Need help with Low voltage cutoff circuit

[DongBeetle]

Hi all
I've been asked to adapt a small logitech sundsystem to run off an 12v lead acid battery, and for that I want to add a low voltage cutoff circuit to protect against over discharge of the batt.
I've come up with this circuit, and I would like to know if it looks ok?
Also would it better the stability if the LM393 is supplied with a constant 5V from the AN7805, would that even work?!?
Regards.

 

[alec_t]

It looks ok, but could you perhaps simplify it by just using a zener diode instead of the 7805 to give the reference voltage for the comparator (I don't believe the cut-off voltage is critical)?
The comparator performance won't be greatly affected by small changes in its supply voltage, so the battery voltage should be fine as the supply. Besides, the FET benefits from having >5V on its gate for reliable turn-on.

 

[DongBeetle]

Thanks for the fast reply and good advise.
I might look into using a zener insted.
Regards.

 

[ericgibbs]

hi,
I would agree with alec ref the zener, the 7805 is an over kill.

The downside to using the 5V to power the LM393 is that the +Vin, Vset input would be higher than the permissible Vinp to the LM393 powered by +5V.

E.

 

[DongBeetle]

So I've made the changes you suggested, and it seems to work on the breadboard, here is what I've got now.
Regards.

 

[DongBeetle]

One more thing
Sorry to be such a novice, but can you please explain how to calculate the hysteresis, this image is from the datasheet, and to me it looks like the formula for calculating hysteresis!?
Does it apply to my circuit, and how?
Regards.

 

[alec_t]

Yes, those are the hysteresis formulae which apply to your circuit. Th1 and Th2 are the upper and lower switching thresholds respectively.
BTW, I would reduce the value of R5 to, say, 4k7 or 3k3 to allow a bit more current through the zener for a more sharply defined zener voltage.

 

[ericgibbs]

Yes, those are the hysteresis formulae which apply to your circuit. Th1 and Th2 are the upper and lower switching thresholds respectively.
BTW, I would reduce the value of R5 to, say, 4k7 or 3k3 to allow a bit more current through the zener for a more sharply defined zener voltage.

hi,
What are you setting the battery switch OFF voltage too.??

I would recommend not less than 11V, certainly no lower than 10.8V.

Your divider chain with the potentiometer needs recalculating.

E.

EDIT:
Run a LTSpice simulation for you, can you follow it OK.?

these two images show the Set pot at either end of its range.

 

[ericgibbs]

hi Dong,

This image should give you a better idea.

It shows the Vbattery falling from 12V down to 10V and rising back up to 12V

The switch OFF point for the falling voltage is around 10.8V and the switch back ON is around 11.4V

Note the chain resistor values and the hysteresis value.
IMO a 10K pot is too coarse for a fine adjustment.[ its set for 35% rotation for the simulation]

E.

 

[DongBeetle]

Thanks alot Eric!
I had noticed a problem, and I was messing about with different resistors when I saw your reply, I have set it up like in your simulation and it seems to be working just fine.
If I was to change the 10K pot for a 5 or 1K, how would it affect the other resistor values?
Regards.

 

[MikeMl]

Hi Mr Beetle,

Here is a circuit I have used. Note the parts count. Note that it needs no relay to switch up to several Amps. It turns back on automatically when the battery is recharged. It uses an LM/TL431 shunt regulator, which is a part about as useful as a 555. Like Eric's, the circuit has a bit of hysteresis to prevent cycling as the battery reaches the trip point.

The time scale used in the simulation is hugely shortened for convenience. Plotted is V(bat+) red and V(load) green. Note turn-off voltage, and turn-on voltage.

 

[ericgibbs]

Thanks alot Eric!
I had noticed a problem, and I was messing about with different resistors when I saw your reply, I have set it up like in your simulation and it seems to be working just fine.
If I was to change the 10K pot for a 5 or 1K, how would it affect the other resistor values?
Regards.

hi Dong,
This uses a 5Kpot and two 8.2K resistors.

