• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Build a li-ion 36V battery pack charger with a Boost regulator

evaleto

New Member
Hi,

I'm wondering if is ok to use a DC/DC Boost Step-up regulator (like UC2577-ADJ (the simplest), LTC3862-2 or LTC3873-5) as charger for a lithium ion battery pack (42V 2A charging specs). I want to build a simple charger with those specifications :
  • input DC source (input 18-24V/4A),
  • output a precise 42V +/- 1% for charging ,
  • output current limiting (empty battery is like 0 ohm wire, a shortcut!! ) - I don't know if DC/DC can control this
  • output shutdown when Icharge=1/100C (of battery capacity). - by adding a Attiny + FET
  • reverse polarity protection with a Schottky diode
For the current limiting implementation, my guess is the power source will do the Job (any laptop power has ~20V 4A limitation ).
For safety I have to know if this development plan is right. The result will be publish as an opensource project.

Any help will be appreciated,

Olivier
 

Sceadwian

Banned
You can't build a 'simple' charger for a 36V lithium pack and expect safe results, each cell needs to be monitored individually and has to have some method to ballance control individual cells during charging. For a pack of that size, multipoint thermal monitoring would be recommended as well. Some 'simple' chargers for packs that size rely on PTC resistors to balance the pack, but this is just cheap engineering, an accidental long slow overcharge of a pack that relies on this 'simple' method could explode, all it would take is one bad cell to cause this long slow overcharge.

Single Lithiums at 3.6 volts (nominal) are trivial to charge easily, even 2-3 maybe even 4 series packs are safe with PTC current limiters. But you're talking about a 10 series pack, the only way to do that right is to monitor each individual cell and control charge rate for each cell individually.
 
Last edited:

evaleto

New Member
Thank you Sceadwian,

You mean I have to split my charger in two stages; one for power (my previous post) and one for battery protection and cell balancing with the bq77910a or bq77pl900?
I'm still wondering about the power stage regulation. For battery charging I need a precise 42V +/- 1% and a current limitation. I would like to know if those boosts regulators can achieve part of that LTC3873-5, uc2577-adj.

My guess is to plug the simplest boost regulator with the bq77910 (or bq77pl900) .
 

WTP Pepper

Active Member
The simplest solution for this charger is to charge each cell individually. As already said, you need charge balancing on this number of cells so you don't have a single point of failure.

Build a simple single cell charger using one of many Li charger ICs (MAX1811 comes to mind but there are others), but power it from a DC-DC convertor (5v output) that has an isolated output. These standalone small modules are readily available.

Power all of these chargers in parallel, but series wire the cell output.
Use the LED output from the individual charger ICs to monitor the whole pack and spot early failures of individual cells.

Using this method you don't need any complex boost circuit as you only need enough volts to power a single DC-DC convertor, but need a decent current supply to run 10 of them in parallel.
 
Last edited:

evaleto

New Member
Waw,

I was looking for a solution like you suggest without success. There is no (positive) information about the stackable implementation. I didn't find any publication about that!!! Could you tell me more about this implementation, is there any doc about that (which dcdc, which isolation)?
Moreover, for my battery pack, I also need a battery protection (over/under voltage, shortcut, temp fuse, ...). Should I implement protection with the same way of the charging ? Is there any single ICs that do charge and protection?

About the balancing, the bq77910a manage protection and balancing for 10 cells. I thought I only need to provide 42V and 2A to manage the charge.

Cheers,
-- olivier
 

Sceadwian

Banned
About the balancing, the bq77910a manage protection and balancing for 10 cells. I thought I only need to provide 42V and 2A to manage the charge.
You should have mentioned you were using the bq77910a in the first place that changes everything! You can't expect useful responses if you don't provide all the information! Your original post said absolutely nothing about the IC you were planning on using.

Yes, the bq77910a is exactly what you need to be using as it does all the things that are required for pack monitoring and balancing, although you will have to configure it to fit your needs, an improper configuration is just as bad as using no protection circuit. The IC however does not control charging, it simply monitors cell balance and warns of over/under voltage during operation, the charger itself will have to be turned off by some other method or the pack will charge until an over-voltage is detected and the protection IC kicks in (this would not be ideal but it will get you to a 1st stage charge condition if set right, delays and hysteresis settings would have to be set properly)

This isn't a charge control IC though, just a protection/balance IC. It monitors each cell for over/under voltage, you could set the 'over' voltage to a safe level which would be 4.1 volts (for common lithium polymers see your battery makers specification) but this is not what the IC is designed for. Basically you would program this IC for the max/min voltages allowed by the battery maker and set it up to balance cell voltages during charging, but you would still need an external charger.

The simplest 'full feature' Lithium charger you can design is actually not that complicated. You use a constant current source of about 1C (C=battery capacity in amphours) at this point the voltage is still monitored for being in range for the protection IC (it can't protect itself from externally applied voltages) until the total cell voltage equals 4.1-4.2 volts. Once this peak voltage has been reached the charger switches over to the second stage charge which is constant voltage until the current decreases to approximately 1/10th of the initial charge current, at which point the charge must be terminated.

