The care and feeding of a123-based packs...

Batteries, Chargers, and Battery Management Systems.
Post Reply
User avatar
fechter   100 GW

100 GW
Posts: 14714
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Post by fechter » Nov 23 2007 11:18am

"One test is worth a thousand opinions"

User avatar
GGoodrum   100 MW

100 MW
Posts: 2908
Joined: Sep 20 2007 2:03pm
Location: South Orange County, CA
Contact:

Post by GGoodrum » Nov 24 2007 1:12pm

Wow, you are right, but I wish I would have known just to read the first page, and the last few posts. :) Lots of tangents in the middle. :?

Did you ever take this any further? I definitely want to do something like this for a123 and LiFeBatt-based packs, but wonder how simple this sort of circuit could be made. Physical size is important as well. Could something like this be done without using big transformers, caps or inductors?

With a123s, I would parallel 2-4 cells first, and the connect them in series strings from 12-20 cells. For LiFeBatt-based setups, I think the same 12-20 cell strings would be desirable. This would mean having anywhere from 11-19 of these circuits. That's a bunch more than what would be required for 3-5 SLAs in series, hence the need to make these as small and as least expensive as possible.

Nonetheless, I think we found the trail that leads to my elusive "holy grail" :D

--Gary

User avatar
fechter   100 GW

100 GW
Posts: 14714
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Post by fechter » Nov 25 2007 11:13am

Tangents on Endless Sphere? Never.

With 20s packs, the cost might get prohibitive. The basic topology should work though. At a high enough switching frequency, the inductor could be fairly small.

Here's a similar idea:

What if each cell had a voltage monitor that started up a switcher when the cell got up to voltage. The switcher would pump energy into the primary of a torroidal transformer, drawing current from the cell. The secondary would have enough turns to source current to the entire string (or sub-section) through a rectifier.

The transformer could have multiple primaries, one for each cell, but a single secondary. Again, at a high enough frequency (mhz?) the torroid could be physically small and not have many turns, keeping the cost down. This way, you could have 4,6 or maybe 10 cells run off a single transformer.

At high frequencies, you could get a lot of power through a small torroid.
Efficiency should be quite good if everything is designed carefully.

Any kind of transformer or inductor tends to be expensive in a circuit, so this may not be the cheapest solution. It may be cheaper than a bunch of single cell chargers though.

There might be a way to do a switched capacitor charge pump too. I haven't seen any of these that work at high power levels. I suppose there's a reason for this.
"One test is worth a thousand opinions"

User avatar
lawsonuw   1 kW

1 kW
Posts: 397
Joined: Oct 30 2007 12:04am
Location: Wisconson USA

Post by lawsonuw » Nov 25 2007 9:25pm

fechter wrote:There might be a way to do a switched capacitor charge pump too. I haven't seen any of these that work at high power levels. I suppose there's a reason for this.
I'm betting Inductors are preferred because they naturally limit current.

Anywho, back to the PowerCheq. In the patent they list on the product page they discuss prior art. (kind'a in the middle of the text) One method is exactly as you state Fetcher, another method omits the high voltage secondary and uses a primary coil per cell to shuffle power around. These methods are non-modular, but the second method would work just fine with extra channels.

I've seen some pretty big toroidal cores on Digikey. (~1.5in OD) Winding your own transformer using one of them shouldn't be *too* hard.

Marty

User avatar
lawsonuw   1 kW

1 kW
Posts: 397
Joined: Oct 30 2007 12:04am
Location: Wisconson USA

Post by lawsonuw » Nov 25 2007 9:41pm

Snickers wrote: At the end of charging, when, BMS is starting to equilibrate cells. There is “smallâ€￾ current oscillation in both direction between BMS and cells already full.
Silicium would like to cancel these oscillations. He has tested with one inductance between cells and BMS, but it’s become more critical!
With a bigger cable between BMS and cells, this phenomena is decrease.
He thinks…we must introduce a filtering (R,C) somewhere…!
How to solve this problem?
I've re-posted silicium's .pdf schematic and I'm using the circuit surrounding CELL0 for all my component names:

Ok, U1 plus Q3 and Q1 have a *VERY* high gain while they are active, so it's not surprising that they oscillate. C1 provides damping to the circuit. try boosting it in 10x steps. (Q3 holds the voltage of it's gate relatively stable when it's on, so C1 may need to get really large) R3 + RV1 + R2 also effect the damping, upping these values by 10x may help. Total gain can also be reduced by adding a resistor between U1's output and Q3's gate. I'd personally start at one ohm and work up from there. Swapping Q3 for the appropriate low voltage FET may also help. This helps because a FET would allow it's gate voltage to change more, thus allowing C1 to have a greater effect. Whew! that's all I can think of now.

Marty
Attachments
BMS A123 x12 eng.pdf
Circuit schematic mirrored here
(99.54 KiB) Downloaded 207 times

User avatar
GGoodrum   100 MW

100 MW
Posts: 2908
Joined: Sep 20 2007 2:03pm
Location: South Orange County, CA
Contact:

Post by GGoodrum » Nov 25 2007 10:34pm

You are right, 20s becomes unwieldly. I do the packs as 10s, and then use two in series.

So, which approach would be simpler, switching caps, or the inductors?

-- Gary

User avatar
rsisson   100 W

100 W
Posts: 124
Joined: Oct 18 2007 11:43am

Post by rsisson » Nov 26 2007 5:10pm

Wow what a thread...

Just finished reading the whole thing....

My issue is that I want a HIGH POWER system. The pack I am looking at is 48V 40Ahr. Even at .5C the charging current is 20A..ouch.

I tend to look at things from the INSIDE OUT...

If we want to limit an idividual cell to 3.7 volts, put a zener accross it. A BIG zener. We can pick the Voltage, 3.7 to 4.1 and that is the voltage...done...

If we stack a whole lot of "Zenered" Batteries to make the pack, at some point ALL of the zeners will be conducting...and the pack will be ballanced and the zeners will be smoking...

Ok..how about a 2-stage zener interface
at 3.7 volts Zener #1 does one function (lights a light)
at 4.1 volts zener #2 does something more dramatic (clamps the cell)

A taper charger could be set so that at (no of cells) x 4.1v the current is minimal.. ie a constant VOLTAGE charger

at anything LESS than Max charge, the charger will be trying to put its maximum output into the slacking cells, the charged cells will already be clamped by the Zeners... At cells x 4.1 you could force the current to a safe "float" value

Now this idea has a few drawbacks...
+ No current limitations In or Out
+A truely dead cell could sink a LOT of power and do nasty things
+No External indication of what is going on...
+Charger must be smarter
+Expensive, a 20A Transistor may be $5+
And a few Plus's
+No current limits in or out (Zener or Tip35 limit)
+Can be mounted on Large (prismatic) cells directly
+No external wires neded
+Very limited part count at the battery

If Zeners are not available in high enough ratings, a simple Tip35 Transistor, a resistor and a Low power zener will do the same (with the zener voltage adjusted for the Vbe drop)

If you add things from the INSIDE-Out, the circuits and design stay simpler...

Zeners (on digikey) go 3.0, 3.1,3.3, 3.6, 3.7, 3.9, 4.3)
3.9v + .2v Vbe for a power Transistor would be near perfect

Am I missing something???

User avatar
Snickers   10 mW

10 mW
Posts: 21
Joined: Nov 22 2007 2:45pm

Post by Snickers » Nov 26 2007 6:39pm

lawsonuw wrote: Ok, U1 plus Q3 and Q1 have a *VERY* high gain while they are active, so it's not surprising that they oscillate. C1 provides damping to the circuit. try boosting it in 10x steps. (Q3 holds the voltage of it's gate relatively stable when it's on, so C1 may need to get really large) R3 + RV1 + R2 also effect the damping, upping these values by 10x may help. Total gain can also be reduced by adding a resistor between U1's output and Q3's gate. I'd personally start at one ohm and work up from there. Swapping Q3 for the appropriate low voltage FET may also help. This helps because a FET would allow it's gate voltage to change more, thus allowing C1 to have a greater effect. Whew! that's all I can think of now.
Marty
Thanks for your help Marty
Silicium will exchange the capacitor C1 just to see... :roll: He have allredy made some differents test with several transitor gain. But, it doesn't really help...
He will focus on activation of 500 mA current limit when the first cells is full. There is currently no delay to switch "on" the current limitation. It produces a short time current falling. A ramp arround 10-100 ms to decrease current should be perhaps helpfull in order to temperate current transition.
The problem could also come from from power supply oscillation.
I remain, these oscillations are not critical, but it wil be better if he could cancelled them...

@rsisson, Wouhaou You will lose full power from the charger inside zener diodes or transitors when the batterie will be full 20x48=~1 kw !

User avatar
rsisson   100 W

100 W
Posts: 124
Joined: Oct 18 2007 11:43am

Post by rsisson » Nov 26 2007 7:44pm

I would have (partially) agreed with you until I ready the Metric Mind site (worth reading if you haven't' yet) and his Thunder-Sky top "Vclamps" which are almost identical to what I proposed, and he built and is using them.

He has 96 cells "Clamped" to 4.3v each feeding his Civic... Yup that's a lot of Power... 100+ Mile range WOW...

http://www.metricmind.com/ac_honda/main2.htm is a good place to start...

His charger starts out Dumping Massive amounts of Current (8.2A's) into his 90Ahr pack, and then switches over to Constant Voltage and then the current trickles down to ZERO BEFORE all of Zener's turn on... and the way he designed it, even if ALL of the zeners were turned on, his CLAMP can take 18A's

In my Case, if I used a charger that only put out 81.7V MAX,(19 cells) the current when ALL of the Zeners were on would be (should be) ZERO... now it never will be zero, as some cell will always be a tiny bit low, so it will get the difference between 4.3 and its current voltage up to the limit of the charger... all of the other batteries will be clamped at 4.3v. A good charger will not try to pump many kilowatts into a cell that is .1v off...and most that I know couldn't do it if they tried...

Now this application really only works on HIGH POWER cells, as the Clamp isn't cheap, but I am sure it can be made to work for smaller applications, such as the A123's, but there is a fair difference between those and the Thunder-Sky's.

From his BLOG...

At this point, it looks like slowing down the charger is not even needed - 9.5A on charge gets distributed between full cells and clampers such that the heat sinks aren't even warm. .....To check functionality of all the clampers at once I've created special test charge profile - 3A charging current and no voltage limit (practically I set it to 450V, way above sum of all the clamper voltages. If the pack is fully charged, this profile causes all the clampers to go into regulation mode since they are the only devices setting voltage limit per cell, so all LEDs should lit. Another benefit - I can adjust the clamping voltage at the current the charger provides at the end of charge - I set 4.25V @ 3A. Actually IV curve is so sharp, that it was not really necessary: at 4.23V there is no regulation, and at 4.27V there is full blast 10A bypass - more than my charger outputs.

User avatar
GGoodrum   100 MW

100 MW
Posts: 2908
Joined: Sep 20 2007 2:03pm
Location: South Orange County, CA
Contact:

Post by GGoodrum » Nov 26 2007 8:13pm

rsisson wrote:Am I missing something???
I'm not sure, but at a minimum I'm guessing a big honkin' heat sink. :o

I love simple, but if it truely was this easy, why hasn't it been done already?

Until this all gets sorted out, I'm going to be using my "interim" solution, which is my 10 individual 2A VoltPhreaks chargers to do the final top-off/balancing, and a surplus Tyco telcom power unit that will deliver a steady 17A from 12-53V, to do the heavy lifting. It gets cutoff as soon as the first cell block hits 3.7V.

I love your PVC car, BTW. I read your thread on it. I'm thinking of doing the same but I want to use two 5-series motors and run them on an 80V setup (24s a123s or LiFeBatts...).

-- Gary

User avatar
rsisson   100 W

100 W
Posts: 124
Joined: Oct 18 2007 11:43am

Post by rsisson » Nov 26 2007 8:54pm

Thanks for the feedback on the car...

Next iteration will have two motors..need the power for two people and Groceries... that's why I'm looking at at BIG pack...

My point is that it HAS been done, by Victor over at Metric Mind... His 400V 90Ahr 36KILOWATT setup puts us all to shame... Think about that, 36 KILOWATTS for an HOUR... OUCH!

As for heat sinks, this scheme uses lots of small ones vs one big one. He had 96 heat sinks and they didn't get warm... but again, as all of the clamps turn on, the CURRENT should be tapering off to near Zero... yea if you crank up the voltage AND current, you might be dissipating a lot, but spead over all of the clamps, not just one...

He was kind enough to make public his BMS clamp plans, and they are easy enough to replicate... they even look cool...

Don't know how I would change things for an A123 pack... it wouldn't need to be quite so robust as the current limits are lower, might be able to make the entire thing on one board and make it outboard for medium sized packs...

Take a look at his BMS Blog at
http://www.metricmind.com/ac_honda/bms.htm

It is impresive as hell....

User avatar
GGoodrum   100 MW

100 MW
Posts: 2908
Joined: Sep 20 2007 2:03pm
Location: South Orange County, CA
Contact:

Post by GGoodrum » Nov 26 2007 10:30pm

rsisson wrote:Thanks for the feedback on the car...

Next iteration will have two motors..need the power for two people and Groceries... that's why I'm looking at at BIG pack...

My point is that it HAS been done, by Victor over at Metric Mind... His 400V 90Ahr 36KILOWATT setup puts us all to shame... Think about that, 36 KILOWATTS for an HOUR... OUCH!

As for heat sinks, this scheme uses lots of small ones vs one big one. He had 96 heat sinks and they didn't get warm... but again, as all of the clamps turn on, the CURRENT should be tapering off to near Zero... yea if you crank up the voltage AND current, you might be dissipating a lot, but spead over all of the clamps, not just one...

He was kind enough to make public his BMS clamp plans, and they are easy enough to replicate... they even look cool...

Don't know how I would change things for an A123 pack... it wouldn't need to be quite so robust as the current limits are lower, might be able to make the entire thing on one board and make it outboard for medium sized packs...

Take a look at his BMS Blog at
http://www.metricmind.com/ac_honda/bms.htm

It is impresive as hell....
Yes, I agree, most impressive. I got lost in his site for several hours last week. There is a thread here somewhere that referenced that BMS design, which is how I found out about what Victor did with that beast of a Honda. I'd love to do something like that, but I don't have all the skill sets he does to tackle that without a lot of help/

Baby steps... :wink:

For your car app, you might also look at LiFeBatt cells. These are new, but they are also LiFePO4-based. the big advantage is that the cells are 10Ah, or about 4 times as big as the a123s. These will simplify doing bigger packs, such as what you are talking about. They seem to be a lot more robust than the current crop of Chinese LiFe imports, and are manufactured in Taiwan, instead of the mainland. They are the only ones offering any sort of warranty for raw cells.

One problem with using two motors in the back, is that we need some sort of electronic differential, to handle turns. To start, I'm going to try using two throttles, one on each grip.

-- Gary

User avatar
rsisson   100 W

100 W
Posts: 124
Joined: Oct 18 2007 11:43am

Post by rsisson » Nov 26 2007 11:16pm

Just a quickie on two motors...

Supposedly 2 crystalyte controllers can SHARE a throttle...

Minor differnces in turns should not be a problem so long as everything ELSE is matched... ie motors, tires, pressure, etc...

User avatar
rsisson   100 W

100 W
Posts: 124
Joined: Oct 18 2007 11:43am

Post by rsisson » Nov 26 2007 11:23pm

The problem is I KNOW I need at least 20Ahrs... 13Ahrs of SLA barely gets me 5 miles (with my lead foot/hand)

13A - 5 miles
20A -8 miles or more because they are lighter
40A -16 miles with some leftover for extra weight perhaps...