The plots show the 5K pot being set for 10% thru 90% in 10% steps.

If I was building from new, I would use Mikes circuit.

E.

 

[DongBeetle]

Thank you Mike and Eric
So I tried building Mike's circuit, but I could not make it work.. maybe the p-fet I used (NEC J449) doesn't work for this..
I'm gonna stick with the other circuit matching Eric's last changes, it seems to be working..
I have salvaged the parts so it won't cost me anything, my biggest problem is my variable power supply, it's about 50 years old and build by my grandad.. it's a bit unstable
Regards.

 

[DongBeetle]

A little update..
I found out that with the 220K hysteresis resistor the system had a tendency to resonate at the tippoint, so I changed it to 100K wich has made it stop, I'm not sure what the turn on voltage is now, but it seems to be tipping at around 11,8-12V?!

 

[MikeMl]

...
So I tried building Mike's circuit, but I could not make it work.. maybe the p-fet I used (NEC J449) doesn't work for this...

The J449 is the problem. It has a threshold voltage of ~-6V, and my circuit only applies a Vgs of ~-7V. Also, the Ron is >0.55Ω when the gate is driven to -10V.

The PFET I used has a threshold of -2V, and is well turned on by the time Vgs gets to -4V. It has an on resistance of 0.013Ω with a Vgs of -4.5V @ an Id of 10A.

 

[ericgibbs]

A little update..
I found out that with the 220K hysteresis resistor the system had a tendency to resonate at the tippoint, so I changed it to 100K wich has made it stop, I'm not sure what the turn on voltage is now, but it seems to be tipping at around 11,8-12V?!

hi Dong,

If Grandpa's power supply is being used as a 'test battery' it is possible that the power supply may have a high 'ripple' voltage level, which when superimposed on the DC could cause the relay to chatter.

By increasing the hysteresis, you could get to a condition where the cut OFF voltage occurs at say 10.8V and the ON voltage is > 12V, so the circuit will not switch back ON.

E.

 

[DongBeetle]

Hi Eric
I don't know the prober name for my "power supply" it's a small box I put between the battery and circuit, basically it's a big porcelain variable resistor, varying the voltage output.
About the hysteresis resistor, at the moment I use 100K, so I've decreased the value.. and the battery will most likely be fully recharged after each time it's been cut off, but do you see any problems there?

And mike, I thought as much about the J449, but unfortunately it's the only p-fet I have at hand..
Regards.

 

[WTP Pepper]

Hi all
I've been asked to adapt a small logitech sundsystem to run off an 12v lead acid battery, and for that I want to add a low voltage cutoff circuit to protect against over discharge of the batt.
I've come up with this circuit, and I would like to know if it looks ok?
Also would it better the stability if the LM393 is supplied with a constant 5V from the AN7805, would that even work?!?
Regards.
View attachment 74263

My first concern is that once the battery is cut-off, you still have several components that can further discharge the battery. The 7805 being a major culprit with it's rather high quiescent current. There are better like the LM2985 5V part.

Take a look at the MAX8212. It's designed for this with an inbuilt reference. I have used this without issues before for this very application.

 

[DongBeetle]

well Pepper, the 7805 has been replaced with one resistor and zener diode, and my multimeter reads 3,2mA after cutoff wich I think is pretty low, I haven't done any calculations on it, but it should take a while to discharge the battery (28Ah) to a dangerous level.. right!?!..

 

[MikeMl]

My first concern is that once the battery is cut-off, you still have several components that can further discharge the battery. ...

In Mike's circuit, the quiescent current after the device switches off is 313uA. Since the input bias current into the REF input of TL431 is only 2uA, the voltage sampling network consisting of R5, U1, and R2 could be redesigned to make the quiescent current of less than 100uA. R6 would have to be adjusted, too.