GGoodrum
1 MW
Sorry, I had a phonecall. 
I just enabled it. The page is here: http://www.tppacks.com/products.asp?cat=26
-- Gary
I just enabled it. The page is here: http://www.tppacks.com/products.asp?cat=26 -- Gary
-
Howdy! we're looking for donations to finish custom knowledgebase software for this forum. Please see our Funding drive thread
Even Newer 4 to 24-cell Battery Management System (BMS)
- Thread starter GGoodrum
- Start date
greasypants
10 W
- Joined
- Mar 10, 2007
- Messages
- 98
Just ordered two boards, thanks for all the hard work. I am planning a 28s2p of PSI cells on my next bike (assuming I can find a place to put them all). Would I be able to add a cell to the 24 cell board for a total of 28 cells without an additional power supply and us one big charger or would it be better to split the pack into one 12s2p and one 16s2p packs. Also what voltage battery chargers would be required for the three different size batteyr packs. Thanks again .
ahambone
10 W
- Joined
- Oct 6, 2008
- Messages
- 71
My order is in... looking forward to the project!
Gary, Richard - you guys rock!
If everything goes according to plan I will be using this on a (30+p)24s pack for an EV conversion. We will see how the construction of that pack goes over time . Even if I go lead-acid on that project I still have several uses for the BMS boards - too many HWC 18650 and A123s around to not have a good BMS!
Cheers,
--adam
Gary, Richard - you guys rock!
If everything goes according to plan I will be using this on a (30+p)24s pack for an EV conversion. We will see how the construction of that pack goes over time
. Even if I go lead-acid on that project I still have several uses for the BMS boards - too many HWC 18650 and A123s around to not have a good BMS!Cheers,
--adam
ldoolit
10 µW
I just ordered one board. Thanks so much for putting this thing together!
I don't see any connectors in the BOM. Am I blind?
My cells are from ThunderSky, they have slightly higher voltage ratings than traditional LiFePO4.
On the LVC side, I plan to substitute a 2.7 Volt TC54VC2702EZB for the 2.1 Volt TC54VC2102EZB.
On the shunt control side, I will rejigger the voltage divider. Right now the set point is 2.5*(180+75)/180=3.54V.
I will change out the 75K for 100K, giving me 2.5*(180+100)/180=3.89V. Does this sound plausible?
- Larry
I don't see any connectors in the BOM. Am I blind?
My cells are from ThunderSky, they have slightly higher voltage ratings than traditional LiFePO4.
On the LVC side, I plan to substitute a 2.7 Volt TC54VC2702EZB for the 2.1 Volt TC54VC2102EZB.
On the shunt control side, I will rejigger the voltage divider. Right now the set point is 2.5*(180+75)/180=3.54V.
I will change out the 75K for 100K, giving me 2.5*(180+100)/180=3.89V. Does this sound plausible?
- Larry
GGoodrum
1 MW
Wow, the first batch is sold out already. 
There appears to be lots of lurkers out there.
I've just enabled a bunch more, but these won't ship until later this week. I put a note in the description about this.
-- Gary
There appears to be lots of lurkers out there. I've just enabled a bunch more, but these won't ship until later this week. I put a note in the description about this.
-- Gary
ldoolit said:
It doesn't work out exactly like the math would indicate since we're using very high resistances in the divider to reduce standby power consumption. With the parts indicated in the schematic, I'm getting about 3.68v average. Some are as high as 3.7v. If you want to go higher than 3.68v, then calculate based on scaling the existing values.
voicecoils
1 MW
quick question, If I buy one board, can I build it up as two 12 cell BMS's? If so, do I need to order 2 * 12 channel BOM instead of one set of 24 channel BOM? I notice not everything is doubled...
Cheers!
Cheers!
You would need to buy two boards since there is only one control circuit on each board.
If you wanted to run the two 12 cell units together but just split the cell circuits into two packs, then you could do it with one board. In this case you would need to run 4 wires from the main board to the other one and charge them both at the same time with a 24 cell charger.
If you wanted to run the two 12 cell units together but just split the cell circuits into two packs, then you could do it with one board. In this case you would need to run 4 wires from the main board to the other one and charge them both at the same time with a 24 cell charger.
Here's a picture of the ver. 2.1 prototype. In the newest version of the boards, the junk with the red wires on the left has been incorporated into the board. As you can see, the control circuit is combined with the cell circuits for cells 1 to 4.
methods
1 GW
Does anyone have a "home brew" method for conformal coating?
I am thinking something like 10 coats of a polycrylic or something like that.
If I dont try something like that I may try a thin heat-shrink case that is sealed on both ends.
Not worried about heat as my packs tend to stay balanced.
My Project Mayhem bike will be rain-proof
-methods
I am thinking something like 10 coats of a polycrylic or something like that.
If I dont try something like that I may try a thin heat-shrink case that is sealed on both ends.
Not worried about heat as my packs tend to stay balanced.
My Project Mayhem bike will be rain-proof
-methods
rkosiorek
100 kW
it's less expensive to just use the right stuff to begin with.
12oz cans of the comercial stuff are about $25.00CAD and just spray on. you have a choice of Urethane, Acrylic or Silicone.
i have used the urethane and HT silicone varieties.
http://www.techspray.com/newinfo/2106.pdf
http://www.techspray.com/newinfo/2103.pdf
prepare the boards by washing with Triclo or 99% Isopropyl alcohol. rubing alcohol is not god enough. you can usually get the pure stuff at the pharmacy. it will be behind the counter so you'll have to ask the pharmasist for it. helps if you don't look like you're going to drink it when you ask though.
just spray on at room temperature. it will dry in about 24hrs. or better yet pop it into a 120F to 150F oven for 15 to 20 minutes.
it can be removed with the standard eletrical industry conformal coating solvents.
in Canada you can get them at Electrosonic.com and in the States mouser carries them but will not ship them outside of the continental USA.
rick
12oz cans of the comercial stuff are about $25.00CAD and just spray on. you have a choice of Urethane, Acrylic or Silicone.
i have used the urethane and HT silicone varieties.
http://www.techspray.com/newinfo/2106.pdf
http://www.techspray.com/newinfo/2103.pdf
prepare the boards by washing with Triclo or 99% Isopropyl alcohol. rubing alcohol is not god enough. you can usually get the pure stuff at the pharmacy. it will be behind the counter so you'll have to ask the pharmasist for it. helps if you don't look like you're going to drink it when you ask though.
just spray on at room temperature. it will dry in about 24hrs. or better yet pop it into a 120F to 150F oven for 15 to 20 minutes.
it can be removed with the standard eletrical industry conformal coating solvents.
in Canada you can get them at Electrosonic.com and in the States mouser carries them but will not ship them outside of the continental USA.
rick
rkosiorek
100 kW
NO!! There is a difference between Conformal Coatings and POTTING COMPOUND.
Conformal coatings are designed to protect the board from weather, moisture, vibration. but they are applied in a relatively thin layer. much like a thick paint. they can be removed to allow service.
Potting compounds like those used on the DeWalt BMS are designed to be a permanent seal. they are a thick very hard and solid substance. they are usually so thick and viscous that they do not readily pour unless heated. they are more or less permantly moulded onto the part. often to make sure that the compound fills all of the voids during molding and to release any air bubbles or pockets the parts are put onto motion or vibration tables to shake out all of the bubles while the compound is hardening. occassionaly centrifugal casting is used. these componds are usually hardened either by heat or ultraviolet light. the cure heat is much higher than the molding temperature. the molding temp is around 100F to 120F and the cure temp is 150F to 180F
rick
Conformal coatings are designed to protect the board from weather, moisture, vibration. but they are applied in a relatively thin layer. much like a thick paint. they can be removed to allow service.
Potting compounds like those used on the DeWalt BMS are designed to be a permanent seal. they are a thick very hard and solid substance. they are usually so thick and viscous that they do not readily pour unless heated. they are more or less permantly moulded onto the part. often to make sure that the compound fills all of the voids during molding and to release any air bubbles or pockets the parts are put onto motion or vibration tables to shake out all of the bubles while the compound is hardening. occassionaly centrifugal casting is used. these componds are usually hardened either by heat or ultraviolet light. the cure heat is much higher than the molding temperature. the molding temp is around 100F to 120F and the cure temp is 150F to 180F
rick
i think the intent was to reuse existing connectors.
if you need connectors you can buy them or if you are up to getting old computer motherboards and desoldering and cutting the wires from old computer supplies they are a good source of connectors.
you can also use the 24 to 20 pin connectors used in some newer computer power supplies too.
if you need connectors you can buy them or if you are up to getting old computer motherboards and desoldering and cutting the wires from old computer supplies they are a good source of connectors.
you can also use the 24 to 20 pin connectors used in some newer computer power supplies too.
ldoolit said:
wybornd
10 µW
rf said:Looks like you really need a back-order capability on you web page, Gary.
Then you'd know to order two-thousand boards instead of two hundred.
Richard
Whew that went fast, 17 hours of timezone difference is a killer ... I think if the back order capability is possible, it may save Gary a thousand PM's and emails from guys like me who missed out.
Anyway, I will wait patiently till a new batch is available.
Cheers Dean
BTW: Well done Gary and Richard et al
PJD
100 W
I have been running an earlier version of this BMS for a few months now, and the actual shunt activation voltage for the 75/180 resistor pair is about 3.67 to 3.70 volts. So, your 3.89 volt value will probably be correspondingly higher as well.
At risk of wandering a bit off-topic, the consensus is (or was?) that 3.65-3.70 charging voltage limit is also a good idea for Thundersky cells too.
As far as the recommended 2.5 volt minimum for Thunderskys. I am a Thundersky user, and I stayed with the 2.1 volt LVC part myself rather than going up to 3.7 volts. I think it is unlikely that their 2.5 volt minimum applies to higher discharge rates (2C to 3C) or at colder temperatures. I suspect the real situation is probably more like SLA battery performance data - the actual "do not go below" voltage is lower at higher discharge rate and lower temperature. I have done several near and below freezing temperature rides so far, and at 26-32F, even a fully-charged pack will sag to about 2.85 volts per cell at a 2.25C rate. This went down to 2.65 volts per cell once discharged about a 50% SOC. So, I would have already encountered cutouts in some situations in uphill traffic if I had used the 2.7 volt part.
ldoolit
10 µW
I will recompute based on about 2 microamps of input bias current for the FAN431L, and target about 3.9Volts. I tend to engineer in compromises.PJD said:
2.1 Volts sounds terribly low to me, even under load. I have the luxury of not needing to ride in temperatures that low. The bike is garaged before my morning ride, and afternoons around here don't normally go below 45F. I think I'll take my chances at 2.7V. I can always swap them out if it becomes a problem. Using my old SLA pack I learned how much power it takes to stay with traffic: extrapolating to these Thunderskys it works out to between 2C and 3C. More is betterPJD said:
I'd like to get up to 5C starting from a green light.In the long run, I'd like a smarter LVC that knows about the current, and could tell the difference between current-induced sag and and a discharged cell. But I have to get back on the road sooner than I can put something like that together.
Thanks for your input there.PJD said:
The charging voltage for Thundersky cells is spec'd higher but there seems to be little risk in staying a bit low on the votlage as long as you can get a full charge. In fact it will probably make the cells last longer.
Have you measured the capacity of your pack? That would be a good indication if it works OK.
I'm also curious if your issue with balancing time has changed any with repeated use? What kind of balancing times are you getting now?
Since you are the first one I know of that's using it with Thundersky cells, you certainly earn a guinea pig award.
Even if you do not need a conformal coating, it would be a very good idea to glue the capacitor cans to prevent them from vibrating and breaking off. Silicone RTV works OK. Other glues would work too.
Where the cell wires attach to the board is another vibration prone spot. Glue can help here also, or zip ties to keep the wires from moving.
PJD
100 W
Richard,
I now have about charge 35 cycles using the BMS on the two scooters with the 48 volt 40AH Thundersky cells.
At end of charging with 8-amp chargers, the shunt indicator LEDs are generally all flickering with uniform brightness within about 15 minutes of the first LED going on. The bi-color power LED then transitions gradually from red-orange to greenish-yellow over about 45 additional minutes. The charger's amperage declines from 2 amps to about 0.5 to 0.6 amps over this same time. I am assuming that this reflects a combination of the cells passing less and less current along with the regulator throttling back to just a bit over the shunting current.
Cell voltages when shunting vary from about 3.665 to 3.70 volts - but the first cells to start shunting may rise as high as 3.72-3.73 volts before settling down to 3.69 volts once the current cuts back.
I have been considering uniform brightness shunt LED's, along with a greenish yellow bi-color LED, as my criteria for "fully charged".
The BMS's are mounted on the underside a hardboard access cover with 1/4-inch holes drilled over each shut resistor for ventilation. In the current cooler season, the charger can be left connected (to this non-cutoff equipped earlier version) for at least 2 hours without apparent problems - but the part of the board over the resistors and shunt transistors gets quite warm to touch.
I now have about charge 35 cycles using the BMS on the two scooters with the 48 volt 40AH Thundersky cells.
At end of charging with 8-amp chargers, the shunt indicator LEDs are generally all flickering with uniform brightness within about 15 minutes of the first LED going on. The bi-color power LED then transitions gradually from red-orange to greenish-yellow over about 45 additional minutes. The charger's amperage declines from 2 amps to about 0.5 to 0.6 amps over this same time. I am assuming that this reflects a combination of the cells passing less and less current along with the regulator throttling back to just a bit over the shunting current.
Cell voltages when shunting vary from about 3.665 to 3.70 volts - but the first cells to start shunting may rise as high as 3.72-3.73 volts before settling down to 3.69 volts once the current cuts back.
I have been considering uniform brightness shunt LED's, along with a greenish yellow bi-color LED, as my criteria for "fully charged".
The BMS's are mounted on the underside a hardboard access cover with 1/4-inch holes drilled over each shut resistor for ventilation. In the current cooler season, the charger can be left connected (to this non-cutoff equipped earlier version) for at least 2 hours without apparent problems - but the part of the board over the resistors and shunt transistors gets quite warm to touch.
drunkencat129
100 W
- Joined
- Aug 5, 2008
- Messages
- 116
i want thiss bms lol im willing to put my packs on the line for it lol how much and where do i get itfechter said:
i need 2 20cell bms setups i got crazy lipo packs and a lime pack from yardworks that i need a bms badly for help me out please
PJD said:
OK, that sounds very good. I also noticed the cell voltages ran a bit higher during charge, then backed off slightly when the charge finished. I also noticed that the shunt LEDs stay on a little while after disconnecting the charger as part of the surface charge is drained off.
I over-rated the shunt resistors quite a bit to help reduce maximum surface temperature. Ideally they should never exceed 100C on the surface, but they can withstand much more.
The shunt transistors get pretty toasty too, but same thing applies.
Drunkencat,
nobody has ever tried this with Lipo yet, but I don't see any reason why it should not work well, especially with the auto cutoff circuit at end of charge. To make it work, the value of one resistor in each cell circuit would need to be changed to bring up the max charge voltage.
I don't know, what is the ideal voltage for Lipo
100c is 212F the point that water boils.
i was wondering if http://www.vishay.com/docs/30213/cp.pdf style resistor or
style could be used because it would be easier to attach a heat sink to them to get rid of the heat better?
correct me if i am wrong the other thing is maybe it is better to let the resistors get hot because heat cuts down on the efficiency causing more power to be wasted as heat otherwise it may take longer to bleed off the power to balance the cells.
i was wondering if http://www.vishay.com/docs/30213/cp.pdf style resistor or
correct me if i am wrong the other thing is maybe it is better to let the resistors get hot because heat cuts down on the efficiency causing more power to be wasted as heat otherwise it may take longer to bleed off the power to balance the cells.
fechter said:PJD said:
OK, that sounds very good. I also noticed the cell voltages ran a bit higher during charge, then backed off slightly when the charge finished. I also noticed that the shunt LEDs stay on a little while after disconnecting the charger as part of the surface charge is drained off.
I over-rated the shunt resistors quite a bit to help reduce maximum surface temperature. Ideally they should never exceed 100C on the surface, but they can withstand much more.
The shunt transistors get pretty toasty too, but same thing applies.
Drunkencat,
nobody has ever tried this with Lipo yet, but I don't see any reason why it should not work well, especially with the auto cutoff circuit at end of charge. To make it work, the value of one resistor in each cell circuit would need to be changed to bring up the max charge voltage.
I don't know, what is the ideal voltage for Lipo
methods
1 GW
drunkencat129 said:
I am setting up for 2 lipo boards right now. If you search back you will find where we calculated what the correct resistors would be for a 4.2V upper cut off. It is a simple 1 resistor change from the BOM
For LVC, the bad news is that you can only get a 2.7V part. The BOM calls out a 2.1V part. The ideal LVC for a lipo is 3.0V. Some go down to 2.7V but I think that is a little reckless. There is a 3.0V part but it is SMT so you would have to play games at solder time.
For my Lipo board I am going to order both the 2.7V LVC parts and the 3.0V SMT parts. If the SMT parts solder down well I will use those. If not, I will use the 2.7V parts for now until I can source the 3.0V parts in low quantities. Moot point as I never, ever, ever hit LVC on my lipo packs.
note: Some will argue that a 2.7V cutoff is more appropriate for Lipo but I would argue that these days there is no reason to run a lipo with less than a 20C discharge rate and a cell with that low of internal resistance should not ever need to sag below 3.0V. If it does, I would consider that abuse. All high quality RC aircraft speed controls protect the battery with an LVC of 3.0V. Some Chinese ESC's go down to 2.7V. I would say that 95% of lipo users cut off at 3.0V If you are using very low C lipo (like 2 - 3C) then you may consider a 2.7V cutoff more appropriate.
As far as where do you get them? http://www.TPpacks.com in the bike section. They are about $50 bones and the parts are about $80 bones. Money well spent
For my 24S 9Ah lipo pack it is the perfect solution.-methods
Similar threads
