How to disconnect device from 18650 battery when voltage drops below 3V?

[dusan]

I've made small audio amp (1 npn common emitor preamp, lm386 for output, 0.5W speaker). I'm powering it from 9V battery but I would like to redesign it to use single 18650 battery (3.7V). I've read they should not go below 3V so how do I disconnect the device when battery drops below 3V?

Should I just compare battery voltage to reference and only when it is above 3V connect battery via mosfet to my device? But then the comparator will still be connected to battery, no?. I just don't know how to do this I've never done this and I don't know what to search for.

I've googled but anything I found is consumer advice not actual engineering advice.

I have sense that this should be something trivial, obvious, something when I see it I think "of course", but you know wheel is obvious in hindsight but very hard before. Is there a book with 1000 of problems and solutions like these?

I haven't done anything EE related in 20  years so consider me beginner-hobbyist.

[james_s]

Probably the easiest thing to do is just get a single cell protection PCB, this will provide protection against overcharging as well. They are cheap and readily available.

[Peabody]

Or just use a protected 18650.

[Skashkash]

There are single cell protection circuits.
 Usually designed around a DW01 and two p-fets.
https://cdn.sparkfun.com/assets/learn_tutorials/2/5/1/DW01-P_DataSheet_V10.pdf

 If this is a "one off" there are some modules that combine the charger IC and protection circuit in one.
  https://www.amazon.com/Makerfocus-Charging-Lithium-Battery-Protection/dp/B071RG4YWM

 I use these modules in simpler projects.  They just combine the popular cheap charger chip with an equally cheap protection circuit.

[dusan]

Thanks, I think I'll use that module with combined charger and protection

[Calvin]

Hi,

I´m also searching for quite a while now for such modules or subassemlies ... but from 5V USB to 2 or more series connected 18650 Cells. Charger with protections (OV, OC, UV and load disconnect)
Anyone know where to find something?

regards
Calvin

[james_s]

That's going to be more complex since it will require a boost converter. 5V from USB is enough to charge a single cell with a simple linear controller IC but not enough for 2 cells. What you want may exist but I have not yet encountered a simple little integrated module.

[Calvin]

Hi,

Yes, exactly .... all I found so far are single-cell devices.
The MP2672A of PowerMonolithics would be fine, a 5V-USB Boost charger plus BMS + several protection features, but it misses on the undervoltage battery protection. 
It discharges the batteries into the load infinitely ... such a complex device and they ´forgot´ to include two more PMOS ... Now, how silly is that??
Maybe it might be possible though to use two DW01, one for each battery cell?

regards
Calvin

[Terry Bites]

?? https://www.amazon.com/Li-ion-Lithium-Battery-Protection-Circuit/dp/B07W6L7F3X?th=1

[Calvin]

Hi,

yes, there seem to exist millions of these devices on amazon -which I btw. rather refuse to buy from.
They all seem to have a fixed UV-treshold (overdischarge treshold) of <3.0V ... which is imho too low to be of practical use.
3.2V to 3.5V would be healthier for the battery.
Has anyone found devices with a higher overdischarge treshold?

egards
Calvin

[james_s]

You can buy the exact same devices direct from China too, ebay, aliexpress, dealextream, etc.

3V is not unreasonable as a protection disconnect, ideally you shouldn't really run the battery down that far if you can avoid it, but it will prevent damage. IIRC the cutoff voltage is usually designed into the IC used.

[tanveerriaz]

use a old phone bat.  protection circuits. you get it easy and free

[tooki]

use a old phone bat.  protection circuits. you get it easy and free

I wouldn’t. You don’t know what it’s configured for, and your own battery may need different parameters.

[Peabody]

Hi,

yes, there seem to exist millions of these devices on amazon -which I btw. rather refuse to buy from.
They all seem to have a fixed UV-treshold (overdischarge treshold) of <3.0V ... which is imho too low to be of practical use.
3.2V to 3.5V would be healthier for the battery.
Has anyone found devices with a higher overdischarge treshold?

I would just point out that hundreds of millions of protected lithium batteries use the DW01 protection chip, which shuts off at 2.5V, and turns back on at 3V.  The 2.5V is usually under load, and then it recovers to 3V after protection is triggered.  So it wasn't really 2.5V with no load.  Many batteries will cycle until they are no longer quite able to recover back to 3V.  But they will be close to 3V at no load.  So it's not quite as bad as it seems.

