Author Topic: Looking to (re)build a circuit to discharge my lipo batteries to like 3.9-3.7  (Read 5148 times)

0 Members and 1 Guest are viewing this topic.

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
These guys make a pretty slick little device that drops the voltage of batteries down to their storage level.

RC stuff comes in 1S, 2S, 3S, 4S.... configurations which are 1 in series, 2 in series, etc.  So the battery drain level is different for each configurations.

$5 is not a bad price for these guys but I also have a bunch of 1S batteries that I need to discharge to storage voltage.

Can anyone look at this and napkin up me a schematic and/or explain how they are doing this?

That's either a shunt resistor or more likely a higher wattage load resistor (I have a few of those), with a small driver circuit for an LED.  Are they basically using that LED as the driver for the transistor running to that load resistor which once it drops below a certain voltage (set with the other resistors) the LED goes off and the drain stops happening?

 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
the 3 terminal device might be a TL431 or similar shunt regulator
as soon as the supply voltage gets below the value programmed by the resistors divider it goes off and let just 0.1-1µA current through it

here is a simulation with a capacitor used as a rough battery model
898428-0

the green trace is the battery voltage dropping to about 7.6V (60% charge) at the end and the red one is the discharge current
« Last Edit: December 29, 2019, 06:46:06 am by not1xor1 »
 
The following users thanked this post: gilligan

Offline eblc1388

  • Frequent Contributor
  • **
  • Posts: 400
  • Country: gb
not1xor1 has shown you a 2S circuit and I'll show you the 1S discharge circuit, similar to his. I added the LED for indication. The LED will go off when the discharge process is completed.

The main discharge current is through the 68R resistor and the LED. This will continue until the cell voltage has dropped to the desired cutoff level, selected by user as per the value of resistor R5.

There is is still some small current drain when the LED turns off so the circuit is not meant to be connected to the cell for storage purposes as it will slowly drain the cell down.

 
 
The following users thanked this post: gilligan

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
A single transistor could perhaps replace the TL431 if you can live with a few % of discharge voltage variation over typical room temperature range. Also, a low voltage TL431 variant like TLV431 or many others could be useful.

It is interesting however that the original circuit clearly uses some other method, because resistor values just don't make sense.

These are, as anyone can see: 31Ω, 5100Ω and 3.6/5.1/8.2MΩ, depending on battery voltage. They must be using some clever obscure IC here.

edit
Let's do some math.
Reference input current of TL431 is 2µA typical, flowing into the chip. This times 3.6MΩ plus 2.5V makes 9.7V, 3.2V per cell.
TL431 output saturates down to 2.5V, leaving between 7V and 10V across the resistor, for 30~40mA of discharge current and up to 400mW dissipation in the resistor.

For 4S it's 3.17V / 33~23mA / 460mW and for 6S 3.15V / 20~15mA / 450mW.

Could it be? :scared: :-DD
« Last Edit: December 29, 2019, 11:14:29 am by magic »
 
The following users thanked this post: gilligan

Online Siwastaja

  • Super Contributor
  • ***
  • Posts: 9335
  • Country: fi
It takes ages (days to weeks) to discharge through the small resistor and TL431, so you need to leave it mounted. Then there's the risk you forget it connected, and because the quiescent current is fairly high with that resistor divider in there, there's a risk of destroying the battery by overdischarge during storage.
 

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
It takes ages (days to weeks) to discharge through the small resistor and TL431, so you need to leave it mounted. Then there's the risk you forget it connected, and because the quiescent current is fairly high with that resistor divider in there, there's a risk of destroying the battery by overdischarge during storage.

What do you recommend?
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Many decent LiPo chargers for RC have an option to charge/discharge 1S to nS battery packs to storage level in a much shorter amount of time. I have a Imax B6 around, it's relatively cheap and does the job.

As said above, I wouldn't leave one of those circuits plugged to a battery for extended periods of time because it will eventually discharge it completely (due to the resistive divider). Now if the recommended use it to monitor the circuit and unplug it when the LED turns off, as it may take hours to get there, I think it's pretty unpractical. I think the fact it's cheap and stand-alone would really only make sense if you could leave it plugged to the battery and forget about it until you use the battery again...

An equivalent circuit that would draw minimal current once below the threshold would require more care and probably a specialized IC (with fixed internal divider) which I doubt they used here, or more components...


« Last Edit: December 29, 2019, 04:13:13 pm by SiliconWizard »
 
The following users thanked this post: gilligan

Offline NiHaoMike

  • Super Contributor
  • ***
  • Posts: 9321
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
What about add a buzzer or bright flashing LED that activates when the voltage is reached?
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
Try this with TL431:

anode to B-
cathode to B+ through 22Ω
reference to B+ through 330kΩ

I don't necessarily think this is a great solution but it looks like what they are doing so perhaps it works :-//

It may have the advantage of drawing very little current after discharge is complete.

Certainly watch it while it approaches 3.1V and disconnect the battery if it doesn't stop by 3V. It should stop around 3.16V.
Maybe use a battery with protection circuit for testing :-DD
 

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
I got ya...

So, what about this... what if I put some small relays across the battery... drops below the threshold and click the relay let's go.  The coil of the relay would also help drain the battery.
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
I don't think relays can be expected to have well defined turn-off voltage which will be repeatable and stable over long time. These are mechanical devices with springs, pivots and all sorts of unreliable stuff.

I would first try the TL431 circuit described above, simply because I suspect that this is how those things that you posted work. But test it first. Or test it on a lab PSU, that would be quite trivial.
 
The following users thanked this post: gilligan

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
A relay... that would be absolutely not accurate, not by a long shot. The added fun would be that it would probably get into an oscillating state when the voltage gets close to the min operating coil voltage... furious clicking... Fun stuff. :-DD
 
The following users thanked this post: gilligan

Online tunk

  • Super Contributor
  • ***
  • Posts: 1099
  • Country: no
Would an LED in series with a resistor and one or two diodes work?
 

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
Would an LED in series with a resistor and one or two diodes work?

The original has an LED and a fat resistor.
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
A single transistor could perhaps replace the TL431 if you can live with a few % of discharge voltage variation over typical room temperature range. Also, a low voltage TL431 variant like TLV431 or many others could be useful.

TL431 is the cheapest and unlike other chips withstand up to 100mA of current

Quote
TL431 output saturates down to 2.5V,

saturation voltage should be a little less than 2V
 

Online Kasper

  • Frequent Contributor
  • **
  • Posts: 793
  • Country: ca
A relay... that would be absolutely not accurate, not by a long shot. The added fun would be that it would probably get into an oscillating state when the voltage gets close to the min operating coil voltage... furious clicking... Fun stuff. :-DD

The clicking could notify the user the drain is complete ;)
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
It takes ages (days to weeks) to discharge through the small resistor and TL431, so you need to leave it mounted. Then there's the risk you forget it connected, and because the quiescent current is fairly high with that resistor divider in there, there's a risk of destroying the battery by overdischarge during storage.

Days to weeks ?  :scared:

Lets make a rough calculation (assuming 100% efficiency and constant voltage battery).
A 2000mAh cell, at 100% capacity holds an energy of 4.2V * 2A * 60 minutes * 60 seconds = 30.24kJ
To discharge it to 60% of capacity you have to take away 40% of energy, that is 30.24kJ * 0.4 = 12.096kJ.

A TL431 withstands 100mA of current. Lets make it 80mA of average current (I didn't care much of that in the circuit I previously attached), now that means that at 4.2V and 80mA the needed time would be : 12.096e3 / (4.2*80e-3) = 36e3 seconds, that is 10 hours.

So unless I made some coarse mistake and unless you want to discharge the battery pack of an electric SUV I think it would take quite less than days or weeks.  :phew:

The schematic I posted was just a quick guess of that in the photo. I do not know if they use a dedicated chip or if that is just the usual careless design with a generic shunt regulator.

In any case, with just few more components, it is easy to design a few µA leakage circuit (around 1µA in LTspice). It would cost no more than 1-2 euros/USD and might be a good DIY project.  8)

899058-0

notes:
- all resistor values should be changed according to the number of cells and the type of LED, do not forget about the maximum MOSFET gate voltage: a voltage divider or zener is needed in case of more than 4 cells.
- I just chose the first low Vth MOSFET model in LTspice, I've no idea about the price and other features, but there are plenty of suitable MOSFETs
- C2 is there to start the circuit. It might cause TL431 instability (oscillations) and would require a discharge button, so it might be better to just replace it by a (momentary ON) push button with a series resistor (or better just use a bit overrated power resistor for R3)
- do not forget to take into account TL431 and resistor tolerances.

