Author Topic: Looking to (re)build a circuit to discharge my lipo batteries to like 3.9-3.7  (Read 5146 times)

0 Members and 1 Guest are viewing this topic.

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
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 ::)

The resistors divider is calculated for 3.8V and in any case 4.2V * .1A = 420mW so just a minimal difference.
In any case I never pretended to exactly replicate that circuit.

BTW here is a new simplified version with just one BJT and 150mA of average discharge current. Due to the BJT saturation voltage the resistor value should be adjusted.
(The BJT may still be replaced by a P-MOSFET)
« Last Edit: December 31, 2019, 09:13:13 am by not1xor1 »
 

Online Siwastaja

  • Super Contributor
  • ***
  • Posts: 9334
  • Country: fi
Apart that you are re-quoting Siwastaja's plain nonsense,

Not nonsense - no capacity was originally defined, so it could be anything. I was thinking a pack somewhere around 10Ah, and discharging down to somewhere around 30-40%, this would be days. Your calculation also slightly increased the discharge current from the schematic you posted first.

Your calculation is correct with your assumptions. Increase the resistor divider values to further decrease quiescent current, and the circuit is fairly good for your example case.
« Last Edit: December 31, 2019, 04:44:45 pm by Siwastaja »
 

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
Awesome stuff guys... I'm certainly learning things.


FYI, these are LI-PO batteries not LI-ION just for clarity sake in the need to properly store them at the proper voltage.
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
Awesome stuff guys... I'm certainly learning things.


FYI, these are LI-PO batteries not LI-ION just for clarity sake in the need to properly store them at the proper voltage.

to be sure you have to check the cell manufacturer datasheet
 
The following users thanked this post: gilligan

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
Apart that you are re-quoting Siwastaja's plain nonsense,

Not nonsense - no capacity was originally defined, so it could be anything. I was thinking a pack somewhere around 10Ah, and discharging down to somewhere around 30-40%, this would be days. Your calculation also slightly increased the discharge current from the schematic you posted first.

Your calculation is correct with your assumptions. Increase the resistor divider values to further decrease quiescent current, and the circuit is fairly good for your example case.

OK... I may have been too rude
please accept my apologies and ... BTW have a happy new year  :)
 
The following users thanked this post: Siwastaja, gilligan

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
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 ::)



The resistors divider is calculated for 3.8V and in any case 4.2V * .1A = 420mW so just a minimal difference.
In any case I never pretended to exactly replicate that circuit.

BTW here is a new simplified version with just one BJT and 150mA of average discharge current. Due to the BJT saturation voltage the resistor value should be adjusted.
(The BJT may still be replaced by a P-MOSFET)
(Attachment Link)

Quick question... how did you get the TL431 into LTSpice?

I'm VERY new to LTSpice and I'm looking and I think I get it, but everything I read/watch says "pick something similar" and I don't know which component would be similar to the Zener with Reference. :(
 

Online MarkF

  • Super Contributor
  • ***
  • Posts: 2763
  • Country: us
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
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.

Yep. They also claim "Brings a 1300mah lipo to storage voltage within 24 hours", so this is clearly advertised as well that it will take hours.

Yes they also claim "Low leakage current so the device can be left plugged in", but without actually seeing the schematic, I still kind of doubt it. Of course "low leakage current" doesn't mean much without any figure. To many non-EE people, a few tens of µA is a very low current... (and it's achievable with a simple resistive divider).

 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Quick question... how did you get the TL431 into LTSpice?

I'm VERY new to LTSpice and I'm looking and I think I get it, but everything I read/watch says "pick something similar" and I don't know which component would be similar to the Zener with Reference. :(

You can find the TL431 Spice model on TI's website for instance. https://www.ti.com/lit/zip/sloj174

Sorry I don't feel like explaining in details how to create a new part in LTSpice from a Spice model, but this has been explained before so you can look that up in this forum and on the Internet in general, you'll find tutorials (if someone else feels like explaining this again, feel free ;D )

 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
Actually there is a whole little universe of TL431 models from various sources and with various deficiencies.
What most of them have in common is not trying to model the IC at transistor level but using simplifications which may not simulate weird corner cases faithfully.

For example, the model you linked seems to have no mention of temperature and how it affects not only reference voltage (because it barely does) but also reference current, which OTOH can make a difference if high resistance divider is used. And the chip will get hot if you pass 100mA through it.
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
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.

Yep. They also claim "Brings a 1300mah lipo to storage voltage within 24 hours", so this is clearly advertised as well that it will take hours.

Yes they also claim "Low leakage current so the device can be left plugged in", but without actually seeing the schematic, I still kind of doubt it. Of course "low leakage current" doesn't mean much without any figure. To many non-EE people, a few tens of µA is a very low current... (and it's achievable with a simple resistive divider).

in any case it is an easy DIY project, there is no need to care about the vendor claims.

I just provided some ideas.
Some improvements may come from using a LMV431 (typical off current is just 1nA, max 100nA) and by carefully selecting the MOSFET (no need for a high power one which likely has high leakage).

I'm confident it would be not too difficult to achieve a discharge of 100-200mA and a leakage current of less than 1µA within a couple of euros/USD of cost.

Unfortunately I can't make any test now as my lab is not available (I'm very slowly working on a ventilation system with an air purifier with active carbon and up to H13 filter).
 
The following users thanked this post: gilligan

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
Actually there is a whole little universe of TL431 models from various sources and with various deficiencies.
What most of them have in common is not trying to model the IC at transistor level but using simplifications which may not simulate weird corner cases faithfully.