Yea Yea, I know I can't really such 13Ahrs out of the 13Ahr SLA's at heavy load, but it gets me a place to start... and gives me a "magnitude" to play with...

rf   1 kW

1 kW
Posts: 381
Joined: Sep 18 2007 8:02pm

single cell charger?

Post by rf » Nov 27 2007 7:54pm

So what's an easy way to build a cheap, small, light, efficient, 10 to 20 amp charger for a single A123 cell? Can it be done for ten or fifteen bucks?

Perhaps we could put something simple together, then refine the hell out of it. :-)


Richard

User avatar
GGoodrum   100 MW

100 MW
Posts: 2908
Joined: Sep 20 2007 2:03pm
Location: South Orange County, CA
Contact:

Re: single cell charger?

Post by GGoodrum » Nov 28 2007 1:23pm

rf wrote:So what's an easy way to build a cheap, small, light, efficient, 10 to 20 amp charger for a single A123 cell? Can it be done for ten or fifteen bucks?

Perhaps we could put something simple together, then refine the hell out of it. :-)


Richard
Bob Mcree found one here, but you would still need some way to limit the voltage to about 3.7V, and let the current taper off (constant voltage mode...).

In my mind, it is still pretty hard to beat the 2A CC/CV VoltPhreaks chargers for about $10 a piece. Later this week I'm going to be testing a circuit based on the LVC design, that will control a surplus Tyco 17A constant current supply used by the phone companies, to "boost" the first part of the charge cycle up to a total of 19A. Once any one of the cells first hits 3.7V, the circuit cuts off the boost supply, and the 2A chargers finish up topping off the cells. Here's what the board layout looks like:

Image

Anyway, I still believe that individual cell charging is the best way to make sure each cell/cell block is fully charged, but unless the chargers are built into the pack, you still need to bring out the individual wires. The "bucket brigade" balancers, like the PowerCheq widget, are a clever idea but the problem I have with them is that all the cells end up at the level of the weakest cell/block. It has been my experience that after some use, all the cells have variable internal resistances and slight capacity differences. What I do is try and at least group the ones with similar "resting" voltages, after charging, together in parallel blocks. These blocks all seem to end up at different "full" voltage levels. With individual cell/block charging it doesn't really matter if they are different, just so long as they are as full as they are going to get.

I'm still not sure exactly how these shunt regulator type BMS designs handle charging. I get that they use a FET to bypass the cell when the cutoff is first hit, but if it stopped there the cell would only be at about the 80-85% charge level. I'm guessing the cell voltage would drop as soon as the cutoff is hit, but how far does it need to go for the FET to turn off? In any case, it sounds like the FET will oscillate on and off, until the cell voltage can stay above the voltage that would cause the FET to cutoff, which I guess says the cell is fully charged. Is this the way this is supposed to work? :?

-- Gary

User avatar
GGoodrum   100 MW

100 MW
Posts: 2908
Joined: Sep 20 2007 2:03pm
Location: South Orange County, CA
Contact:

Post by GGoodrum » Nov 28 2007 2:38pm

rsisson wrote:The problem is I KNOW I need at least 20Ahrs... 13Ahrs of SLA barely gets me 5 miles (with my lead foot/hand)

13A - 5 miles
20A -8 miles or more because they are lighter
40A -16 miles with some leftover for extra weight perhaps...

Yea Yea, I know I can't really such 13Ahrs out of the 13Ahr SLA's at heavy load, but it gets me a place to start... and gives me a "magnitude" to play with...
I think the reason you are getting such low "mileage" is because with smaller capacity SLAs, you are getting some pretty excessive voltage drop under load. That is causing you to "burn" Ah at a higher rate than it would if you used higher capacity SLAs. Switching to any Lithium-based chemistry, in the same capacity range, will provide significant range improvements over what you are getting now, but a123s will have the most improvement, due to their extremely high "C" rating. Each 2.3Ah cell has the ability to be discharged at up to 70A continuous, and a whoping 120A for 10-second "bursts". :shock: That works out to a continuous C rating of about 30C and a burst rating of 52C. This is for a single string of 2.3Ah cells. put four of these in parallel and those numbers for up to 280A continuous and 480A burst. I think most SLAs have a proble delivering current at even a 1C rate, which is why you are seeing such a big voltage drop.

