Advantages and disadvantages Headway vs. Prismatic ...

I want to built a 72V16Ah Lifepo4 24S battery pack and want to use it with a 40A ebike.ca controller on my EV. I now have to make a choice between cylindrical or prismatic cells with soft, or with hard shell. So far I have seen that price, weight and compactness is a little in the disadvantage for Headways, but many professional users swear by them. Why is that ?
I have searched the ES forum, but couldn’t find a thread briefly dealing with only this subject. Can anyone of you experts, experienced users or gurus tell me what the advantages are of both types of cells compared to each other with regard to robustness, weight, durability etc. and any other aspect important to know ? What should make you choose for one rather than for the other ?
Thanks.

If you are looking for more details, kindly visit DIY Lifepo4 Battery Pack.

For more new energy storage cabinetinformation, please contact us. We will provide professional answers.

My other friend has a similar setup, but the pack has an LVC-only board. He uses this bike a lot less frequently, and charges it with a no-name 3A charger. Two weeks ago I checked this pack and even though there has been zero balancing, the cells were still amazingly close in voltage. The difference here is that he has never discharged this pack down below about 50%. My first friend, on the other hand, uses about 8-9Ah each way, and charges both at work and at home.

My LiPo setups behave similarly, in that when I keep the discharges to something less than about 50-60%, they will stay very closely balanced. I can go 10, even 20 cycles, and there just isn't much difference. If, on the other hand, I drain them down to 80-90%, I do start to see some more significant differences, as much as 100-150mV sometimes. As Jeremy points out, I think a lot of this is do to fact that the cell IR changes more as the cell gets closer to the end-of-capacity.

In any case, I'm zeroing in on a simple approach for a "safe" LiPo setup, both for discharging and for charging. I'm now back to having a combo LVC/HVC board that also serves as a parallel adapter board for connecting multiple LiPo packs together in parallel, using the JST-XH balance tap pigtails that are on each Turnigy/Zippy pack. Here's what one looks like:




There's a two-wire opto-coupled signal output that is used for both LVC and for HVC signals. Multiple 6s or 8s LVC/HVC boards can be "daisy-chained" together, as shown here:



This simple combo signal is used two ways. For LVC protection, it is simply connected to either the ebrake signal on a controller, or is tied directly (using a current limiting resistor...) to the throttle signal. Basically what happens is the throttle signal is cut, briefly, if any cell tries to go below the LVC set point (2.1V for LiFePO4 and 3.0V for LiPo...). As soon as the load is removed, by killing the throttle, the cell voltage will recover up above the trip point and the circuit will release the pull-down of the throttle. If the throttle is still "engaged" and the load returns, the LVC circuit will "hit" the throttle again. This cycle will repeat at about a 1-2Hz rate until the throttle is backed off far enough that the load demand won't trip the LVC circuit. It should be noted that the cell voltage will recover back up to close to the "nominal" voltage, with no load on the system. Anyway, this has proven to be very effective way to keep cells from being over-discharged, under load, which is the only sure-fire way of killing both LiPo and LiFePO4 cells.

There is still the case, however, where a controller could be left on, causing a slow drain on the cells. This can drain a cell "dead", down to 0V even, but in most cases, this slow "trickle drain" won't cause the permanent cell reversal that basically makes the cell a dead short, that over-discharging under load will do. For my own setups, this hasn't been a problem, because I use big 75A Andersons between the controller and the pack, and I've just gotten into the habit of always disconnecting the pack after a ride. Richard and I are looking at doing a "smart" active cutoff module, though, that can prevent the case of a slow drain from leaving things connected, killing cells. The reason it is "smart" is because it can differentiate between LVC trips under load, and the LVC activation from slow drains. We used to have an active cutoff module, that worked quite well, even with 150-200A setups, but what I found is that it is not good to cycle master power to the controllers, when the LVC starts tripping. With the new module, trips under load will still hit the ebrake input and/or pull down the throttle signal. Slow drain trips will cut pack power completely, until manually reset. The intent is that this module can be added in a setup later, and make use of existing LVC signals.

The HVC signal, which can share the same opto line, is used in a similar fashion, but during the charge process. If a cell gets full quicker than the rest, it will hit that voltage "knee" where the voltage will start to rise at a faster rate. For LiFePO4 cells, this is around 3.60-3.65V, and for LiPo it is around 4.15-4.20V. With balanced cells, they all will hit this point at roughly the same time. If the charger's CV mode is set at this point, it will hold all the cells at this point, which will cause the current to slowly drop, until it gets all the way down to zero. Actually, when it gets down to something like 1/20-1/50th of the original charge current, the cell is about as full as it will get. Anyway, this is what happens when the cells are balanced. If not balanced, one, or more, cells can get full faster than the rest, and will hit the "knee" before the rest. This can cause the voltage for this cell to rise quickly to the point of cell damage, or worse, while the charger is still not in the CV current limiting mode. The HVC circuits on each channel will trip if this happens, and this signal is used by a "charge controller" to throttle back the current by however much is needed to keep the errant cell voltage from going any higher. Think of it as individual cell CV modes.

On the full BMS, the charge controller logic is on the same board. When we started doing the LiPo variants, I split the charge controller circuit out into its own small box. This standalone charge controller also performed one othr function and that was to monitor the charge current and cut it off when it dropped below an adjustable set point. Recently, Richard came up with a clever way to add charge current limiting to "dumb" power supplies, like the inexpensive MeanWells. These popular supplies aren't originally meant to used as battery chargers, so most lack a true constant current (CC) mode, having a simple "hiccup" mode on the front-end, for overload protection. Richard's small widget mounts to the output terminal strip, and ties into a sense input via a single hookup wire that is tacked to the top of a resistor, next to the voltage adjustment pot. There is an adjustment pot that can be used to set the desired charge current from 0 to the max the supply is rated for. Since this worked so well, we then added to the current limiter functionality, turning it into a complete Charge Controller. In addition to the charge current limit, there is also another pot for setting the low current/end-of-charge trip point, which is adjustable from 0-1A. It also has the input for the HVC signal, which it will use to throttle the current should a channel's HVC circuit start tripping.

So, it is possible now, I think, to make a reliable "commuter" LiPo pack that not only has individual cell LVC protection, but also can be safely bulk-charged with an inexpensive MeanWell supply. Balancing can be handled a number of ways. The DIY crowd can build one of the CellLog-based balancers, which can provide the simplest solution, in terms of ease-of-use, or there are a number of RC-type balancing solutions that can work quite well. Until recently, it was hard to get an RC-based off-the-shelf charging/balancing solution that didn't require a lot of the sort of "pack manipulation" described by Dogman. The other problem I've had with most RC balancers and/or balancing chargers, is that they don't provide much in the way of balancing current, most being limited to 100-200mA. This might be fine for most "single-p" RC packs, but for a 15-20 Ah ebike pack, it can take literally half-a-day to balance a pack. Our original "full BMS" unit managed 500mA, and the newer v4.x version, and the new CellLog-based balancers, all handle a bit over 1A.

I have, however, just gotten a hold of a couple of new Hyperion balancing chargers that have a number of useful features for us. For one, they can handle pack setups up to 14s LiPo, which makes it quite simple to use with a "typical"12s LiPo commuter pack. Another improvement in these, over most other RC units I've used, is that it has 500mA of balance current, and uses 12-bit A/D converters for greater balancing accuracy. Finally, the input voltage can be as high as 30V, which means I can drive one of these from the same S-350-24 MeanWell supply I use for bulk charging. They have a max charge rate of 20A. Here's a link to info about these chargers:

HYPERION EOS 1420i NET3 CHARGER - 1S-14S, 20A MAX, 550W

I'm still a Hyperion dealer, so if these work out, I'll start offering these on my site, with a big discount for E-S members. I need to do some connection adapters, but there should be no reason you couldn't plug this into the same balancer plugs coming from the LVC/HVC boards.

Not sure I've cleared anything up here, or muddied the waters even further, but I will say that I don't give a rat's ass about cricket! :lol:

-- Gary

Okay, if our English-speaking cousins are done with the cricket-bashing ( :lol: ), I'd like to weigh in here on a couple of things. First of all, I think the most reliable/enjoyable commuter setup I have is a 9C with a 12-FET Lyen controller. Although I haven't quite got there yet with a "turnkey" LiPo commuter setup, we are getting pretty close. What I've done for a few friends who I've done these 9C setups for, is a 48V 16-cell PSI LiFePO4 pack. One still has an original v2.6 full BMS board, and is charged with a 4A Soneil charger. Last time I checked, the cells were all reasonably balanced, and they appear to still have pretty close to their original 10Ah capacity. My friend has been using this for a 7 mile commute about three times-a-week, for about two years now.My other friend has a similar setup, but the pack has an LVC-only board. He uses this bike a lot less frequently, and charges it with a no-name 3A charger. Two weeks ago I checked this pack and even though there has been zero balancing, the cells were still amazingly close in voltage. The difference here is that he has never discharged this pack down below about 50%. My first friend, on the other hand, uses about 8-9Ah each way, and charges both at work and at home.My LiPo setups behave similarly, in that when I keep the discharges to something less than about 50-60%, they will stay very closely balanced. I can go 10, even 20 cycles, and there just isn't much difference. If, on the other hand, I drain them down to 80-90%, I do start to see some more significant differences, as much as 100-150mV sometimes. As Jeremy points out, I think a lot of this is do to fact that the cell IR changes more as the cell gets closer to the end-of-capacity.In any case, I'm zeroing in on a simple approach for a "safe" LiPo setup, both for discharging and for charging. I'm now back to having a combo LVC/HVC board that also serves as a parallel adapter board for connecting multiple LiPo packs together in parallel, using the JST-XH balance tap pigtails that are on each Turnigy/Zippy pack. Here's what one looks like:There's a two-wire opto-coupled signal output that is used for both LVC and for HVC signals. Multiple 6s or 8s LVC/HVC boards can be "daisy-chained" together, as shown here:This simple combo signal is used two ways. For LVC protection, it is simply connected to either the ebrake signal on a controller, or is tied directly (using a current limiting resistor...) to the throttle signal. Basically what happens is the throttle signal is cut, briefly, if any cell tries to go below the LVC set point (2.1V for LiFePO4 and 3.0V for LiPo...). As soon as the load is removed, by killing the throttle, the cell voltage will recover up above the trip point and the circuit will release the pull-down of the throttle. If the throttle is still "engaged" and the load returns, the LVC circuit will "hit" the throttle again. This cycle will repeat at about a 1-2Hz rate until the throttle is backed off far enough that the load demand won't trip the LVC circuit. It should be noted that the cell voltage will recover back up to close to the "nominal" voltage, with no load on the system. Anyway, this has proven to be very effective way to keep cells from being over-discharged, under load, which is the only sure-fire way of killing both LiPo and LiFePO4 cells.There is still the case, however, where a controller could be left on, causing a slow drain on the cells. This can drain a cell "dead", down to 0V even, but in most cases, this slow "trickle drain" won't cause the permanent cell reversal that basically makes the cell a dead short, that over-discharging under load will do. For my own setups, this hasn't been a problem, because I use big 75A Andersons between the controller and the pack, and I've just gotten into the habit of always disconnecting the pack after a ride. Richard and I are looking at doing a "smart" active cutoff module, though, that can prevent the case of a slow drain from leaving things connected, killing cells. The reason it is "smart" is because it can differentiate between LVC trips under load, and the LVC activation from slow drains. We used to have an active cutoff module, that worked quite well, even with 150-200A setups, but what I found is that it is not good to cycle master power to the controllers, when the LVC starts tripping. With the new module, trips under load will still hit the ebrake input and/or pull down the throttle signal. Slow drain trips will cut pack power completely, until manually reset. The intent is that this module can be added in a setup later, and make use of existing LVC signals.The HVC signal, which can share the same opto line, is used in a similar fashion, but during the charge process. If a cell gets full quicker than the rest, it will hit that voltage "knee" where the voltage will start to rise at a faster rate. For LiFePO4 cells, this is around 3.60-3.65V, and for LiPo it is around 4.15-4.20V. With balanced cells, they all will hit this point at roughly the same time. If the charger's CV mode is set at this point, it will hold all the cells at this point, which will cause the current to slowly drop, until it gets all the way down to zero. Actually, when it gets down to something like 1/20-1/50th of the original charge current, the cell is about as full as it will get. Anyway, this is what happens when the cells are balanced. If not balanced, one, or more, cells can get full faster than the rest, and will hit the "knee" before the rest. This can cause the voltage for this cell to rise quickly to the point of cell damage, or worse, while the charger is still not in the CV current limiting mode. The HVC circuits on each channel will trip if this happens, and this signal is used by a "charge controller" to throttle back the current by however much is needed to keep the errant cell voltage from going any higher. Think of it as individual cell CV modes.On the full BMS, the charge controller logic is on the same board. When we started doing the LiPo variants, I split the charge controller circuit out into its own small box. This standalone charge controller also performed one othr function and that was to monitor the charge current and cut it off when it dropped below an adjustable set point. Recently, Richard came up with a clever way to add charge current limiting to "dumb" power supplies, like the inexpensive MeanWells. These popular supplies aren't originally meant to used as battery chargers, so most lack a true constant current (CC) mode, having a simple "hiccup" mode on the front-end, for overload protection. Richard's small widget mounts to the output terminal strip, and ties into a sense input via a single hookup wire that is tacked to the top of a resistor, next to the voltage adjustment pot. There is an adjustment pot that can be used to set the desired charge current from 0 to the max the supply is rated for. Since this worked so well, we then added to the current limiter functionality, turning it into a complete Charge Controller. In addition to the charge current limit, there is also another pot for setting the low current/end-of-charge trip point, which is adjustable from 0-1A. It also has the input for the HVC signal, which it will use to throttle the current should a channel's HVC circuit start tripping.So, it is possible now, I think, to make a reliable "commuter" LiPo pack that not only has individual cell LVC protection, but also can be safely bulk-charged with an inexpensive MeanWell supply. Balancing can be handled a number of ways. The DIY crowd can build one of the CellLog-based balancers, which can provide the simplest solution, in terms of ease-of-use, or there are a number of RC-type balancing solutions that can work quite well. Until recently, it was hard to get an RC-based off-the-shelf charging/balancing solution that didn't require a lot of the sort of "pack manipulation" described by Dogman. The other problem I've had with most RC balancers and/or balancing chargers, is that they don't provide much in the way of balancing current, most being limited to 100-200mA. This might be fine for most "single-p" RC packs, but for a 15-20 Ah ebike pack, it can take literally half-a-day to balance a pack. Our original "full BMS" unit managed 500mA, and the newer v4.x version, and the new CellLog-based balancers, all handle a bit over 1A.I have, however, just gotten a hold of a couple of new Hyperion balancing chargers that have a number of useful features for us. For one, they can handle pack setups up to 14s LiPo, which makes it quite simple to use with a "typical"12s LiPo commuter pack. Another improvement in these, over most other RC units I've used, is that it has 500mA of balance current, and uses 12-bit A/D converters for greater balancing accuracy. Finally, the input voltage can be as high as 30V, which means I can drive one of these from the same S-350-24 MeanWell supply I use for bulk charging. They have a max charge rate of 20A. Here's a link to info about these chargers:I'm still a Hyperion dealer, so if these work out, I'll start offering these on my site, with a big discount for E-S members. I need to do some connection adapters, but there should be no reason you couldn't plug this into the same balancer plugs coming from the LVC/HVC boards.Not sure I've cleared anything up here, or muddied the waters even further, but I will say that I don't give a rat's ass about cricket! :lol:-- Gary

Contact us to discuss your requirements of bms 2.0. Our experienced sales team can help you identify the options that best suit your needs.