For example, the model you linked seems to have no mention of temperature and how it affects not only reference voltage (because it barely does) but also reference current, which OTOH can make a difference if high resistance divider is used. And the chip will get hot if you pass 100mA through it.

As soon as you use a MOSFET to switch off the current through the resistance divider you have to care only about the off current of the shunt regulator (LMV431 is probably one of the best choice) and of the MOSFET

Unfortunately I do not have LMV431 spice model, in any case I put the TL431 one in the zip file with a new simulation (see attachment). About 190mA of discharge current and still around 1µA of off current.
 
The following users thanked this post: gilligan

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Actually there is a whole little universe of TL431 models from various sources and with various deficiencies.
What most of them have in common is not trying to model the IC at transistor level but using simplifications which may not simulate weird corner cases faithfully.

For example, the model you linked seems to have no mention of temperature and how it affects not only reference voltage (because it barely does) but also reference current, which OTOH can make a difference if high resistance divider is used. And the chip will get hot if you pass 100mA through it.

Yes, looking at it, the TI model looks pretty crude. ;D
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
I think that if OP is still with us and serious about building something, the best way is to just put together a prototype on a breadboard and test it in practice and tweak until it works.

An easy and safe way to test is with a lab PSU or adjustable LM317 circuit powered from some USB charger or similar thing. This way you could quickly simulate behavior at various voltages, see how it reacts to heating up etc. If the TL431 blows up, not much is lost ;)
« Last Edit: January 02, 2020, 09:42:39 pm by magic »
 

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
I think that if OP is still with us and serious about building something, the best way is to just put together a prototype on a breadboard and test it in practice and tweak until it works.

An easy and safe way to test is with a lab PSU or adjustable LM317 circuit powered from some USB charger or similar thing. This way you could quickly simulate behavior at various voltages, see how it reacts to heating up etc. If the TL431 blows up, not much is lost ;)

I'm still here and eagerly waiting for parts to arrive to do just that!

Till then I'm playing in LTSpice (one, to see how it all works together and two, to understand LTSpice finally!)
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Just the info from the vendor: https://www.analog.com/en/technical-articles/ltspice-simple-steps-to-import-third-party-models.html

Looks like creating parts from Spice models has gotten a lot easier than what it used to be, and I was still using the old, "manual" way of editing/creating files by hand... so I did learn something as well today. ;D

Long story short:
* Open the Spice model file in LTSpice. (Ex: TL431.mod) It will open as a Spice file (basically text editor mode);
* Select the subcircuit name you want to create a part from (ex: 'TL431' on the subcircuit line '.SUBCKT  TL431 1  2  3');
* Right-click on the selected name, a menu pops up, select "Create symbol";
* A confirmation message appears, choose 'Yes';
* A symbol now appears in the symbol editor. You can customize its shape and rename the pins that correspond to the subcircuit nodes by default (only change the pin names, not the Spice order!);
* Save it. You can now close the symbol and the Spice file windows. The new part is available for use without having to restart LTSpice.

Note: the symbol editor in LTSpice is terrible. Even more so than the schematic editor. You've been warned. ;D
 

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
Just the info from the vendor: https://www.analog.com/en/technical-articles/ltspice-simple-steps-to-import-third-party-models.html

Looks like creating parts from Spice models has gotten a lot easier than what it used to be, and I was still using the old, "manual" way of editing/creating files by hand... so I did learn something as well today. ;D

Long story short:
* Open the Spice model file in LTSpice. (Ex: TL431.mod) It will open as a Spice file (basically text editor mode);
* Select the subcircuit name you want to create a part from (ex: 'TL431' on the subcircuit line '.SUBCKT  TL431 1  2  3');
* Right-click on the selected name, a menu pops up, select "Create symbol";
* A confirmation message appears, choose 'Yes';
* A symbol now appears in the symbol editor. You can customize its shape and rename the pins that correspond to the subcircuit nodes by default (only change the pin names, not the Spice order!);
* Save it. You can now close the symbol and the Spice file windows. The new part is available for use without having to restart LTSpice.

Note: the symbol editor in LTSpice is terrible. Even more so than the schematic editor. You've been warned. ;D

Yeah, that's what I ended up doing.

It's quite frustrating and all the tutorials give examples where they can pick known components... like a diode or op-amp.

Found one that showed an 8 pin chip and realized i could just redraw it to be what I wanted.

Very clunky.
 

Offline not1xor1

  • Frequent Contributor
  • **
  • Posts: 716
  • Country: it
It's quite frustrating and all the tutorials give examples where they can pick known components... like a diode or op-amp.

Found one that showed an 8 pin chip and realized i could just redraw it to be what I wanted.

Very clunky.

in the zip file I attached in my previous message there are both a TL431 model and an asy (symbol) file.
 
The following users thanked this post: gilligan

Offline gilliganTopic starter

  • Contributor
  • Posts: 38
  • Country: us
It's quite frustrating and all the tutorials give examples where they can pick known components... like a diode or op-amp.

Found one that showed an 8 pin chip and realized i could just redraw it to be what I wanted.

Very clunky.

in the zip file I attached in my previous message there are both a TL431 model and an asy (symbol) file.

Yes, thank you very much!

I have built a 1S, 2S, 3S, and 4S circuits with your help.

I'll post them soon for feedback as I'm sure I screwed some things up. LOL
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf