Article Next leap in battery performance for ebikes?

[knut7]

Are there any batteries by any of the usual manufacturers (Bosch, Brose, etc) that are not a 10s or 13s config? Or in other words, are any of these batteries not 36V or 48V?

I make those claims based on my experience with it, which I discussed with Marco Sonderegger. He confirmed the battery being on it's limit and ramping down at low SOC. He said they didn't go 21700 because they couldn't make it fit the water bottle shape. My intention of bringing up this example was not to prove that 48V is superior, but to say it can have advantages in some situations. It allowed Specialized to build a battery as small as possible while still being powerful enough.

I explained how I understand things, seems you interpreted it as me contradicting you. As I've said from the beginning, advantages from using thinner cables doesn't mean anything in this case. That's not something I'm suddenly admitting.

I've had electrophysics classes but my engeneering degree is not electrical.

 

[csj]

Are there any batteries by any of the usual manufacturers (Bosch, Brose, etc) that are not a 10s or 13s config? Or in other words, are any of these batteries not 36V or 48V?

That depends on your definition of "usual manufacturers". A more interesting question is why you think that matters? Limiting it to only one of two voltages helps your argument none, and further limiting it to "usual manufacturers" doesn't help either. There are, however, voltages other than 36V and 48V, but again that's not relevant.

I make those claims based on my experience with it, which I discussed with Marco Sonderegger. He confirmed the battery being on it's limit and ramping down at low SOC.

Unless your "experience with it" is technical and involves actual product knowledge, and you are capable of understanding those issues better than you've demonstrated here, your experience doesn't matter.
As for Marco Sonderegger, what you claim he "confirmed" does not support your assertion (that a 10s battery couldn't be made to work). Also, my understanding is that he's a product manager. Is he an engineer? Does he know inner working details?

He said they didn't go 21700 because they couldn't make it fit the water bottle shape.

If he said that then he's disqualified from commenting. My eMTB uses 24 2170 cells arranged in a water bottle shape and attached to the frame using a water bottle cage. That's just absurd. Also, I sincerely doubt he claimed 2170 wouldn't fit in a 10s config which is the topic here. 10 cells are easier to fit than 13.

My intention of bringing up this example was not to prove that 48V is superior, but to say it can have advantages in some situations. It allowed Specialized to build a battery as small as possible while still being powerful enough.

Yes, that's what you've said, but it's wrong and I've explained why it's wrong. I've also refuted any claim that 48V has added a capability to the SL that you couldn't otherwise get.

I've had electrophysics classes but my engeneering degree is not electrical.

I would suggest, then, that you determine how IR of the pack is affected by configuration, then reconsider your positions based not only on voltage and current but IR (which is responsible for the heat you've errantly characterized) and see where things are different than you've assumed. What I've presented is first semester EE course stuff and is not at all controversial. What you've presented only makes sense when you increase voltage by increasing cell count. I understand your position, I'm trying to explain where it is wrong. Take away the extra cells and all the advantages you think you see suddenly disappear.

 

[RobertR]

And no one has any idea on amp draw these motors max draw is ?

 

[Flatslide]

And no one has any idea on amp draw these motors max draw is ?

The Dyname 3.0 motor has an advertised 1200W peak power. The pack fully charged is 54V, so allowing for voltage sag under load, say, 48V, that would be a short term current draw of 25A or thereabouts. This is one way to make 108Nm of torque . It is also a way to make the rear wheel step out sideways at inopportune moments lol.

 

[knut7]

@csj I've ridden more than 40 different emtbs, all of them 36V or 48V systems. Bosch, Shimano, Yamaha, Giant, Brose/Spesh, Dyname, Panasonic all use 36V and/or 48V. Perhaps some Chinese or less used brand offer different voltages, but basically all the bikes we discuss here are 36V or 48V. I really don't know why it matters, didn't you ask why we're limited to 10s and 13s configs?

The Specialized 1.1 SL motor will deliver 240W continous. Let's say it has a maximum consumption of 300W. I don't know the cutoff voltage for all systems, but Bosch has 30V cutoff limit. So let's just go with that number. At minimum SOC, the batteries need to output 10A to feed the motor 300W. So that should in theory work with a 10s1p battery, the Samsung 35E is rated at 8A but will deliver 10A. And the NCR18650GA cells are rated at 10A. At 10A, the Samsung 35E has delivered 1900mAh before it hits 3.0V. This is hardly satisfactory, it's just 70Wh of the theoretical 126Wh available in a 10s1p battery. This is at about 25*C. I used real world measurements of the cells rather than going by the specs sheet. Measured by a guy I've communicated a bith with. 3500mAh 18650 Li-ion cells' discharge capacity test (Sanyo NCR18650GA vs Panasonic NCR18650GA vs LG MJ1 vs Samsung INR18650-35E)

These bikes are meant to be used at lower temperatures too. When we tested the Levo SL, we had temperatures between about 8 and 15*C. And the extender battery was getting really hot. We didn't measure the temperature, but it was significantly above body temperature. 45*C or even more. That's not ideal for cycle life. And this observation is what made me ask Marco about the the choice of cells/voltage. I think his answer makes sense and I don't doubt they tried using the 21700 cells in the extender. I didn't ask if they tried building the extender at both 36V and 48V though.

 

[csj]

I really don't know why it matters, didn't you ask why we're limited to 10s and 13s configs?

I asked why it matters because it doesn't, and I asked because your response to my examples is that only 10s and 13s matter.
I chose other configs because the math is more obvious, but I can argue with 10s and 13 because, again, it doesn't matter. For you, this issue is something you are using to avoid the real discussion.

The Specialized 1.1 SL motor will deliver 240W continous. Let's say it has a maximum consumption of 300W. I don't know the cutoff voltage for all systems, but Bosch has 30V cutoff limit. So let's just go with that number. At minimum SOC, the batteries need to output 10A to feed the motor 300W.

There's no reason to believe that a system will provide its "maximum consumption" at "minimum SOC". Furthermore, we KNOW that the SL will draw a maximum current of 7A, not 10A, and I suspect you know this too.

So that should in theory work with a 10s1p battery, the Samsung 35E is rated at 8A but will deliver 10A. And the NCR18650GA cells are rated at 10A. At 10A, the Samsung 35E has delivered 1900mAh before it hits 3.0V. This is hardly satisfactory, it's just 70Wh of the theoretical 126Wh available in a 10s1p battery.

Ignoring that you've deliberately misstated the current draw of the SL so as to exceed these cell ratings @ 10s, you can build 10s1p batteries with other cells, namely a Samsung 50E which tests reliably to 15A by more reliable testers than your friend. Also, and this gets far afield, the fact that you'd choose to cite Thunderheart rather than Mooch is consistent with the problem here.
Also, I will note that in your contrived example that the range extender itself would only offer 90Wh of a "theoretical" 160Wh, hardly satisfactory. I think most readers, when realizing this, immediately understand how deliberately dishonest your example is. The range extender will get a lot more than 50% of its rated range because it is NOT used in the manner you suggest.

These bikes are meant to be used at lower temperatures too. When we tested the Levo SL, we had temperatures between about 8 and 15*C. And the extender battery was getting really hot. We didn't measure the temperature, but it was significantly above body temperature. 45*C or even more. That's not ideal for cycle life. And this observation is what made me ask Marco about the the choice of cells/voltage. I think his answer makes sense and I don't doubt they tried using the 21700 cells in the extender. I didn't ask if they tried building the extender at both 36V and 48V though.

Lower temps don't change the argument, the pack is getting hot is because it is too small. A 10s pack with larger cells would be better, not worse.
How hot the pack gets has nothing to do with pack voltage but rather the power requirements being made of the limited number of cells, something I have said here too many times already. If you haven't gotten it yet, maybe you should stop doubling down on bad information and just admit you don't understand. You continue to ignore my original discussion of IR which is definitive on this issue.

Also, you should realize that posting unverifiable comments from someone of unknown technical qualification does not make your argument stronger. The fact that you "don't doubt" Marco's unverifiable claims that 2170 can't be used in a water bottle form factor says a lot, it says that you can't be bothered to consider whether it can because doing so would be damaging to your position. Do you need evidence of how absurdly wrong that is, or can you operate a calculator enough to prove it to yourself? Here's some help:

From Wikipedia: "The standard size of bottle cage holds a bottle 2-7/8 inches (2.875") or 73 mm in diameter and five inches (5") or 127 mm tall, or with an indentation that far from the bottom for the tab on the cage to engage."

From Wikipedia: A 7 cell config can be optimally arranged a 3 times the cell diameter, a 5 cell config at 2.701 times the cell diameter.

Therefore, using 2170 cells a 13s pack requires 63mm and a 10s pack requires 57mm, both comfortably smaller than 73mm (and a cage supports larger). My 24 cell pack uses a 3 x 8 cell arrangement which requires 70mm diameter minimum YET it still fits INSIDE a Topeak bottle cage toolkit. IDK who is lying here about 2170 unsuitability for bottle cage form factors, but it's not even close to a problem. My bottle cage form factor battery is 430 Wh and 1.8 Kg, it whoops the range extender soundly.

Finally, Knut, it's clear at this point that this isn't about getting to the facts but rather about maintaining your image of expertise despite evidence to the contrary. It is not my goal, or my interest, to help you learn anything, it's my goal to provide useful information to others than might be reading threads in order to learn. Hopefully I've done that, and you can choose to learn from it or not.

 

[RCDallas]

Hey Knut7, Thanks for your videos! I enjoy the thoughts and opinions of a fellow eMTB enthusiast. About the agressive comments above from @csj I understand what you are saying and agree that a 48v system will be more efficient and the battery will charge quicker at the rated 4 amp output of the charger than a 36v system. This guy is an antagonist...

Keep up the good work and looking forward to your next videos.

 

[Waynemarlow]

@csj

The Specialized 1.1 SL motor will deliver 240W continous. Let's say it has a maximum consumption of 300W. I don't know the cutoff voltage for all systems, but Bosch has 30V cutoff limit. So let's just go with that number. At minimum SOC, the batteries need to output 10A to feed the motor 300W. So that should in theory work with a 10s1p battery, the Samsung 35E is rated at 8A but will deliver 10A. And the NCR18650GA cells are rated at 10A. At 10A, the Samsung 35E has delivered 1900mAh before it hits 3.0V. This is hardly satisfactory, it's just 70Wh of the theoretical 126Wh available in a 10s1p battery. This is at about 25*C. I used real world measurements of the cells rather than going by the specs sheet. Measured by a guy I've communicated a bith with. 3500mAh 18650 Li-ion cells' discharge capacity test (Sanyo NCR18650GA vs Panasonic NCR18650GA vs LG MJ1 vs Samsung INR18650-35E)

We are using the Samsung 35E batteries in a 13S2P and 14S2P config and are finding that they won't perform to the manufacturers spec sheets at lower cell voltages. The rated 8A is like all manufacturers documentation, in a laboratory test under ideal conditions, but in the field that 8A draw is a limitation as the battery voltage falls away really to quickly towards the end of the battery's charge and under load will trip the BMS cell low voltage levels. We are getting typically only 280 - 290 Wh with the 14S2P packs.

However we have just gone over to the Samsung INR21700-40T which each cell has a 35A output, have to say these batteries really do pack a punch and do not drop voltage levels under power levels we are using ( we limit the engines we are using to about 15A ), on the first few rides these batteries are looking really good, at 4000mah per cell, they don't hold the highest of density but the length of time the battery will hold there voltage above 2.9V is really impressive.

 

[csj]

We are using the Samsung 35E batteries in a 13S2P and 14S2P config and are finding that they won't perform to the manufacturers spec sheets at lower cell voltages. The rated 8A is like all manufacturers documentation, in a laboratory test under ideal conditions, but in the field that 8A draw is a limitation as the battery voltage falls away really to quickly towards the end of the battery's charge and under load will trip the BMS cell low voltage levels. We are getting typically only 280 - 290 Wh with the 14S2P packs.

However we have just gone over to the Samsung INR21700-40T which each cell has a 35A output, have to say these batteries really do pack a punch and do not drop voltage levels under power levels we are using ( we limit the engines we are using to about 15A ), on the first few rides these batteries are looking really good, at 4000mah per cell, they don't hold the highest of density but the length of time the battery will hold there voltage above 2.9V is really impressive.

Yes, but 40T's don't offer the same energy density that high energy cells like the 35E do, so packs with satisfactory range weigh considerably more. Cells like the 40T are more popular among the higher power DIY crowd.

Modern high energy cells are rated at 2C - 3C maximum current draw but should not be used at greater than 1/2C average draw. For higher currents, other chemistries work better, as you've just said. This is why the Spec range extender is not really a well designed pack, it is intended to be used at 1C - 2C draws but uses a cell specifically not designed to work well in that application.

 

[csj]

This guy is an antagonist...

If providing correct information is antagonistic, I'm proud to be an antagonist. Nothing isn't improved with some blind tribalism though, right?

 

[knut7]

@csj I'm claiming 10A for a 10s1p system, right? That's 6,25A for 48V, so I am slightly under-estimating the current draw at 36V. We can forget about using the 18650 cells in a 10s1p config. And Spesh claim they couldn't do it with 21700 cells, so they had to go 13s1p 18650 cells. Which was my original claim. Neither of us know what limitations they had to work with in this project. So you are free to claim they're lying.

I mention outside temperature because it's relevant to the IR / max discharge current.

 

[Waynemarlow]

You guys are forgetting about voltage, low voltage means a sluggish and poor performing motor, so do you pack a battery with Samsung 35E's that has some of the best density ( W/h's ) that can only hold its best voltage levels to 2/3rds of its rated density at the voltage level we use, or do you use the same identical 18650 size battery that can pull higher voltages 9/10th's of its rated capacity, which is less than the 35E's. On paper you would think that the 35E's would be best, in the field we are finding the old Samsung 30Q cells as good if not better, but have a lower rated capacity.

As a get me home battery you could build a 18650 battery as a 10S1P but you would have to select the battery to have probably a minimum 15 - 20 A draw rather than the 8A as the 35E's, they are available but at about only a bit over 100Wh's the cost would be prohibitive. Anyway you can get a 13S2P battery giving about 300W'h's into a small case 190mm x 70mm x 80mm, why does it have to be bottle shaped. The old argument of lighter bike ( to lift over sytles sorry umm to lift over trees ) with 2 small batteries ( one in your rucksack ) is coming back full circle from where we started, certainly the sub 20kg bikes we are using ( including a 300W/h battery ) are just so much more versatile for what we do than the 25kg monster trucks that we seem to be being offered.
There is another emerging cell format, 20700, such as the Sanyo NCR20700B 4250mAh - 15A its a tad smaller and has a pretty good draw off rating. Its only 1 mm smaller but ?

 

[Waynemarlow]

Umm but the real world is showing what both Knut and CSJ are saying to be true, that the 35E battery isn't up to snuff for a single 10S1P. You can't get around bad design but now if Spec hadn't been suckered into the max density loop ( yep I did as well, bought over 200 over the little ferkers ) and had selected a different battery with a better C rating, they would have had success but with a trade off of lower W/H's.

Why the ferkin big tirade that took all of probably an hour to write, far better to be off on the bike.

 

[wepn]

Umm but the real world is showing what both Knut and CSJ are saying to be true, that the 35E battery isn't up to snuff for a single 10S1P. You can't get around bad design but now if Spec hadn't been suckered into the max density loop ( yep I did as well, bought over 200 over the little ferkers ) and had selected a different battery with a better C rating, they would have had success but with a trade off of lower W/H's.

Why the ferkin big tirade that took all of probably an hour to write, far better to be off on the bike.

I'd say observations & from what I've read from csj, I'd say accurate ones. If you want to communicate your expertise & be a credible voice, a friendly & respectful approach is effective. Self-righteousness - really not. Batteries become secondary when you're being a dipwad.

 

[csj]

I'm claiming 10A for a 10s1p system, right? That's 6,25A for 48V, so I am slightly under-estimating the current draw at 36V.

It's hard to tell what you're claiming, but 10A for a 10s pack would not 6.25A for a 13s, it would be 7.7A.

We know the SL draws 7A max. That means a 10s1p, for the same power, would require 9A. Yes, that would be worse for 18650 cells, but there's no reason larger cells couldn't be used, regardless of what some marketing guy told you. I've shown that's the case, and have a pack that proves it is, but you aren't interested in considering that, you seem only interested in finding some argument you can win.

What's interesting here, though, is that this particular argument that you focus on proves my original point, that it's the difference in cell count that accounts for the entire difference in performance. You want to hang your hat on the fundamental difference between 13 cells and 10 as though that proves that higher voltage offers performance advantages. It does not; only more, or larger, cells does. The fact is that Spec could have dropped the voltage to 7s and used a 14 cell pack, resulting in still better performance in the same bottle cage form factor despite a lower system voltage. Also, you could actually consider the bulk of my original post(s) rather than focus on this range extender sideshow.

We can forget about using the 18650 cells in a 10s1p config. And Spesh claim they couldn't do it with 21700 cells, so they had to go 13s1p 18650 cells. Which was my original claim. Neither of us know what limitations they had to work with in this project. So you are free to claim they're lying.

No we can't and you saying so doesn't mean it's true. Furthermore, "neither of us know" why "they had to go" 13s. I would wager it wasn't because of the range extender at all.

 

[csj]

Umm but the real world is showing what both Knut and CSJ are saying to be true, that the 35E battery isn't up to snuff for a single 10S1P. You can't get around bad design but now if Spec hadn't been suckered into the max density loop ( yep I did as well, bought over 200 over the little ferkers ) and had selected a different battery with a better C rating, they would have had success but with a trade off of lower W/H's.

Why the ferkin big tirade that took all of probably an hour to write, far better to be off on the bike.

No one is using a 10s1p configuration so your point is moot. Spec uses a 13s1p config which is arguably also inadequate. For all other packs with at least 20 cells and low power usage, you should rethink your position on 35E-type cells.

 

[csj]

I'd say observations & from what I've read from csj, I'd say accurate ones. If you want to communicate your expertise & be a credible voice, a friendly & respectful approach is effective. Self-righteousness - really not. Batteries become secondary when you're being a dipwad.

It's not clear who you are directing this to, or why you think it doesn't demonstrate that you are the "dipwad". Why you think "respect" is the issue, or that it isn't present, says more about you than it does about the topic. Engineers don't play politics and people who want to learn do not benefit from deference to wrong answers. The ultimate respect is respect for knowledge, something you think should "become secondary" based on your personal feelings.

 

[Waynemarlow]

No one is using a 10s1p configuration so your point is moot. Spec uses a 13s1p config which is arguably also inadequate. For all other packs with at least 20 cells and low power usage, you should rethink your position on 35E-type cells.

Yup we worked that out some years back, the 35E's work really well in the 3P 39 cell packs which we originally built, where the amperage can be drawn across all three batteries, we just built some smaller packs for a get me home battery of the 35E's to put in our rucksacks and were a little disappointed by their performance over similar packs built with 30Q's. It wasn't until we put a logger on the packs to see why they were kicking out the BMS voltage shut downs that we realised the 35E's were not living up to specs, not by much and if you accept that, the packs still have a larger W/hs than the older batteries, you just have to dial back how much ooomph you are asking of the battery in the last part of their use, if its get me home slow or not have assistance or just dial back the assistance level, its a no brainer.

