Author Topic: Battery pack with switches? Can it be done?  (Read 1193 times)

0 Members and 1 Guest are viewing this topic.

Offline mrburnzieTopic starter

  • Regular Contributor
  • *
  • Posts: 139
  • Country: cs
Battery pack with switches? Can it be done?
« on: May 31, 2019, 10:17:43 am »
Hi everyone!

I'm working on a project, where my electronics run off of 4.2V Li-ion batteries. The electronics are a couple of IC's and a 12V motor.

The IC's are powered by an LDO that drops the 4.2V to 3.3V, and the 12V is run by boosting the voltage from 4.2V to 12V.
My problem is that the efficiency isn't great when we look at the boost IC.

The reason I have a 4.2V battery pack (1S3P) is because I have a small solar panel that slowly charges the battery pack.

To solve the efficiency problem, I have come up with an idea of switching the wiring of the battery pack. Let me explain:



This is the wiring I have come up with (but for 1S2P it would be the same for 3P,4P ect).



This would be the wiring when I need 4.2V to power the electronics (which are in sleep mode) and the batteries are charging. 


This would be the wiring, when I switch off SW1 and SW3. The idea is that the electronics would be powered still off-of the first 4.2V battery, and the motor would be connected to the new 8.4V (or in 3S 12.6V) source.

Of course the switching order is crucial, so we don't short the batteries.

Is this a good solution, or should I go with the old solution (boosting the voltage for the motor)  :-/O ?

The motor would draw around 2A-3A.

Edit: the motor would be on for max 10s and would be switched over to the second image. So the 8.4V/12.6V source would be avaiable for only 10s.

« Last Edit: May 31, 2019, 10:31:59 am by mrburnzie »
"Talk is cheap, show me the code"

Anyone need of freelance software/hardware developer, hit me up!
 

Online Siwastaja

  • Super Contributor
  • ***
  • Posts: 9327
  • Country: fi
Re: Battery pack with switches? Can it be done?
« Reply #1 on: May 31, 2019, 11:05:45 am »
No, next to impossible to get this work sanely.

Improve the boost efficiency by redesigning (use synchronous topology), or go for fixed 3s, and buck it down if you have some parts that require lower voltage than the motor. If the 3V3 consumption is low enough, even direct LDO from the raw 3s voltage might be just fine.
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 3382
  • Country: gb
Re: Battery pack with switches? Can it be done?
« Reply #2 on: May 31, 2019, 11:34:00 am »
There is a good reason no-one does this, as Siwastaja says it's very difficult to make this work reliably.  With a mechanical switch you have to solve the contact timing problem, get it wrong and you'll likely weld the switch contacts together and catch the thing on fire.  Then you need to find a way to prevent plugging your 4.2v charger in when the switch is set to 3S mode...
 

Offline mrburnzieTopic starter

  • Regular Contributor
  • *
  • Posts: 139
  • Country: cs
Re: Battery pack with switches? Can it be done?
« Reply #3 on: May 31, 2019, 12:58:23 pm »
okay, let's say that the switches are mosfets controlled by an MCU. The charger would be also connected through a mosfet, so we turn it off when it goes into 3s mode. What would the problem be then? (besides it being unreliable)
"Talk is cheap, show me the code"

Anyone need of freelance software/hardware developer, hit me up!
 

Online Siwastaja

  • Super Contributor
  • ***
  • Posts: 9327
  • Country: fi
Re: Battery pack with switches? Can it be done?
« Reply #4 on: May 31, 2019, 01:15:23 pm »
The cells would be hard to guarantee staying in balance, so separate voltage monitoring for each series cell (a full blown cell-level BMS, that is) would be necessary.

Then, do you know what happens when you connect two cells with different voltages in parallel? They blow up. So you need an active DC/DC equalizer which automatically transfers charge between cells before they can be hard-connected in parallel. This equalization would take minutes, so very slow switching. The end result is expensive as a space station and provides absolutely no benefit over any simple solution.

You actually can pull it off as you describe, but without many expensive and complex safety systems it's bound to fail catastrophically one day. Even if you were willing to take the risk, why bother? I totally fail to see any benefit it offers over a simple one-battery system. Voltage conversion is cheap and efficient today.

For example, instead of putting many protected, current-limited smart MOSFET switches, change your current boost converter to a higher current synchronous type and you're done.
 

Offline mrburnzieTopic starter

  • Regular Contributor
  • *
  • Posts: 139
  • Country: cs
Re: Battery pack with switches? Can it be done?
« Reply #5 on: May 31, 2019, 01:21:19 pm »
Alright, thank you for the detailed explanation!
"Talk is cheap, show me the code"

Anyone need of freelance software/hardware developer, hit me up!
 

Offline ajb

  • Super Contributor
  • ***
  • Posts: 2785
  • Country: us
Re: Battery pack with switches? Can it be done?
« Reply #6 on: May 31, 2019, 05:00:41 pm »
Dewalt does this with their FlexVolt power tool batteries, the cells can be switched between a nominally 20V configuration for use in standard 18/20V tools or a nominally 60V configuration for higher voltage tools.  You do want to take care to match the cells, ideally by choosing cells from the beginning that are intrinsically well matched and also doing some balance monitoring to ensure they stay that way. 

An alternative approach to avoid having to build a high power boost converter for the motor may be to leave the battery in its higher voltage configuration, and boost the solar panel output instead (assuming that your solar output is lower than your motor draw, which it sounds like is the case).  Another advantage of this solution is you can do MPPT to improve the efficacy of your solar charging. 
 

Online Siwastaja

  • Super Contributor
  • ***
  • Posts: 9327
  • Country: fi
Re: Battery pack with switches? Can it be done?
« Reply #7 on: May 31, 2019, 08:36:21 pm »
An alternative approach to avoid having to build a high power boost converter for the motor may be to leave the battery in its higher voltage configuration, and boost the solar panel output instead (assuming that your solar output is lower than your motor draw, which it sounds like is the case).  Another advantage of this solution is you can do MPPT to improve the efficacy of your solar charging.

Definitely this. Power conversion is minimized, assuming the solar power is fairly low even in direct sunlight, a very smal boost will do.

The only thing left is to minimize the 3V3 usage (assuming it's just MCU, by utilizing sleep modes properly) so that you can use a low quiescent current LDO directly off the 3s pack.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf