Trigger uses this.

When the voltage drops below 6v, pin 1 goes high and lights the LED marked 'R' by putting a max of ~4 mA through it, so just put a red LED in with the same polarity.
When the LED is extinguished the battery is low or the circuit is unpowered or the battery is below around 2v, but with a red/green LED you mostly know the circuit is working.

There are two opamps in the same package.

 

[kuksul08]

In this case hysteresis snaps the opamp output from high to low or vice versa without dithering regardless of how slowly the input voltage is changing.
A Schmitt Trigger uses this.

When the voltage drops below 6v, pin 1 goes high and lights the LED marked 'R' by putting a max of ~4 mA through it, so just put a red LED in with the same polarity.
When the LED is extinguished the battery is low or the circuit is unpowered or the battery is below around 2v, but with a red/green LED you mostly know the circuit is working.

There are two opamps in the same package.

Thanks again :)

So let me get this straight to fully understand this-

1K resistor on pin 1: current limiting for LED(s). 6V-2V / 1000Ohm = 4mA
R/G LED: I could use a red and a green LED facing opposite directions instead of a special Red/Green
10K and 4.7K resistors on pins 1/3/6: creates hysteresis (somehow)
10uF cap: prevents oscillating in the circuit by dampening out source voltage fluctuations?
5.6K resistor: ?
10K resistor and 10K pot on same branch: Does this set the voltage at which it switches over to Red, hence why it's adjustable?
5.6K resistor: ?
LM336Z-2.5: ?

Any idea what power ratings these resistors should be?

 

[xul]

1.
1K resistor on pin 1: current limiting for LED(s). 6V-2V / 1000Ohm = 4mA


Yes. With 8v in and a 2v drop on the LED and assuming the opamp output swings 'rail to rail' the power dissipated is [8-2]^2/1000 = 36 mW. This is small enough that you should pick the cheapest resistor that will fit in the space allocated.


2.
R/G LED: I could use a red and a green LED facing opposite directions instead of a special Red/Green


Yes.


3.
10K and 4.7K resistors on pins 1/3/6: creates hysteresis (somehow)


With the 10k pot wiper at the top and 6 v in, pin 1 will be at ~0v and pin 3 will have 2.5 - {[10k][2.5v - 0v]/[~4.7M]} = 2.495v so the hysteresis in this direction is about 5 mV.
With 4v in, pin 1 will be at ~4v and pin 3 will have 2.5 + {[10k][4v - 2.5v]/[~4.7M]} = 2.503v, so the peak to peak hysteresis is about 8 mV.
Evidently the designer of this circuit is not expecting much electrical noise.


4.
10uF cap: prevents oscillating in the circuit by dampening out source voltage fluctuations?


This circuit draws different current depending on what it's doing. With non-zero connecting wires this translates to voltages which this high gain amplifier may interpret as valid input signals, and depending on the phasing this circuit may oscillate.
Since the voltage across a capacitor cannot change instantly this cap provides a point of zero AC impedance. Single Point Ground is a related concept.
Not all circuits may need this cap but it is good practice to put it in as insurance against undesirable operation.
If this circuit were switching a tungsten load I'd use a cap and a single point ground.


5.
5.6K resistor: ?


Establishes a reasonable current through the 2.5v voltage reference.
With no current into any opamp input and 8v in this current is [8-2.5]/5600 = ~1 mA.


6.
10K resistor and 10K pot on same branch: Does this set the voltage at which it switches over to Red, hence why it's adjustable?


Yes.
With the wiper at the top of the lower 10k, and knowing that the wiper voltage is always equal to the reference voltage this circuit would fire at 5.0 v.


7.
LM336Z-2.5: ?


Is or was L[Linear]M[Monolithic] IC voltage reference by National Semiconductor.
They have, or used to have, application notes for probably every one of their products and these notes took up a whole book. These notes are probably on the web.


8.
Any idea what power ratings these resistors should be?
See above.

 

[kuksul08]

Thank you so much, I really appreciate it :wave:

 

[xul]

My pleasure. It was a trip down memory lane for me.