When used as a basic charger using the over voltage protection this IC will at best give you first stage charging, which is 80% of the cells capactiy. To get the last 20% of capacity out of the cell/pack you need to use the secondary method, however it is not good for the cell to remain at this stage, so quiet honestly only doing the first stage of a Lithium charge will increase pack life a lot and only slightly effect capacity. To what extreme's you need to take charge/discharge considerations into effect is determined by the application itself which you have described nothing about.
 
Last edited:

Sceadwian

Banned
Isolation is not required as the IC he's using balances the pack. All he needs is a constant current source with a voltage limit for first stage charging. Again this can be done (but it's not ideal) using just the IC he's chosen. One of it's little brothers is probably in a lot of inexpensive lithium power tools.
 

evaleto

New Member
You should have mentioned you were using the bq77910a in the first place that changes everything! You can't expect useful responses if you don't provide all the information! Your original post said absolutely nothing about the IC you were planning on using.
Yep, Sorry for this misunderstanding. In my mind I did a split for the charging and the protection circuit. I'm newbies :) ... after reading the datasheet of the bq77pl910 I was wondering if a simple voltage regulation with current limiter will be enough to charge my pack. And this was the subject of my post. Can I use a simple boost regulator to regulate a precise 42V, I guess the current limitation will be done by the power supply before my module. I will read more about the balancing management by the bq77pl910, It seems not to be like I tought.

To what extreme's you need to take charge/discharge considerations into effect is determined by the application itself which you have described nothing about
I'm building an open source electric bicycle with a ultra small package and small weight. I decided to start with the battery to control the size, the layout and the charge.

What, would be, in your mind, the ideal to charge and protect 10 serial lithium ion battery?

Thank you for helping.
 
Last edited:

Sceadwian

Banned
What cells are you using for the battery? That will be the defining point. You can not use cheap cells in a pack even with a ballancing cirucuit.
 

ChrisP58

Well-Known Member
One thing to remember about boost converters is that they have no control when the output voltage is less than the input. And while your 36 volt pack should never be less than 30 volts, there may be fault conditions where it could be. If you don't provide some other means to block the current from the source supply, you could end up damaging a deeply discharged pack, instead of charging it.

The simplest is to use a buck converter that is higher than the 42 volt target, or use a converter topology that allows a controlled voltage both above and below the input voltage. A SEPIC converter would work, or a single inductor buck-boost converter. See TIs LM5118 or Linears LTC3789 for examples of the latter.
 

Evalon

New Member
Hi evaleto - this is not to talk you out of this but you mention in one of your posts that you are a newbie. To do what you are intending you may have to do PCB layouts (many of the ICs are quite small), get into the different characteristics of different Lithium-ion batteries, experiment with prototypes that may - or may not - work first off, and read and understand datasheets that sometimes contain highly specialized information.

Li-ion batteries are tricky battery-types in terms of providing "exactly" the kind of charging and discharging conditions that fulfill their needs (sceadwian touches upon this) so that they are charged close to 100% (if so desired) and also last long. Many companies make chargers specifically for these purposes - also for bikes ... Anyway, should you decide to go for a finished pack there's a forum that is very actice in this field, and my guess is that if you post a thread there they can help you find a suitable vendor. They probably also have finished DIY designs for battery packs that already work. The forum is:

http://www.pedelecforum.de/forum/forum.php

It's in German but I have posted there in English with no problems of understanding and much help!

Best of luck in your endeavors whatever you chose to do ;-)

Jesper
 
Last edited:

Sceadwian

Banned
ChrisP58 said:
you could end up damaging a deeply discharged pack, instead of charging it.
This can not occur, the protection IC would prevent it.

WTP Pepper said:
He does need isolation if he chooses to use the configuration I posted.
You're not keeping abreast of his previous posts, the configuration you recommended is not required because the charge control IC he's selected already does all the work for balancing during both charge and discharge cycles and functions as both high/low voltage shut off, over current, short circuit, and even thermal overload protection. It doesn't do anything sophisticated for charging, however if it is programmed with conservative timeout reset values for the over voltage shut off, it will allow charging from a constant current source that's rated for a few volts over what is required. Once the voltage goes 'over' which would actually be set to the 80% charge voltage of 4.1 it would shut the pack off for a long delay, it would periodically restart and repeat the charge to 4.1 volts and shut off again, it would perpetually keep the pack at an 80-85% charge state which is ideal for Lithium Cell lifespan. You lose about 20% of the cells total capacity for discharge but this prolongs pack life and requires no advanced charging ciruitry.

Evalon, Lithiums aren't really that tricky to charge in single cells, and the control IC he picked essentially turns the entire pack into what appears to the outside world to be a single Lithium cell.
 
Last edited:

Sceadwian

Banned
Good quality cells with detailed datasheet's, always helps! Your specification for a 2amp current limited supply that can provide as much as 44 volts is appropriate for the simplest possible charging. The optimal voltage for a 10 cell Lithium pack is 42 volts, but you need a couple extra volts of headroom to account for the slight power drop from the balancing system itself, the protection IC will prevent cell over-voltage from occurring.

