thedarlington
100 W
very nice evolution the projects had, too bad its so intricate id be scared to solder one of these but awesome job you all are doing
"Zephyr" - Finally, the "v4" Fechter/Goodrum/Hecker BMS...
- Thread starter GGoodrum
- Start date
How important is it to heat sink the shunts?
I've got 24 cells with a 720W charger.
I mounted some caps on the bottom. I can't fit the board in the case with the shunts touching. I hadn't thought to install the shunt resistors to compensate.
Thanks.
I've got 24 cells with a 720W charger.
I mounted some caps on the bottom. I can't fit the board in the case with the shunts touching. I hadn't thought to install the shunt resistors to compensate.
Thanks.
chenwood said:
If the charger voltage is set so the shunts are barely on at end of charge, the heating should not be too bad. If you have badly imbalanced cells and most of the shunts run full blast, there will be quite a bit of heat. You can try it and see how it goes. You may also be able to lay down a strip of aluminum over the resistors to fill the gap. Use silicone glue to hold the strip to the resistors. This may be a better approach anyway, as the resistors will be better protected when sliding the board into the case.
Zenid
100 W
Already sorted. I found a batch of replacement green/red LEDs on ebay. They're a bit cheap and nasty, but I've installed one and it's doing its job - red is back.fechter said:
I'm not bothered about whether the LEDs light, I thought that they were indicating when a given cell had reached the 3.6V 'full charge' point it's been given, and that therefore they would all be lit by the time it switched off. Since this isn't important, I've dialed the pot a couple of hundred ohms to the right again.fechter said:
Yes, I'm very happy with it. After a ten mile spin, I'm fully recharged again in just under an hourfechter said:
(at 9A with my EMC-900). Under my old regime it would take about 4 hours! See my blog for full details of the setup.
http://zenid10.wordpress.com/2011/07/14/thats-lifepo4/
One slightly unnerving thing I've noticed with the pot setting though: With each cycle, it seems that the pot needs turning further and further round to the right to make switch off at the right time. Could this just be the Lithium cells "bedding in" - something to do with cell chemistry or whatnot, or is there some other explanation for this oddity? Like you say it's important that they don't stay on, as they unit gets quite hot at the end of a charge if they do (like when the pot is set too low).
Gregb
100 W
Zenid have you tested the voltage on the cells with a VM at the end of charge with it still on?? are they 3.6V?
Zenid said:
I'm not sure. Most batteries do behave a bit differently on the first few cycles and need to be 'broken in'.
Otherwise there could be some drift in the charger voltage setting. You can check that with a voltmeter.
There is going to be a little variation in the shunt voltages with changes in temperature also. I'm not sure how much variation to expect, though. I've never seen them drift by more than a few mV. With extreme temperature changes both the shunts and the charger (and cell behavior) could vary significantly.
You might experiment with increasing the cutoff setting to see how much difference in capacity you really get. Normally we assume when the charge current gets down to C/50 or so, they are as full as they're going to get. If you cutoff at a slightly higher current, you'll have a bit less charge but this may improve the cycle life of the cells as long as you don't need 100% of the full capacity.
Ideally you'd want to have an ammeter on the charger to check the current. For a given pot setting, the cutoff current should remain fairly consistent.
Gregb
100 W
Fechter, had a chance to check that board in the revised reference??
steveo
100 kW
Can anyone please advise what voltage should be set on my charger for a 24s lipo battery
100.8v???
I had my psu set to that and the led is remaing orange green colour,
Also does the jumper plug need to be on the bms board, or is it only plugged in for testing?
Thanks
Steveo
100.8v???
I had my psu set to that and the led is remaing orange green colour,
Also does the jumper plug need to be on the bms board, or is it only plugged in for testing?
Thanks
Steveo
steveo
100 kW
steveo said:
The EOC disable jumper is only installed for testing. Normally it is removed.
100.8v should be about right for the voltage.
The LED only goes red during HVC activation. You can test this by shorting the HVC line when the circuit is on.
100.8v should be about right for the voltage.
The LED only goes red during HVC activation. You can test this by shorting the HVC line when the circuit is on.
steveo
100 kW
fechter said:
Hey Fechter
I got the main led to turn green near end of charge, still fine tuning it
I have a (4 - 6s packs) 1p 24s pack running on bms, 2 packs have about 1ah more then the other 2
When running it on the bms, the packs with more charge charged up to 4.25-4.27v approx, the 2 other packs sitting around 4.10v range
The main led went green to cut out the chargers, my question for ya is why are my discharge resistors not getting warm during balance/ discharge? Im getting heat near the led part but thats it, if i leave the bms going eventually the higher voltage cells will come down to match the lower cells, it could be the draw from the leds doing the work perhaps lol
Thanks
-steveo
You should probably try testing the shut voltages without cells attached. Using a light bulb in series with your supply, you should be able to light up a portion of the shunts (bypass the control circuit). If you go back a few pages, you can see the diagram I made for Zenid. When the shunt LEDs are partially lit, measure the voltage across each one to make sure it's in the right range.
No heat indicates the cells are not up to the shunt voltage. You could probably increase the charger voltage slightly. No heat is really one of the design features we worked hard to get. A little heat would be ideal. If the shunt voltages are set too high for some reason, this could also explain it.
Sounds like you have the control circuit working OK.
No heat indicates the cells are not up to the shunt voltage. You could probably increase the charger voltage slightly. No heat is really one of the design features we worked hard to get. A little heat would be ideal. If the shunt voltages are set too high for some reason, this could also explain it.
Sounds like you have the control circuit working OK.
steveo
100 kW
fechter said:
Hey fechter
A new video for you, i did the best i could based on how you advise me to test it
http://www.youtube.com/watch?v=oCbRv_-_K98&feature=youtube_gdata_player
Maybe i should lower my chargers so that im charging the cells to 4.15 instead of 4.20, it seems based. On my testing shunts activate at 4.15v
Let me know your thoughts, still no heat at all on discharge resistors even with some cells at 4.25v at top of charge
Steveo
Looks like the shut activation voltages are in the right ballpark.
The cells should not be going over this voltage by any significant amount, so I'm not sure what's going on there.
You can sometimes have enough voltage drop in the tap wires to make a difference, but this would only be at full current. Measuring on the board connections vs. the cell end of the wires should largely eliminate this.
It almost sounds like the BD136 transistors are not turning on. It's pretty confusing which way to install them, so they could possibly be backwards. The schematic and datasheet would be one way to double check this.
Another test would be to use your light bulb setup but only feed a portion of the cells so that there will be more current available. If the cells are going significantly over 4.15v it indicates a problem. A current limited bench power supply is really the best way to test this.
The cells should not be going over this voltage by any significant amount, so I'm not sure what's going on there.
You can sometimes have enough voltage drop in the tap wires to make a difference, but this would only be at full current. Measuring on the board connections vs. the cell end of the wires should largely eliminate this.
It almost sounds like the BD136 transistors are not turning on. It's pretty confusing which way to install them, so they could possibly be backwards. The schematic and datasheet would be one way to double check this.
Another test would be to use your light bulb setup but only feed a portion of the cells so that there will be more current available. If the cells are going significantly over 4.15v it indicates a problem. A current limited bench power supply is really the best way to test this.
steveo
100 kW
fechter said:
Bd136 is backwards!, good call, on my board the writing on the transistor is visable, on the photo on the instruction it is not, crap... Not looking forward to takin those things out gahhhh
This should be outlind in the instructions!
Ill let u know what happens
Thanks
Steveo
Zeropointbug
1 W
I haven't looked through his whole thread, so I was wondering what cell count is your BMS designs? Would I be able to purchase one now, or the circuit board and build it myself?
I have a 16s A123 15Ah pack (51V)
Thanks
I have a 16s A123 15Ah pack (51V)
Thanks
steveo said:
Yeah, that and the capacitors and a few assorted mismarks on the board. Fixing up the instructions better is on the to-do list.
I think your BD136s should be fine after you turn them around. Good thing you tested. An industrial strength desoldering station would be real handy about now...
Zeropointbug said:
The boards Gary has are 24s. If you want to run 16s, you can cut off the extra part or simply leave it with no parts installed. Any number of cells can be supported by adding or removing sections.
Right now, only bare boards are available from Gary. Parts list you have to order yourself.
Andyh was building boards for people, but we haven't heard anything from him for months. I think we killed him or he went insane soldering all those little resistors. We may need to send out a search party to see what happened to him.
Zeropointbug
1 W
Okay, so would you recommend your board vs. the 30A BMS I have from Cell_man? Which I already have, but am hesitant to use it as the in rush current rfom plugging into controller is too hard on it.
SO, if I don't want to run the BMS when running bike (to controller), and only when charging, this would be okay, correct? (Cell_man's 30A BMS)
SO, if I don't want to run the BMS when running bike (to controller), and only when charging, this would be okay, correct? (Cell_man's 30A BMS)
Zeropointbug said:
That could be risky as there would be no cell level protection during discharge. If you know the actual capacity of the battery and have a way to measure it while riding, then you could most likely avoid overdischarging and destroying the batteries.
If you already have a Cell man BMS, I would stay with it unless it doesn't work. The inrush current from connecting the controller should be avoided with any setup by using some kind of precharge resistor. Big sparks are bad.
The Zephyr board allows the batteries to go straight to the controller, so there is no issue with high discharge current. You still need a main fuse or breaker and the controller should have some kind of current limiting.
intoworldsofuncertainty
100 µW
Hi,
I'm interested in purchasing a "Zephyr" BMS printed circuit board but have found the tppacks website to be under construction. Does anyone know when it will be back on-line?
cheers!
I'm interested in purchasing a "Zephyr" BMS printed circuit board but have found the tppacks website to be under construction. Does anyone know when it will be back on-line?
cheers!
Gregb
100 W
Steveo, have you actually measured the PD across the shunt resistors to see what current is flowing in them ??
Gregb
100 W
I think that is a "how long is a piece of string" question. Due to his provider going belly up with no warning Gary has to rewrite his site......intoworldsofuncertainty said:
StJaux
1 mW
Richard/Gary,
Great work! Been following the post with interest, any chance of us viewing the circuit diagram?
Thanks,
John
Great work! Been following the post with interest, any chance of us viewing the circuit diagram?
Thanks,
John
GGoodrum
1 MW
Gregb said:
Yes, I'm still trying to get my hands around this task.
In the meantime, you can PM me if you want a board. It's $79 for a 16-channel version and $99 for the 24-channel variant. Shipping is about $5 domestically, and about $15 for international.
-- Gary
