Balance Charger Design Idea
Beau Schwabe
Posts: 6,566
I just wanted to get some feedback on this design and what you might think.
In particular I'm interested in any feedback from the cascaded high-side control. Works nice on the simulator, but I haven't had a chance to 'plug it in' just yet perhaps next weekend.
Ultimately I would like to expand this idea to other battery chemistries as a "Universal Chemistry Battery Charger".
In particular I'm interested in any feedback from the cascaded high-side control. Works nice on the simulator, but I haven't had a chance to 'plug it in' just yet perhaps next weekend.
Ultimately I would like to expand this idea to other battery chemistries as a "Universal Chemistry Battery Charger".
933 x 773 - 143K
Comments
1) It appears that the NPN xistors will be saturated; what's Vce for those guys?
2) What's the Li-Poly voltage?
3) Can you elaborate on what you expect the ouptut structure to do?
Regards,
DJ
1) I Assume you mean the NPN's on the Li-Poly side -- Those NPN's will saturate when P0,P1,P2, or P3 are LOW -- They are meant to be PWM'd or pulsed very briefly and act as a 'shunt' to prevent the Li-Poly cell from gaining a higher voltage over the other cells. The Bottom 3 NPN's would have a small resistor to prevent a direct short across the Li-Poly Cell. 2) The Li-Poly voltage is 3.7V per cell (4.4V under a charge) 3) - The charging is done through the two 10 Ohm resistors and the 'Top' most NPN drive transistor on the far left. - The remaining NPN drive transitors are normally 'OFF', and are there to 'throttle' a cell from a voltage imbalance with reference to the other cells. - The PNP's provide a control path to the drive NPN's across their B-C junction - The NPN's between the Propeller and the PNP form a cascade of high-side drivers where the High-Side voltage appears at every PNP Emitter stage. When P0,P1,P2 or P3 are LOW, the PNP turns 'ON' creating a darlington effect with the coresponding NPN drive transistor.How hard would it be to expand this to at least 6S? Also, can it "autosense" to work with whatever size LiPo you hook up?
I am going to watch this post to see how you do this
I want to something like this myself
"...it looks like this will only work with up to 3S LiPo packs" - Yes, that is correct. 3-Cell packs are the limit with the provided circuit.
"...but 4S to 6S packs are common sizes" - for most of what I would use, 11.1V (3-Cell) would be plenty, I suppose 14.8V (4-Cell) and 22.2 (6-Cell) would be useful also.
"How hard would it be to expand this to at least 6S?" - Not very hard at all, the circuit is step-and-repeatable to some degree, The 10K control resistor to the PNP might need to increase, and you would need to adjust the resistor divider to the ADC, but it should be fine with a MCP3208 instead of the MCP3204 provided in the schematic.
"...Also, can it "autosense" to work with whatever size LiPo you hook up?" - A simple voltage check from the ADC across the cell would determine the presence or not of a cell. In the case of an absent cell, the 'SHUNT' transistor could be energized to complete the charging circuit.
Loopy Byteloose,
Yes, the pre-charge phase would be PWM controlled, and yes, I agree, that some sort of an 'OR'ed temperature sensing could be managed.
The mention of thermally sensing the cells makes a lot of sense.
Regards,
DJ
Why monitor? Because something like this?
http://forums.parallax.com/showthread.php?120396-Never-leave-a-plugged-in-battery-charger-unattended!&highlight=Charger
http://www.youtube.com/watch?v=z3o_2mwRPdw
Jim
Um, yes.
DJ
...and yes, I did see Phil Pilgrim's post and his dealings with batteries. I have had similar problems with NiMh leaking while under charge, resulting in a run-away charger failure. Fortunately I caught it in time because the smell wasn't right.
Secondly,
I have had one very negative experience with Li-Po. I purchased a spare battery for an R/C airplane that was about $50USD. A few weeks later I put it on the recharger that came with the airplane kit and it swelled up like a pregnant guppy. It didn't explode, but one wonders how close it was to doing so.
Since then I have had simlar swellings with Lead Gel cells that I have put on a charger than wasn't really matched to their capacities. Now I am building a charger than has multiple current choises and varies the voltage out. But it also includes a thermal sensor that shuts the charge cycle completely down until the battery cools off whenever overheating occurs.
One can simply use a themistor, a BasicStamp and a solidstate relay with any charger to set a thermal shutdown. So it seems so silly that nothing is out there on the market with this safety feature. I am completely fed up with buying sealed rechargible batteries and then prematurely destroying them with inappropriate chargers. Having put a lot of thought into the problem, heat remains the best indicator of abusive charging - after all, this is more about chemistry than about electronics.
The balance charger is something that I have often wondered about, but couldn't seem to come up with a wiring plan that might work. If I had a good charger, I'd be more trusting of using these rather big and expensive battery packs.
Beau Schwabe (Parallax);960909]Roy Eltham,
"How hard would it be to expand this to at least 6S?" - Not very hard at all, the circuit is step-and-repeatable to some degree, The 10K control resistor to the PNP might need to increase, and you would need to adjust the resistor divider to the ADC, but it should be fine with a MCP3208 instead of the MCP3204 provided in the schematic.
"...Also, can it "autosense" to work with whatever size LiPo you hook up?" - A simple voltage check from the ADC across the cell would determine the presence or not of a cell. In the case of an absent cell, the 'SHUNT' transistor could be energized to complete the charging circuit.
The series structure with the NPNs in parallel with the cells--hmmm. It definitely needs resistors to prevent direct shorts, and no accidents with the algorithm! The resistors should probably be in series with the batteries, not the transistors, no?
The algorithm? Top switches all have to be ON in order to deliver current to the bottom cell of the totem pole. So supposing the bottom cell is the weakest, the top two cells have to be discharged while the bottom cell is charged up to what did you have in mind? Maybe to equal the weakest among the top two. Subsequently the weakest among the top two is charged until the top two are equal. Then back to the bottom one to charge it, and at the same time the top two have to be discharged. Now they are all equal. Great! Where from there? Move up or down the totem pole Starting at the bottom seems like a losing proposition. Every increment of charge delivered to the bottom one is at the same time extracted from the top two. Then the algorithm has to go up the totem pole and replace that. Same thing with starting at the top. Every increment of charge put in at the top is removed when the algorithm moves down the totem pole. I'm getting dizzy!! What am I missing, Beau?
I updated the original schematic at the top of this thread and attached to this one.
There is no special algorithm based on the position of the cell.
What are the part # for Q 1 to Q 10
Thanks
MJE150032G Datasheet Reference:
http://www.fe.infn.it/~barion/docs/MJE15032G.pdf
The STX690A would be a better solution than the MJE150032G:
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00225079.pdf
Also, the ZTX1048A would be a better solution than the MJE150032G:
http://www.diodes.com/datasheets/ZTX1048A.pdf
Q3,Q6,Q9, and Q12 could be generic 2n3904's. 2mA to 3mA maximum current requirement.
Q2,Q5,Q8, and Q11 could also be generic 2n3906's. 20mA maximum current requirement.
The PNP/NPN composite is known as the Sziklai pair connection. (complementary to the Darlington pair with NPN/NPN or PNP/PNP).
I still have few concerns about the circuit.
Maybe the balancing PWM can be driven through capacitors from the Prop, so that by no programming "accident" or failure can the batteries be left short circuited.
The sense resistors put a constant load on the batteries even when they are sitting idle. The worst thing is to return to a project that was left on the shelf and find the batteries dead! It is only about 0.6mA from the top battery, but that does add up over time. Not only that, unequal sense current from the batteries leads to unbalance. That is less of a problem of course with high capacity packs where 0.6 mA is next to nothing in the short term.
Commercial Lithium battery packs most often already include an integrated charge balancing PCB, and also an unforgiving undercharging shutoff and a polymer fuse to prevent sustained high currents.