BTW - AFAIK - Li-ion batteries are not affected by the Ni-Cd memory effect and last longer if the discharge current is in the order of several tenths of mA rather than A
« Last Edit: December 30, 2019, 05:56:44 am by not1xor1 »
 
The following users thanked this post: thm_w, Siwastaja, magic

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
A relay... that would be absolutely not accurate, not by a long shot. The added fun would be that it would probably get into an oscillating state when the voltage gets close to the min operating coil voltage... furious clicking... Fun stuff. :-DD
Relays have hysteresis, the battery would need a bit of ESR for that. But a relay oscillator is an intriguing idea nevertheless ;)

saturation voltage should be a little less than 2V
Fair point, it may be limited only by saturation of the reference input transistor. I didn't know.
It doesn't change the calculations much, except for my 1 cell version which should definitely increase cathode resistance then.
 

Offline eblc1388

  • Frequent Contributor
  • **
  • Posts: 400
  • Country: gb
It takes ages (days to weeks) to discharge through the small resistor and TL431, so you need to leave it mounted.

With a slight change, one can also control the discharge current by using different value of R2. The attached example shows a discharge current of 1.2A. The cell voltage would drop very quickly within minutes to hour. Care in the choice of PMOS which should be logic level for it to work.



 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
It takes ages (days to weeks) to discharge through the small resistor and TL431, so you need to leave it mounted.

With a slight change, one can also control the discharge current by using different value of R2. The attached example shows a discharge current of 1.2A. The cell voltage would drop very quickly within minutes to hour. Care in the choice of PMOS which should be logic level for it to work.

Apart that you are re-quoting Siwastaja's plain nonsense, there is no difference if the batteries are discharged in half an hour or ten hours.

A quick discharge decreases the battery life and increases cost and component size because of more power dissipation while one can forget to disconnect the battery from the discharger circuit in both cases.

In the second circuit I proposed power dissipation is in the order of hundredths of mW and once the set voltage is reached the current leakage is in the order of µA.
After one month a forgotten 2000mAh cell would lose 0.1-1% of capacity.
« Last Edit: December 30, 2019, 04:58:26 pm by not1xor1 »
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
A relay... that would be absolutely not accurate, not by a long shot. The added fun would be that it would probably get into an oscillating state when the voltage gets close to the min operating coil voltage... furious clicking... Fun stuff. :-DD

The clicking could notify the user the drain is complete ;)

True. ;D

I suspect the clicking would actually not be that furious though, it would likely be more like slow cycles.
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
In any case, with just few more components, it is easy to design a few µA leakage circuit (around 1µA in LTspice). It would cost no more than 1-2 euros/USD and might be a good DIY project.  8)

(Attachment Link)

Yup. Nice.
 

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
The original is designed to drain 6s batteries... Those are typically pretty large mAh... Would this tl431 be able to do that?

I'll likely whip one or two of these up to try... Especially given my batteries are coming in WAY before the rest of my gear lol.
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
The original is designed to drain 6s batteries... Those are typically pretty large mAh... Would this tl431 be able to do that?

I'll likely whip one or two of these up to try... Especially given my batteries are coming in WAY before the rest of my gear lol.

That's not rocket science, just basic physics and math.

If you do not feel confortable with that old brewer Lord Joule, let's just speak about A * h.

A 10000mAh battery can provide 10A for an hour.
If you discharge that at 100mA (the limit for a TL431) it would FULLY discharge in 10000/100 hours. That is 100 hours.
But we need to discharge them up to 60% of capacity so if there were no losses in the conversion between chemical energy and electrical one that would be 40 hours.

Now taking into account the energy conversion losses and a safer (for TL431) current of 80-90mA the discharge time would still be around that, likely 35-45 hours.

And do not forget the original post. The circuits portrayed in that photo are quite likely in that range of discharge current, 50-100mA.
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
They claim 500mW discharge power.

https://www.hobbyrc.co.uk/lipo-suction-4s-xt60

So I wasn't that far off but they also claim overtemperature protection so perhaps it's not TL431 after all. And 3.9V rather than 3.2V termination so I'm not sure how they are doing it with so few components. And low leakage.

Maybe OTP is a polymer fuse ::)
« Last Edit: December 31, 2019, 08:45:54 am by magic »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf