Lead acid battery monitor with 555

[antenna]

A friend of mine lives off grid and often lets his lead acid batteries get too low.  I want to build him a simple battery level alarm with the ability to select the alarm and reset voltage (between 0 and 50% state of charge to turn on alarm and 50-100% state of charge to turn off).  I also wanted the ability to manually silence the alarm.  The 555 seemed like the easiest way to do it.  Before I build it, I wanted to check here to make sure I made no critical mistakes as the circuit I came up with seems "too simple".   I might add a beeper so the alarm isn't running continuously, maybe a chirp every minute, not sure yet.
Thanks!

[antenna]

I know I need to add a capacitor to the + rail after the 100Ω isolation resistor, forgot to draw that.  Should help with false triggering from large momentary current draws too.

[Kim Christensen]

Your circuit needs a resistor between the base of the transistor and pin3 of the 555. (I didn't analyze the rest of it.)
You'd be better off using a comparator set up with a bit of hysteresis. Much simpler than the 555. Plus, a comparator like the LM393 works up to 30Vdc.
Then you'd need only one zener diode (or none if the comparator has a built in reference) and have better control over the trigger and reset voltage levels.

[Spar59]

Your definition of too low a voltage and that which can damage a lead-acid battery are way apart.

Your suggestion of allowing an alarm at 0 to 50% doesn't really bode well unless you set it to 50%.

A lead acid battery if being cycled on a regular basis can only be discharged to 50% if you don't want to seriously reduce its life, yes that does mean you can only use half its rated capacity!

This is the case even with "deep cycle" batteries - they can tolerate deep discharges followed by immediate re-charge such as taking a charged one to a tent for a camping weekend but can't do deep discharge on a daily basis like would be needed on a solar system.

I suggest you first look at the battery specification sheets.

For daily charge/discharge cycles you really want to go with LiFeP04 or similar - a large up front cost but much more usable capacity and much longer life - I learned the hard way and persisted with lead-acid for far too long.

[antenna]

Your definition of too low a voltage and that which can damage a lead-acid battery are way apart.

Your suggestion of allowing an alarm at 0 to 50% doesn't really bode well unless you set it to 50%.

A lead acid battery if being cycled on a regular basis can only be discharged to 50% if you don't want to seriously reduce its life, yes that does mean you can only use half its rated capacity!

This is the case even with "deep cycle" batteries - they can tolerate deep discharges followed by immediate re-charge such as taking a charged one to a tent for a camping weekend but can't do deep discharge on a daily basis like would be needed on a solar system.

I suggest you first look at the battery specification sheets.

For daily charge/discharge cycles you really want to go with LiFeP04 or similar - a large up front cost but much more usable capacity and much longer life - I learned the hard way and persisted with lead-acid for far too long.

I ended up taking the above advice and going with a LM393.  The guy this is for got confused when I explained the set points would be adjustable, so I made them fixed at 11.95v turn on and 12.50 turn off.  That's about 40% SoC, which he routinely drops below.  When it is not sunny, he needs to go start a generator to charge his battery, so I doubt he will be running outside and starting it without it being down to at least 40% (regardless of my countless warnings about sulfation being a serious issue when it sits that low).  The problem with making the setpoint higher is that the voltage may sag and false trigger when he turns on his inverter or keys his radio amplifier, had I made it trigger at 50%, it would probably activate all the time.  There is a capacitor that slows the response to hopefully eliminate that problem, but it will only prevent alarms on very brief drops in voltage.

Anyhow, it is going to stay how I built it as it is now hot-glued into an altoids box.  Thanks everyone!!!

[Benta]

What spar59 said.
If the batteries in question are normal automotive types, you should never go below 70% on a regular basis.
Deep-discharge lead-acid types exist, but they're expensive.

[Zero999]

I don't know much about the optimum battery voltages, but I would opt for a comparator circuit, rather than the 555.

Here's an example using the LM339. U1a is used as a bistable, whilst the other sections are used to sense the high and low thresholds.

A zener & silicon diode for temperature compensation are used as a reference, but I would consider the TL431, or even an L7805L.

[antenna]

I don't know much about the optimum battery voltages, but I would opt for a comparator circuit, rather than the 555.

Here's an example using the LM339. U1a is used as a bistable, whilst the other sections are used to sense the high and low thresholds.

A zener & silicon diode for temperature compensation are used as a reference, but I would consider the TL431, or even an L7805L.

Thank you!  I like that circuit!  The only thing I would change is the reference divider.  As is, the bottom pot would need to be turned to max for 60%SoC 12.18v and the top pot set at just barely above minimum to be set at 90%SoC 12.53v. It would be very difficult to adjust.   

With the 56k/47k divider for voltage sensing, 90%SoC for lead acid of 12.53v would be 5.718v at the divider. Likewise, 60%SoC, or 12.18v would be 5.556v at the divider.  These are the setpoint voltages.  To make the 1k pots less sensitive and easier to adjust, a current of 0.16mA could be chosen so that each pot has a voltage across it equal to the difference in the set point voltages (pots will have equal range centered on initial desired values and easy to fine tune). If the diode regulator is at 6.85v, the total resistance needed for 0.16mA is 42.8kΩ.  At the low set point, the wiper of the low pot needs to be at 5.556v/.16mA=34.7kΩ and the wiper for the high pot needs to be at 5.718v/.16mA=35.7kΩ, so tucking those two 1k pots between a 34.2kΩ tail resistor (where R13 is) and a 6.6kΩ resistor above it (between R11 and the reference node) will put the pots right where they need to be.

The only feature that schematic is missing is an audible alert that can be silenced independent of the alert LED function.

[Spar59]

At least you warned him about battery abuse, so when despite your best efforts to provide him with an alarm he finds his batteries expire more rapidly than he expected you have a clean conscience.

But there is a bit of lattitude as the SOC voltages are normally stated at rest, on load the voltage drops slightly so your alarm will activate at a bit over 40% SOC if they are on load, when the load drops off they will recover back to a slightly higher voltage.

[Zero999]

I don't know much about the optimum battery voltages, but I would opt for a comparator circuit, rather than the 555.

Here's an example using the LM339. U1a is used as a bistable, whilst the other sections are used to sense the high and low thresholds.

A zener & silicon diode for temperature compensation are used as a reference, but I would consider the TL431, or even an L7805L.

Thank you!  I like that circuit!  The only thing I would change is the reference divider.  As is, the bottom pot would need to be turned to max for 60%SoC 12.18v and the top pot set at just barely above minimum to be set at 90%SoC 12.53v. It would be very difficult to adjust.   

With the 56k/47k divider for voltage sensing, 90%SoC for lead acid of 12.53v would be 5.718v at the divider. Likewise, 60%SoC, or 12.18v would be 5.556v at the divider.  These are the setpoint voltages.  To make the 1k pots less sensitive and easier to adjust, a current of 0.16mA could be chosen so that each pot has a voltage across it equal to the difference in the set point voltages (pots will have equal range centered on initial desired values and easy to fine tune). If the diode regulator is at 6.85v, the total resistance needed for 0.16mA is 42.8kΩ.  At the low set point, the wiper of the low pot needs to be at 5.556v/.16mA=34.7kΩ and the wiper for the high pot needs to be at 5.718v/.16mA=35.7kΩ, so tucking those two 1k pots between a 34.2kΩ tail resistor (where R13 is) and a 6.6kΩ resistor above it (between R11 and the reference node) will put the pots right where they need to be.

Please do change the design. It was just a demonstration of how to make a comparator circuit with independantly adjustable high and low thresholds, rather than a real life design.

The only feature that schematic is missing is an audible alert that can be silenced independent of the alert LED function.

The spare comparator in the LM339 can be used as another monostable, similar to how U1a is configured. I don't have time at the moment, but will post a schematic later.

EDIT:
Here's the schematic. I got rid of the PNP on U1A because it's only driving an LED. The alarm now has a PNP output for the alarm.

It's a bit complicated. Why are upper and lower thresholds necessary?

Why not just have a latch and reset?

[floobydust]

I think the operation of the monitor needs more thought unless I've missed how it works.

You don't want fast (outside the window) voltage detection at all. I use a 1-2 minute average in MCU code so transient loads or inrush currents, or PWM charging noise etc. do not false-trigger the monitor.

We don't know what the off-grid system is powering, I was thinking of a toilet fill pump, LED lighting, who knows as far as transient or night-time loads.

Second is after the alarm goes off, somebody acknowledges it by hitting a button - I think it could nuisance re-trigger. With no-load the battery voltage climbs up, and then falls quickly with load. Nobody likes their smoke alarm low-batt chirping at night lol. The 0.35V difference is not much and might be impractical?
Maybe a red blinking LED whenever the battery voltage is high/low (outside a protective window but I think OP only cares about LV) which trips the alarm, once acknowledged stays quiet, like a hush button, forever. It assumes the human remembers it lol.

[antenna]

Regarding that latest schematic, I think a capacitor at the voltage sense divider could have larger resistors and a capacitor to ground slowing the response time.  In the one I already hot glued together, it has a 100Ω resistor and a large filter cap to sort of do the same, but nowhere near as good.  If the one I built don't behave correctly, then he can toss it and I'll buy another box of altoids.

[Zero999]

Regarding that latest schematic, I think a capacitor at the voltage sense divider could have larger resistors and a capacitor to ground slowing the response time.  In the one I already hot glued together, it has a 100Ω resistor and a large filter cap to sort of do the same, but nowhere near as good.  If the one I built don't behave correctly, then he can toss it and I'll buy another box of altoids.

Why do you want a slow response time?

There is no capacitor at the sense divider, in any of the schematics I've posted.

The problem with adding a delay to the voltage sense divider is it will momentarily activate, triggering the alarm, every time the power is applied. C2 ensures, when the power is first connected, the reference voltage is much lower, than the sense divider.

U1d is simply a bistable. When U1a's output goes low, C3 AC couples the negative edge to U1d's +input via D1, which ensures it only responds to a falling edge. U1d's output will go low, turning on Q1. C4 performs a similar function to C2, in that it ensures U1d's output is high when the power is first applied.

To add a delay, C2 will needed to be increased and a capacitor added in parallel with R9. The values will need to be carefully chosen to ensure the sense voltage is always higher than the reference, during power on.

[iMo]

It calls for a small 50cents 8pin MCU
Two pots (2pins ADC inputs), perhaps 5 resistors, an NPN for the buzzer (1pin output), a blinking LED showing status (1pin output), the push-button (1pin input), a 5V regulator (ie 78L05 or similar) for the MCU's power (2pins Vcc Gnd), perhaps 5 capacitors (3x100n - at the two ADC inputs and at the push-button input, 100u @7805 input and 10u at the MCU's Vcc).. Total 7pins.. And the business logic as large as it fits in its flash

[Zero999]

It calls for a small 50cents 8pin MCU
Two pots (2pins ADC inputs), perhaps 5 resistors, an NPN for the buzzer (1pin output), a blinking LED showing status (1pin output), the push-button (1pin input), a 5V regulator (ie 78L05 or similar) for the MCU's power (2pins Vcc Gnd), perhaps 5 capacitors (3x100n - at the two ADC inputs and at the push-button input, 100u @7805 input and 10u at the MCU's Vcc).. Total 7pins.. And the business logic as large as it fits in its flash

A microcontroller will work.

I would recommend putting U1 & U6 in series, since they're there to set the upper and lower thresholds.

To simplify it further, I would be tempted to get rid of the driver transistor and use a piezo disc traducer driven from a square wave between two IO pins.

[iMo]

I would leave the last free pin for the "CALIBRATE" function
A pin header with 10k pullup for a short against GND.
Install the short and power up the MCU -> Led blinks with messages and you set the potentiometers for given percentages (while changing the input voltage). Confirmation of each setting with pushing the button.
Easy 

[antenna]

Regarding that latest schematic, I think a capacitor at the voltage sense divider could have larger resistors and a capacitor to ground slowing the response time.  In the one I already hot glued together, it has a 100Ω resistor and a large filter cap to sort of do the same, but nowhere near as good.  If the one I built don't behave correctly, then he can toss it and I'll buy another box of altoids.

Why do you want a slow response time?

There is no capacitor at the sense divider, in any of the schematics I've posted.

The problem with adding a delay to the voltage sense divider is it will momentarily activate, triggering the alarm, every time the power is applied. C2 ensures, when the power is first connected, the reference voltage is much lower, than the sense divider.

U1d is simply a bistable. When U1a's output goes low, C3 AC couples the negative edge to U1d's +input via D1, which ensures it only responds to a falling edge. U1d's output will go low, turning on Q1. C4 performs a similar function to C2, in that it ensures U1d's output is high when the power is first applied.

To add a delay, C2 will needed to be increased and a capacitor added in parallel with R9. The values will need to be carefully chosen to ensure the sense voltage is always higher than the reference, during power on.

On the other version I built already, I dealt with it alarming on power up by adding a capacitor across the zener reference so the reference takes longer to stabilize than the voltage at the sense pin, so using a capacitor on the sense pin could be used here, as long as the sense voltage takes less time than the reference voltage to charge up.

I appreciate all the suggestions for the next version I, or someone else, may some day build.  Thanks!!!

[_EDMAR_]

I have about a dozen 6v empty cells (approx 24oz) with no Xtra specs. I'm interested in your project, may I have a copy of your schematics?

[Zero999]

You can use any of the schematics I've posted, for whatever reason, including in a commercial product. I can't speak for the original poster.

Thinking about this again.

I've just done a quick search about lead acid batteries and the voltages.

A 12V battery shouldn't be discharged below 10.5V, which would be 5.25V for a 6V battery.

I would recommend two under voltage circuits.

One to cut off the power when the voltage drops below 10.5V. This should be fixed and not manually resettable, with the only way to reset it is to charge the battery above 12V.

And another to sound an alarm, when the voltage drops below a certain point, which can be adjustable and manually resettable.

[antenna]

I dont have my original circuit.  It wasn't adjustable.  Used two comparators with hysteresis to get the right set points, mosfets sampling the comparator output for LED alert and to drive a 555 in flipflop mode to serve as a buzzer switch that can be reset/silenced.  That schematic disappeared from my desk, sorry.  The one posted is much better than the one I had used.