If the protection IC is programmed to engage overvoltage mode (shut off the pack) at 4.3(perhaps 4.2 for increased longevity) volts per cell with a hysteresis of 200mv (or 100mv for 4.2 volts) it will meet Panasonic charging recommendations, although you'll not quiet fully charge the pack, this slight loss in capacity is the tradoff for less sophisticated charging circuitry, again though this extra capacity that can be had also comes with a decrease in pack life as if the cells are peak charged the chemistry degrades slightly over time, Lithiums are happiest at an 80% charge and should be kept as close to that as is possible, and should in general never be discharged bellow 20% or at least should be immediately charged after reaching that state, with an IC like you're planning on using the device should be kept plugged in as much as is possible.

This is one of the reasons it's so important to properly derate the capacity of the cells you chose for the application as if you buy a pack that can only just barely last the length of time you need it to through the extremes of the Lithium cells capacity you will have a reduced pack life due to the stress on the chemistry.
 
Last edited:

evaleto

New Member
thank you all for those informations. I will prepare a schematic/pcb with the implementation of power stage and protection + balancing stage. Finally I will use the uc2577-adj as boost regulator set for 43V (I guess) and the bq77pl900 for battery management. The hard part willbe the implementatio of I2C code .... If you are interrested I can publish results on this thread, as soon it's finished!

-- olivier
 

evaleto

New Member
One thing to remember about boost converters is that they have no control when the output voltage is less than the input. And while your 36 volt pack should never be less than 30 volts, there may be fault conditions where it could be. If you don't provide some other means to block the current from the source supply, you could end up damaging a deeply discharged pack, instead of charging it.

The simplest is to use a buck converter that is higher than the 42 volt target, or use a converter topology that allows a controlled voltage both above and below the input voltage. A SEPIC converter would work, or a single inductor buck-boost converter. See TIs LM5118 or Linears LTC3789 for examples of the latter.
What do you think about this boost regulator uc2577-adj it is really simple to implement,
 

ChrisP58

Well-Known Member
I still discourage use of a boost converter for a battery charger.

I once used a boost converter from a 24 V supply to charge a 30cell NiCd pack. It fit all of the expected operating parameters, since the product shut itself off at 30V (1v/cell) But after I had about 150 systems in the field we started getting complaints from some customers.

It turns out that even though the product shut down, there was some leakage current that continued to discharge the battery pack. Also, the packs were often swapped out and left discharged for some time, further pulling the voltage below 24 volts.

As a result, the open-frame 24V supply that I was using to run the charger would go into hick-up mode when a deeply discharged pack was connected. Since there were other functions powered by that supply, they all shut down too. Eventually, the high current pulses between start up and current limit did charge the battery up to the point where the boost converter could control it, but it was bothersome to the end user as the lights on the panel kept going on and off as the uC kept going through it's start up cycle.

In your case, the protection circuit in the battery will prevent catastrophic failure due to over current charging, but that doesn't mean 'soft' damage or degradation won't happen if it is exposed to short. repetitive current pulses of uncontrolled magnitude.

As for charging to 43 Volts (4.3V/cell) this is beyond what is recommended for the cells. You will get a bit more charge into te pack, and a little more run time per charge, but you will significantly reduce the cycle life of the battery pack.

The protection levels provided by the safety circuit in the pack should never be called into action unless something has failed in either the charger or the end product. The threshold levels in the protection are to prevent catastrophic failure, but degradation can and will happen between the recommended levels and the fault levels.
 

evaleto

New Member
Good to read your advise!
As a result, the open-frame 24V supply that I was using to run the charger would go into hick-up mode when a deeply discharged pack was connected. Since there were other functions powered by that supply, they all shut down too. Eventually, the high current pulses between start up and current limit did charge the battery up to the point where the boost converter could control it, but it was bothersome to the end user as the lights on the panel kept going on and off as the uC kept going through it's start up cycle.

In your case, the protection circuit in the battery will prevent catastrophic failure due to over current charging, but that doesn't mean 'soft' damage or degradation won't happen if it is exposed to short. repetitive current pulses of uncontrolled magnitude.
I have no religion for the implementation except I want to use common power source of laptops. They provide good AC conversion and current limitation (24V6A). I guess that some of your advises don't concern my needs. First I don't have other function powered by my supply on the charge, and the current limitation is already done by the laptop power supply. Am I wrong? What is the *correct* implementation? Is the suggestion from Sceadwian a good solution?

Sceadwian said:
Build a simple single cell charger using one of many Li charger ICs (MAX1811 comes to mind but there are others), but power it from a DC-DC convertor (5v output) that has an isolated output. These standalone small modules are readily available.
Power all of these chargers in parallel, but series wire the cell output.
Use the LED output from the individual charger ICs to monitor the whole pack and spot early failures of individual cells.
Can I power all these chargers in parallel of the protection circuit? And how do I isolate the output of the dcdc convertor?

Cheers,
-- olivier
 
Last edited:

Latest threads

EE World Online Articles

Loading

 
Top