Since even for the worst case, you might pull 60-70A, which even a single string of a123 cells could handle, the voltage drop is not very much. I've hit 5500W/87A peaks on my folding bike, and that is with a 20s4p setup. The way you really notice how strong the a123-based packs are, over pretty much everything else, is that the power feels consistant all the way up to the end of the capacity. With well-balanced cells, it will feel just as strong 10 seconds from the end as it does in the beginning. This actually can cause problems.

Although I've found these to be quite indestructable, form a current delivery and charging point-of-view, the Achilles Heel for these cells is over-discharging. I've abused 100s of cells in RC aircraft over the last year and a half, I've never killed any. The main reason is that once the cells start to dump, you are done flying. Once the power drops, we cut the throttle which removes the load. The resting voltage of the cells is still well above 3.0V per cell, but you won't be able to get it off the ground. With bikes, however, the loads are less, plus it is spread over multiple cells in parallel. It is quite easy to run these down to the point the cells become damaged. The reason it is so easy is because you literally get no warning becasue the power feels the same as it did at the beginning. I managed to kill a few cells this way, until I figured out what ws going on. That's why we did the LVC circuit board, using Bob's simple, but most effective design. As soon as any one cell block drops below 2.7V, the brake inhibit line to the controller is tripped, which cuts off the load.

One other point about a123-based packs is that pretty much all of the capacity is usable. With a 9.2Ah/4p setup of healthy a123 cells, and you run it down to where the LVC trips, you will have used pretty close to exactly 9.2Ah, if not a bit more. With SLAs in these kind of applications, I've heard varying amounts around 60% of the rated capacity is all you get.

-- Gary

User avatar
OLIVIER   100 mW

100 mW
Posts: 35
Joined: Nov 20 2007 11:41pm
Location: Montreal
Contact:

Post by OLIVIER » Nov 28 2007 4:32pm

......effective design. As soon as any one cell block drops below 2.7V, the brake inhibit line to the controller is tripped, which cuts off the load.


-- Gary[/quote]

If/when you have a weak cell in a pack, is it always the same one? Can another cell become weaker?
If you find the weak cell and control the discharge of the pack by the voltage of this one single cell, it would be a simple system. Each pack could have its "monitoring cell".
OLIVIER

User avatar
GGoodrum   100 MW

100 MW
Posts: 2908
Joined: Sep 20 2007 2:03pm
Location: South Orange County, CA
Contact:

Post by GGoodrum » Nov 28 2007 5:27pm

The LVC parts are only about a buck per channel, so you won't save much, even if you do end up with a weak group of paralleled cells. Besides that, If I had a block that was consistantly weaker, I would think about replacing that block. You can tell the weak blocks if you are using the multiple 2A CC/CV chargers like I'm doing, because the weak blocks tak a lot longer for the LED to turn from red to green.

With healthy cells, this isn't so much of an issue. They are all going to be close to the same level. The reason I picked 2.7V is that when the cells hit that point under a 40A load, they will bounce back to a bit over 3.0V once the load is removed. What happens is that you feel the controller cut the power, and then after a second or two, it will come back. If the demand is still there it will cutoff again. If you back off the load it should still give you enough for a short "get home" capability. If, on the otherhand, the load is light, say 5-10A, if the voltage drops so much that the resting voltage is at 2.7V, or close to it, the cells are protected from damage at 2.7V. They are probably okay down to less than 2.5V, but this gives a bit more reserve. At about 3.0V the cells will take a full 2.3Ah back in, so I don't see the need/point of trying to go lower.

The other reason for a relatively high cutoff is that with well-balanced packs, the cells will be strong all the way up to the end, and then they will all pretty much dump at the same time. This can be quite disconcerning, as it feels like something blew, or you hit a kill switch. What you want the LVC to do is to detect when that drop is first starting, and then cut the power.

-- Gary

User avatar
fechter   100 GW

100 GW
Posts: 14714
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: single cell charger?

Post by fechter » Nov 29 2007 9:32am

rf wrote:So what's an easy way to build a cheap, small, light, efficient, 10 to 20 amp charger for a single A123 cell? Can it be done for ten or fifteen bucks?

Perhaps we could put something simple together, then refine the hell out of it. :-)


Richard
Here's one:

http://www.allelectronics.com/cgi-bin/i ... PPLY_.html
Image
Lambda # SCQ12003054848. Input: 100-240 Vac. Outputs: 3.3 Vdc @ 7 A, 5 Vdc @ 3 Amp, 48 Vdc @ 1Amp,-48 Vdc @ 1 Amp. (Max output w/30 CFM forced-air cooling). Open-frame switching power supply, 7" x 4" x 1.73". Molex-type headers for input and output. UL, CSA.
CAT# PS-337

Your Price: $5.00 each

Special Quantity Discounts!
20 to 99 $ 4.50 each
100 or more $ 4.00 each

In stock, ships within 24-48 hours.


OR maybe this one:

http://www.herbach.com/Merchant2/mercha ... y_Code=PWS
Image
+5 Volts 20 AMPS Other 12 Volt Low Current Outputs Input 115/230 VAC, 50/60 Hz (user selectable) Outputs have common returns. Connections via molex style plugs. Unshielded. Size 10-1/2 x 6 x 2-1/2". UL Listed, CSA Certified. Code: TM89PWS5642
Price: $9.95

Shipping Weight: 5.00 pounds

Of course both of these would require some kind of tweaking to get the voltage at the right level, but I think it wouldn't be too hard.
Both of these places had other models. Some had cases.
"One test is worth a thousand opinions"

User avatar
silicium   10 W

10 W
Posts: 67
Joined: Nov 12 2007 2:38pm
Location: Albi, France
Contact:

Post by silicium » Nov 29 2007 10:24am

A new version on my BMS chargeur balancer is ready.
Features:
Automatic start (supress switch)
Current charge up to 10A (8A for me)
Automatic balancing at 0.5A (current limit for low heater)
Automatic stop after balancing (all cell balanced)
LED's status (charge, balancing, stop)
Ready for 12 a 15 cells
Only 2 wires for connect power to BMS
Small power dissipation into BMS (max 1.8W per cell on balancing)
Require only one external power voltage source with current limit
components (for 15 cells):
156 resistors
15 variables resistors
2 resistors pack
37 capacitors
51 Integrated Circuits
66 Transistors
51 Diodes
Total: 378 components :lol:
a lego for Christmas !
Schematic only, board for future 8)
Attachments
BMS A123 x15sr.pdf
(120.81 KiB) Downloaded 339 times
RH205, A123 4P12S with BMS http://vae-tech.forumactif.org/t8-un-bms-home-made (in french)

User avatar
Jozzer   100 kW

100 kW
Posts: 1069
Joined: Apr 19 2007 8:32am
Location: Swindon UK
Contact:

Post by Jozzer » Nov 29 2007 11:26am

Excellent, I know what I'll be doing over chrismas :lol:

thanks Silicium.
Mazda MX-5. 300KW power. Soliton 1 controller, 11" Kostov motor, 20KW/H Turnigy Lipo for 60-100 miles range. 120mph top speed.
Hudson Kindred Spirit 3 wheeler. Twin Agni setup, 300KG 80KW. 100mph top speed (maybe more, but no-one has the guts to try!)
Aprilia RS125, Agni motor, 600A 96v Kelly controller, 6kw/h pack
Dual Agni Ducati.
Countless ebikes and trikes..
http://www.Jozztek.com

User avatar
Doctorbass   100 GW

100 GW
Posts: 7437
Joined: Apr 08 2007 2:24am
Location: Quebec, Canada East
Contact:

Post by Doctorbass » Dec 02 2007 1:29am

Great work Silicium !

We will wait for the board !


Here are some great advices i can give for the new A123 battery pack builder: :wink:

The great advices i can give you are:

These cells are sensitive to unbalance so be carefull about that. If you can, just before to built your pack measure the capacity of each cells with a RC charger that have this option or the CBA-II you would ensure you to maximyze the succes and performance of it. The CBA-II cost 100$ and will save you many trouble dur to bad cells. With that you will be able to built each parallel group with a very close capacity to each other.

If you can't have a BMS charge them using 1 seperated 3.6V charger per parallel group. This way, your pack will be always perfectly balanced during charging and will help to keep a balanced discharge too. I suggest you to check for some great post i did in "the single cell cahrger post" you can find many cheap and powerfull 3.3V adjustable to 3.6V switching power supply.

When building the pack , always keep in mind that the temperature of each cell will influence each of their internal resistor, so try to place the cells so that they will have all the same temperature. I would not recommand some cubic pack because the center cells will always have a different temperature than the surounding cells. Many parallel flat bar group of cell (1 row) will help. You can wrap with thermal shring eack of these group to help to uniformise the temp inside. By this way, all cells of parallel group will have all the same discharge current at the same time.

Also, with heavier pack allways link each independent parallel pack using cable and not some rigid flat bar ot tab. This will avoid vibration to broke the serie connections.

When the pack pound at least 25lbs+ try to place it the lowest possible place, to help the center of gravity.

Doc
CURRENT PROJECT: 2WD duo MXUS/Max-E ebike 32kW
http://endless-sphere.com/forums/viewto ... =3&t=65764
-Fastest speed record 117 km/h on flat
-Fastest 1/4 mile@ 114km/h on flat and 16.316 sec
-Succeded to haul a 19200 pounds schoolbus!
113kmh Gianthttp://www.evalbum.com/3406
Mongoosehttp://www.evalbum.com/1947
YOUTUBE---https://m.youtube.com/user/Doctorbasss
http://twitter.com/DocbassMelancon
I speak FRENCH and english

User avatar
Doctorbass   100 GW

100 GW
Posts: 7437
Joined: Apr 08 2007 2:24am
Location: Quebec, Canada East
Contact:

Re: single cell charger?

Post by Doctorbass » Dec 02 2007 2:07am

GGoodrum wrote:
rf wrote:So what's an easy way to build a cheap, small, light, efficient, 10 to 20 amp charger for a single A123 cell? Can it be done for ten or fifteen bucks?

Perhaps we could put something simple together, then refine the hell out of it. :-)


Richard
Bob Mcree found one here, but you would still need some way to limit the voltage to about 3.7V, and let the current taper off (constant voltage mode...).



-- Gary

About the Dc-Dc converter at ebay, I ordered one for tests

they have the V adjsutement!!

see the PDF here:http://www.sierraic.info/05DATASHEETS/bxb150s.pdf
CURRENT PROJECT: 2WD duo MXUS/Max-E ebike 32kW
http://endless-sphere.com/forums/viewto ... =3&t=65764
-Fastest speed record 117 km/h on flat
-Fastest 1/4 mile@ 114km/h on flat and 16.316 sec
-Succeded to haul a 19200 pounds schoolbus!
113kmh Gianthttp://www.evalbum.com/3406
Mongoosehttp://www.evalbum.com/1947
YOUTUBE---https://m.youtube.com/user/Doctorbasss
http://twitter.com/DocbassMelancon
I speak FRENCH and english

User avatar
EMF   100 kW

100 kW
Posts: 1342
Joined: Sep 30 2007 4:07pm
Location: Curious about The Bikers Bar? Become a member to visit!

Post by EMF » Dec 02 2007 2:25pm

Gary Goodrum, would you mind specifying the size and source of that rubberized shrink material you use on your 4 packs?

Regards,

Dan

Post Reply