[Gyro]

This is going to sound horribly unscientific...

A couple of years back, I obtained a small quantity of standard DW01 LiPo cell protection boards from a fellow member who was disposing of a large number of them for the cost of postage. In the course of experimenting with them, I discovered that increasing the series resistor on the DW01 VCC pin (R1 in the application circuit) from the normal 100R up to around 330k caused the low voltage cut-off point to increase to around 3.1V -3.2V. I did experiment with values up to 1M, but found 220k - 470k to be the useful limits (330k being optimal) beyond which the threshold stopped changing and ultimately became unstable....

...Well I did say it was unscientific! It is completely outside the specified operating conditions of the DW01 IC, but repeatable among my samples. I have used these boards on a couple of portable one-off projects for a cheap and simple cut-off. Standard charging cell voltage resets them to the normal on state.

Two important constraints:

1. The circuit CANNOT be used for charge shut-off protection. This is extremely important - the upper voltage thresholds are similarly raised! The circuit must be in the discharge path only. You should either use a cell with its own protection board or use another standard DW01 based circuit in the charge path for over-charge protection (in addition to the normal charger).

2. The switching FETs are in the -ve rail to the battery. This requires the load circuit ground to be separate from the charger ground. Not a safety issue itself, a common ground will simply bypass the FETs and it will fall back on the standard low voltage cutoff of the protected cell. Separating the grounds is not normally a problem for portable projects where charging and use are separate activities.

Your mileage may very significantly vary - I would never have tried it if I hadn't had these boards hanging around for free - I'm not proud of things that I can't justify from the datasheet parameters.

[Jwillis]

You just want a simple cut off switch for a 3.7V LiIon ?
Low Voltage Cutoff Switch
These will do 1S, 2S or 3S .Cut off at 3V
Might be able to find them cheaper some where else.

[BlackICE]

You just want a simple cut off switch for a 3.7V LiIon ?
Low Voltage Cutoff Switch
These will do 1S, 2S or 3S .Cut off at 3V
Might be able to find them cheaper some where else.

Quite costly, but the current capacity is very high. Most of my uses would be at much lower currents and these parts have too high of a parasitic loss. I'm power small micro controllers with batteries, it is probably possible to use the controllers ADC to monitor the batteries voltage and to turn itself off when the voltage is too low.

[Terry Bites]

If the uv is set externally to the ic then its hackable.

[Jwillis]

You just want a simple cut off switch for a 3.7V LiIon ?
Low Voltage Cutoff Switch
These will do 1S, 2S or 3S .Cut off at 3V
Might be able to find them cheaper some where else.

Quite costly, but the current capacity is very high. Most of my uses would be at much lower currents and these parts have too high of a parasitic loss. I'm power small micro controllers with batteries, it is probably possible to use the controllers ADC to monitor the batteries voltage and to turn itself off when the voltage is too low.

They are costly for sure. For the OP's requirements it should be ok .
Another option  is use a TL431 to reference a Mosfet .

[Peabody]

This is the best solution I've found for using the processor's ADC to measure the battery voltage.  It's just a resistor divider to get the maximum battery voltage down to below Vcc, but the circuit is interrupted by an N-channel mosfet.  So the only time current actually flows through the divider is when the mosfet is turned on, which the firmware can do as seldom as possible, and only for a few microseconds.  Then if it measures lower than you want, you can either switch everything off with a mosfet, or go to sleep.

[Calvin]

Hi,

the Low Voltage Cutoff Switch mentioned before draws several mA of current in Off status.
If that is not a typo it´d be unuseable for that purpose.
A µC is also out of question here(standalone device wo. µC), even though it might be the best in a different situation.
Seems I need to find a low Iq comparator that drives a PMOS. 

regards
Calvin

[David Hess]

Should I just compare battery voltage to reference and only when it is above 3V connect battery via mosfet to my device? But then the comparator will still be connected to battery, no?. I just don't know how to do this I've never done this and I don't know what to search for.

A micropower comparator and reference can be used, microamps or even lower is possible, or if the protection circuit is powered by the output, then when it disconnects the load it will disconnect itself also.