Charge li-ion efficiently for cheap and balancing? go here

[Doctorbass]

Still in link to the "the single cell charging post"

Well.. i see that many people are more and more interested about lithium cells and abandon nimh and sla

But it seems that some still have a problem about charging or bothering with balancing feature problem..... and 3.6 or 4.2V

By my side i really think i have THE solution to solve that for cheap.

I've been writing many post about that. here is one more :lol:

But if you still are interested to charge fast and make it simple read this below:

You only need a little bit of DIY skill and to not be lazy..

That method suggest you to charge each paralle group cell with one charger module.I mean cheap because this method is very versatile and offer you multiples possibles voltages for different chemistry IN ONE CHARGER! and that the performance will exceed most of the comercial charger availlable for ebikes.

THE OTHER MAIN ADVANDAGE IS: that with the same charger you could be able to charge A123 (3.6V) AND normal li-ion cells 4.2V) by simply replacing a resistor value on the dc-dc....not bad!
To do that follow this:

-Determine how many serie cell you have in your entire pack

-Determine in how much time you would like to charge your pack

-Go on ebay and search for : dc-dc 5V

-Try to find the auctions that have availlable at least the same number of dc-dc than the number of serie cell you have.

ex: you have 12 cells: search for 12 item or more batch availlable. often you will find them at 9.99$ each

If your pack have 10Ah, and you want to charge in 2h, you will need dc-dc that have at least 5A

Buy a main power supply that can supply all the dc-dc input at the same time.
see how to calculate below:

let say 12x A123 cell = 43.2V
and that you charge at 5A mean you need that output power: 5A x 43.2V = 216W

DC-DC are efficient to around at least 80% so you will need to multiply by 1.25 that power for the input... that mean 270W of input total on all the combined dc-dc parallel input.

Generally DC-DC work with 36 to 72V on their input and have a regulated voltage and a current limit on the output and input. Some have 12V or 15V input but they are rare.

Ideally a 48V power supply do the job great.
Now for a 270W need, a 350W power supply should be of with a little safe margin.

Now, all you would need is to order the main power supply and the dc-dc that you determined, to get them and assemble these like i suggest here:

first, you will need to begin with the dc-dc voltage adjustment

-search on google the pdf spec of your dc-dc. (use dc-dc, their brand or model and pdf key words)

-buy a 10k 10turn potentiometer

-solder their pin 1 and 2 on the trim down input pins of the
dc dc.

(if your dc-dc have not sensing pins, skip the next step)
-solder the sense (-) pin on the out(-)
-solder the sense (+) pin on the out(+)

-put the 48V juice on the input pin of the dc-dc, plug your voltmeter on the + and - output pins and adjust the voltage to the recommanded cell max charge voltage.. often 3.6 or 4.2V.. try to adjust teh best as you can !

-do that for all dc-dc you need

-if recommanded on the spec add some heatsink on them.. (if you have enough space you can fix all them on the main power supply alluminium top cover if it have a fan it should work fine.

-solder all their voltage input in parallel (all + in parallel) and (all - in parallel)

- solder all the seperate output in serie and solder one wire for each cell tap you have
(ex: you have 12cells, you will need 13 wires)

-try to choose wire size awg 16 for 5A charging, 14 for 8A charging and 12 for 15A charging.

Add multipin connector between battery and the cahrger

It's done!

You now can safely cahrge all the cell you want faster than any other method and have in mind that the balance will be perfect after the charge !

Generally the total cost should vary from:

100$ for a 5A 10s charger balancer,
175$ for a 10A 10s charger balancer
250$ for a 30A 10s charger balancer

I really belive THAT IS THE BEST SOLUTION and that will keep your cells top shape charged at their max capacity without loss produced by seperated balancer that draw precious mAh !

:wink:
Doc

 

[safe]

You are duplicating your threads... :wink:

This one is good though because you answered everything.

You seem to have found "the answer" for a low priced (and powerful) charging solution for LiFePO4.

P.S: You answered my trim question.

Hey Folks... This might be "It".

The next issue is to get some kind of overall review of the DC DC chip products to see what the best prices are and which are better than others. There are many of these products to choose from...

 

[PJD]

Doc,

Can you provide a quick scematic of this - I'm not sure what you mean by connecting all the outputs in series. You mean the outputs of each DC-DC go to the + and - terminals of each cell, right?

 

[Doctorbass]

Doc,

Can you provide a quick scematic of this - I'm not sure what you mean by connecting all the outputs in series. You mean the outputs of each DC-DC go to the + and - terminals of each cell, right?

Yess. your cells are connected in serie to get more voltage.. so the dc-dc are the same... you will need to wire a pair of wire to each dc-dc and tour parallel group cell. but between serie cells, only one wire is ok.

Doc

 

[PJD]

In other words (excuse the sloppy sketch):

This seems entirely too simple to work! Why hasn't anyone else thought of this. Does each DC-DC converter have to be disconnected from the cells when not charging?

 

[recumbent]

I have 15 cells, would i be able to plug all these DC converters in one plug outlet? Or should we rig-up a 30 amp dryer plug.

Can someone show us a picture, or sketch of what this set-up would look like for those of us that don't daydream in electrical devices.

 

[Doctorbass]

In other words (excuse the sloppy sketch):

This seems entirely too simple to work! Why hasn't anyone else thought of this. Does each DC-DC converter have to be disconnected from the cells when not charging?

I would recommand to disconnect, but i dont know the leakage current on their output when they are off..

but putting a shotky diode in serie between the out + and the cell + would solve this issue. (in this case, the sensing + pin should be connected after the diode, (on the same side than the + cell)

I posted a basic schematic representing what i am building initially at 30A output and also a modded version with shotkey.

Doc

 

[docnjoj]

Sorry for the dumb question, but my 30 amp shotkeys have 3 pins, and I only see 2 connections?
David

 

[rkosiorek]

Sorry for the dumb question, but my 30 amp shotkeys have 3 pins, and I only see 2 connections?
David

be a lot less dumb if there was a part number to work with.

if you check the spec sheet for your diodes you will find that each package actually contains 2 diodes. that is why they call them DUAL diodes.

the speck sheet will also tell you if your diodes are COMMON CATHODE (the most common combination used), COMMON ANODE (RARELY used) or SERIES (rarer still) connected. that part of the specification defines how the diode pair is connected inside the case.

very often a mini schematic is printed beside the part number to show how they are connected. this would show 2 diode symbols (which look almost like arrow heads) and how they are oriented.

so if it shows the two arrows pointed towards each other and the common connection is in the center between the 2 arrow heads. it would shows that it is a common cathode pair, the 2 outside pins are individual anodes and the center is the common cathode.

in most cases you can safely connect both of the outside pins together and use it as a single diode. that is unless it is a series diode. need a part number to tell.

rick

 

[Doctorbass]

For great Dc-Dc deal go on ebay and search withdc-dc

Here is a link where my main supplier have alot of them.

http://search.ebay.com/search/searc...ass=goldenopportunities1&fsop=32&fsoo=2&fgtp=

The seller is: Goldenopportunity and all dc-dc are around 10$ each

Doc

 

[frodus]

While I do agree that this is the BULK of the solution, you are forgetting the big part.

Lifepo has a charge profile, I don't see where you've taken that into consideration. The voltage and current on the output should be trimmed.

This may be a good way to slam the batteries, but they just won't last, others have done this type of charging, and while it does balance cells, it does not charge them properly. You need something to trim back the voltage and current.

http://www.powerstream.com/LLLF.htm

Sure, you can "fast charge" but at what cost? You are decreasing the life of your cells!

What you need is a circuit to control the output. You shouldn't just set it at a voltage and let it go, it needs to change depending on the state of the battery. Also, I don't see anything that shuts off the DC-DC converter when one cell is done being charged. How do you stop from overcharging?

I see this as a good power side design, but you need to control it.

And the leakage current on these isn't very good, so disconnect, or use a FET on the output.

 

[frodus]

I have 15 cells, would i be able to plug all these DC converters in one plug outlet? Or should we rig-up a 30 amp dryer plug.

Can someone show us a picture, or sketch of what this set-up would look like for those of us that don't daydream in electrical devices.

you have 2 stages.... the AC-DC power stage (as stated, you need to buy a power supply that plugs into AC that has an output that matches your input of your DC-DC converters) and a DC stage (DC-DC converter input matches output of the AC-DC, and the output is 5V and can be trimmed to a set voltage.

But please remember, do NOT let this charger stay on... you will either need to control the GATE IN pin and run that to a timer, or have some sort of sensing circuit sense when the voltage reaches max, and disables the DC-DC converter, or you need to put the whole thing on a timer.

People have done this plenty, with NiMH, Lead Acid and Lifepo, its called Fast Charging. You're just splitting it up so that at least they're balanced. Lots of people use series fast chargers, for racing...etc. They also call them constant voltage chargers, or automatic chargers. They charge fast, then turn off automatically.... I would recomend you add that functionality so you don't lose cells.

 

[PJD]

How will the output voltage from each DC-DC behave during charging?

For example, my charger normally holds a constant current of 7 amps until the pack voltage reaches 59.5 volts, then taper the current so the voltage stays at 59.5 volts until the current tapers to 300 mA, indicating a full charge, then it switched to float which is 55 volts, allowing whatever trickle of current teh pack wil accept.

Now I hook this gang of Dc DC converters to this charger, with the DC-Dc converters are set to 3.7 volts. Since the maximum input power is 7 X 59.5 = 416 watts, what will the sixteen DC DC converters do to accomodate this limit to the power that have to work with? Will they lower the voltage or lower the current? For example, would the DC DC converters be putting out a constant 3.7 volts, but with a current "ceiling" of (416*efficency))/(16*3.7) =6.0 amp? But, they can't do this, because the charging impedence of a discharged pack would accept much more than 6 amps at 3.7 volts, so the DC-DC converters have to either lower the voltage and raise the amps to maintain the power output. How do thay work?

 

[Doctorbass]

While I do agree that this is the BULK of the solution, you are forgetting the big part.

Lifepo has a charge profile, I don't see where you've taken that into consideration. The voltage and current on the output should be trimmed.

This may be a good way to slam the batteries, but they just won't last, others have done this type of charging, and while it does balance cells, it does not charge them properly. You need something to trim back the voltage and current.

http://www.powerstream.com/LLLF.htm

Sure, you can "fast charge" but at what cost? You are decreasing the life of your cells!

What you need is a circuit to control the output. You shouldn't just set it at a voltage and let it go, it needs to change depending on the state of the battery. Also, I don't see anything that shuts off the DC-DC converter when one cell is done being charged. How do you stop from overcharging?

I see this as a good power side design, but you need to control it.

And the leakage current on these isn't very good, so disconnect, or use a FET on the output.

The charging profile of a li-ion and a li-ion LiFePO4 is CC-CV + a current disable fonction that stop charging when current reach a low level current threshold... ex: 2.3Ah A123 have 50mA threshold for stoping charging process.

cc-cv IS 2 stages... and the dc-dc have a very accurate voltage triming fontion that do the CV fontions nicely. for the CC, that's the internal current limit protection of the dc-dc that do the job.. I tested that and it work great!

I tested that on a 2.5V discharged A123. at 10A charge this cell voltage rise to around 2.9V... so a 10A dc-dc will drop to 2.9V it'S output because the current reached the limit so the current is now controlled and the voltage decreased.. that's the CC part.

I do not claim that this charger should be used to FAST charge cells... I know the li-ion cell caracteristics.. and i know what i'm talking about when talking about li-ion. I know that teh cycle life depend on many factor.. like load curent, charge current, uper and lower voltage limit... and the center of all those: the TEMPERATURE!..

LiFepo4 have a very low internal resistance compared to other li-ion cells.. so that would allow more current on load AND on charge.. A123 can accept 10A without minor deterioration on their cycle life. that's because they have around 10mohms

As a cell have a low internal resistor as it will dissipate less heat while current pass thru it..

I repeat again: that's not the current that dammage a cell.. that's the temperature.. and a factor that determine that is the internal resistance!

So the DC-DC i will use in my project are 30A one that i mooded to 15A by reverse ingeneering them and mooding the reference voltage resistor on the pwm current control chip.I know that this is not all people that would like to do that.. so that's the reason why i recommand to use different dc-dc with lower current.. like some 5A 25W dc-dc.

If you would read more post on that forum you will find that alot of people here have large battery pack close to 500Wh to 1kWh.. and that they often use A123 cell that are stacked like 20s 6p.. so 6 parallel A123 would be able to charge at max 60A.. and a 3p pack is able to hold 30A...

That DC-DC charger Is a great solution for them because no other charger can charge higher than 10A.. some use the voltfreak 2A charger, some the soneil 4-5A charger.. but they are limited to some preseted voltage... using DC-DC offer a flexible solution to ANY number of cells

I would agree that a great simple circuit using some shunt, comparator and voltage reference. like the LM35 or the TL431 that would cut the charge at the end of stage would complete this idea nicely.

Doc

 

[Doctorbass]

For example, would the DC DC converters be putting out a constant 3.7 volts, but with a current "ceiling" of (416*efficency))/(16*3.7) =6.0 amp? But, they can't do this, because the charging impedence of a discharged pack would accept much more than 6 amps at 3.7 volts, so the DC-DC converters have to either lower the voltage and raise the amps to maintain the power output. How do thay work?

Lithium -ion do not need a constant power mode to charge..

The dc-dc will hold the CV stage until the current limit protection is not reached.. so with the lower impedence of the cells on the beginning of charge, the current limit will be activate and the voltage will drop until the impedence rise at least to allow 10A at 3.7V (remember that LiFe would prefer 3.6V instead of 3.7V for their cycle life preservation!)

For sure if each dc-dc need let say 50W when their current limit is reached, and that the 3.6V is reached ( so that would be the max power they could give to your cells is reached) and that you have 16 of them, that would draw alot of power.. and your 59V would need 13.5A...

re-read what i explained above:
Buy a main power supply that can supply all the dc-dc input at the same time.
see how to calculate below:
let say 12x A123 cell = 43.2V
and that you charge at 5A mean you need that output power: 5A x 43.2V = 216W
DC-DC are efficient to around at least 80% so you will need to multiply by 1.25 that power for the input... that mean 270W of input total on all the combined dc-dc parallel input.
Generally DC-DC work with 36 to 72V on their input and have a regulated voltage and a current limit on the output and input. Some have 12V or 15V input but they are rare.
Ideally a 48V power supply do the job great.
Now for a 270W need, a 350W power supply should be of with a little safe margin.


for exemple for my 23Ah 89V (2x 12s18p) pack i will need the full 1100W of an AC outlet!

i will charge at 15A *12 x 3.6V or 4.2V depending on what pack i charge so that would be equal to 756W of charging power

with the efficiency combiend of the dc-dc and the main 48V power supply i will need around 1100W.

 

[frodus]

but you're not dynamically trimming the voltage during constant current.

Constant current and constant voltage "phases" don't occur at the same time. Constant voltage allows the current to do whatever the battery will accept. But when you're in current limit, what are you doing with your voltage? You should be tapering off.... which your design does not do.

I know your design work, I'm just saying it needs refining to be bulletproof. It will work, but it will decrease the life of your cells by pounding them (even at lower currents). What you're doing is just putting a fixed voltage on the battery, and letting it charge up to the current limit of the DC-DC converter, without tapering off voltage.

you need something like this:
http://cdn.vicorpower.com/documents/design_articles/pb_battery-charger.pdf

except for lifepo (they're available) but this schematic shows you what its doing.

 

[frodus]

The dc-dc will hold the CV stage until the current limit protection is not reached.. so with the lower impedence of the cells on the beginning of charge, the current limit will be activate and the voltage will drop until the impedence rise at least to allow 10A at 3.7V (remember that LiFe would prefer 3.6V instead of 3.7V for their cycle life preservation!)

[edited for clarity, appologies]

If yours are going into current cutoff, and folding back the voltage, it might not be good either, because you get hiccupps and spikes when the converter comes back out of limit. You can't say that for all converters they will act the same, because they don't. There are linear current limits, and foldback current limits.

I just don't want to see you get 1/2 the life out of the batteries because they're not being controlled right. I also don't want to see someone else get a diff converter that acts different and expect it to work like yours.

 

[PJD]

So, I need a power supply adequate to support the DC/DC converters at their current limit at the selected voltage. What happens if the power supply is undersized? Do the DC-DC converters simply shut off altogether?

 

[frodus]

So, I need a power supply adequate to support the DC/DC converters at their current limit at the selected voltage. What happens if the power supply is undersized? Do the DC-DC converters simply shut off altogether?

well, you need to know the input to output ratio..... if they're 48V-5V converters... thats about a 9.6:1 ratio @48V, so if they're 10A output at 5V (50W), they're roughly 1A on the input. You should really measure this for your converter. Its different if you use a 48V supply or a 60V supply (60V to 5V is 12:1). take into acount inefficiency, so 125% of that input, would be 1.25A.

i.e. If I've got 12 48-5V converters, and I want to size my 48V power supply right, and my converters are 15A current limited (75W max), that would be roughly 1.56A on the input. 125% of that is ~2A. Multiply that by 12 converters and you get 24A. So, size the power supply for 25A @ 48V. You could have a switch and enable only half the pack at a time, and get a 12-15A, and build a circuit to just use 6 at a time.

You don't want to overload the power supply with DC-DC converters, you might start to smell smoke, or the power supply will start hiccupping and go into current limit, you do NOT want this.

Also, fuse both the battery side of the DC and the wall side of the power supply.

 

[Doctorbass]

The dc-dc will hold the CV stage until the current limit protection is not reached.. so with the lower impedence of the cells on the beginning of charge, the current limit will be activate and the voltage will drop until the impedence rise at least to allow 10A at 3.7V (remember that LiFe would prefer 3.6V instead of 3.7V for their cycle life preservation!)

how is the DC-DC feedback loop decreasing the voltage setpoint by itself? If it goes into current limit, it should still regulate the voltage setpoint... if it doesn't, there's something wrong with the feedback loop in your converters. The vicor modules I have don't do this.... Current limit on the output just limits current, but voltage stays constant. If yours are going into constant current, and folding back the voltage, it might not be good either, because you get hiccupps and spikes when the converter comes back out of limit. You can't say that for all converters they will act the same, because they don't.

I just don't want to see you get 1/2 the life out of the batteries because they're not being controlled right. I also don't want to see someone else get a diff converter that acts different and expect it to work like yours.

Frodus, I know you want to help here and i see you read carefully that subject and this is great, the goal is to wirk on the same direction :wink:

What i dont agree is about the fact that you say that when the CC mode is active, that the voltage should stay the same.. :| :?

ok.. let's examin what you are saying: Ex: I have a sorenson power supply with constant voltage adjust and a constant current adjust. (like many lab power supply)

What will happen if i plug a non inductive, stable, precise......1ohm resistor load on the output... Then i turn the current limit fully clockwise and ajust the voltage to 5V...

you will agree that the current should and must be 5A right?... ok

so if i decrease the current limit to 4A, the voltage must drop to 4V right?.. because i think that if understand what you are saying correctly, the voltage should stay the same.. at 5V ... but it can't because to do that the load would need to change to keep that 5V... you see what i mean? correct me if i'm wrong...

The cell is a variable load.. or an equivalent variable resistor depending on the state of charge...(also, i'm not talking about the capacitive ir inductive part of that load.. only the dc) so i dont talk about inductive here.. i'm talking about charging with DC.. not pulse or frequency like battery desulfator...


So when you say that: "but you're not dynamically trimming the voltage during constant current."

I woud say yes, if the load is the same, the only way to enter into constant current mode , is to vary the current.. resistor law are applied here...

What do you think about that?

Doc

 

[docnjoj]

Thanks Rick K. for the Shotkey lesson! This charger really seems straightforward enough to give it a try!
David

 

[PJD]

So, you don't want to try this with a 48 volt lead acid battery charger as your power supply, correct?

 

[frodus]

Doctorbass, I Just reread my statements, and I'm sorry. I guess I'm confused a little about your converters and how they regulate the output.

I agree, V=IR. I think my statements seem to contradict, Not my intention.... but let me explain my converters. When mine go into overcurrent (meaning they're running over their rated current), they regulate the voltage up to 25% over the max current output (if its 10A, then 12.5A), They are in current limit, but still regulating voltage.

Not until current CUTOFF is the voltage changing. If you're running the converters in current cutoff, you're going to decrease efficiency, the converters will put off more heat, and you won't be getting the most out of them.

It can be in current limit, but still regulate. mine do. But once you hit that knee of 26% or more, voltage decreases, which I didn't understand you were doing. You're operating them up to cutoff.

http://cdn.vicorpower.com/documents/applications_manual/200VIJ00_Sect_4.pdf

again, I still think you need more smarts to control. I'll leave it at that.

 

[docnjoj]

So, you don't want to try this with a 48 volt lead acid battery charger as your power supply, correct?

If U meant me, I want to use my Astro112d as power supply. It does generate 72+ volts at 0 draw. I use a 24 amp, 12 volt power supply. Its gonna take some time to get up the nerve!
David (other Doc, not in electronics!)
oops, not me! Sorry!

 

[PJD]

David,

I actually meant the generic "you"; i.e. "So, a person should not use a 48 volt battery charger..."