We also know we are probably getting less charge discharge cycles by abusing them this way, but battery design is changing so fast that the reduced to 500 charge / discharge cycles we are expecting, is a fair trade off.

Why 48 volts, that is a whole different subject and one that should be discussed, its very technical, but for the less technical among us, it is to get efficiency of the engine at higher cadence levels. There's a whole myriad of technicalities in there of +'s and -'s and way way above my pay grade, but the way forward is 48 volts and perhaps 52 volts ( the real Ebike gurus are now using 72volts but if you put your fingers across the terminals at that voltage then we are into a hearts stopping moment ), which is what we are using on the TSDZ2 engines we are running. The firmware on these motors now calculate total W/h's of the battery you are using ( counting Coulombs ) and have a cunning plan that when you are getting near the end of the battery, it will automatically restrict the Watts to eek out the max life of the battery, its quite cool.

 

[csj]

...the 35E's were not living up to specs...

Not so much missing specs but being used in manner for which they are not intended. The 2 - 3 C rating is the thermal limited continuous current draw, but they are not promising anywhere near capacity (or lifecycles) at that draw! So yeah, what you observe is right, if you want sustained high draw you need another cell like the 30Q. That's not just true of the 35E but all the high energy cells. Also, I made this point in the Francis SL weight-weenie thread, the range extender is a terrible choice for full time operation of the bike for exactly the reasons you're discussing.

Why 48 volts, that is a whole different subject and one that should be discussed, its very technical, but for the less technical among us, it is to get efficiency of the engine at higher cadence levels.

Yes, kind of. The manufacturer needs to decide how to build the motor, including how to construct the windings. High turn counts have some benefits to motor and controller construction but require higher battery voltages to support high cadences. This is almost certainly what drove the 48V decision in the SL, not anything Knut is arguing about the battery. The tail doesn't wag the dog.

The value of specific voltages, like 52, is esoteric and involves specific electrical devices. Such voltages aren't important to low power ebikes like are discussed here. No reason to avoid them either, of course, but 36V vs. 48V vs. 52V are just different design tradeoffs, they are meaningless to the consumer.

...and have a cunning plan that when you are getting near the end of the battery, it will automatically restrict the Watts to eek out the max life of the battery, its quite cool.

It's a neat feature but likely one that manufacturers are implementing already. Coulomb counting is effective, but you can trigger such a feature based solely on battery voltage under load, too.

 

[Waynemarlow]

The value of specific voltages, like 52, is esoteric and involves specific electrical devices. Such voltages aren't important to low power ebikes like are discussed here. No reason to avoid them either, of course, but 36V vs. 48V vs. 52V are just different design tradeoffs, they are meaningless to the consumer.

It's a neat feature but likely one that manufacturers are implementing already. Coulomb counting is effective, but you can trigger such a feature based solely on battery voltage under load, too.

Sorry should have said we already are coloumb counting.

I agree that gearing can change cadence and spin the motor in the sweet spot, but as soon as you use a lot of gearing then efficiency over all drops, computer controlled FOC changes that again, so many parameters to choose from, but then we are all learning about these great little fun machines of ours.

 

[knut7]

It's hard to tell what you're claiming, but 10A for a 10s pack would not 6.25A for a 13s, it would be 7.7A.

We know the SL draws 7A max. That means a 10s1p, for the same power, would require 9A. Yes, that would be worse for 18650 cells, but there's no reason larger cells couldn't be used, regardless of what some marketing guy told you. I've shown that's the case, and have a pack that proves it is, but you aren't interested in considering that, you seem only interested in finding some argument you can win.

What's interesting here, though, is that this particular argument that you focus on proves my original point, that it's the difference in cell count that accounts for the entire difference in performance. You want to hang your hat on the fundamental difference between 13 cells and 10 as though that proves that higher voltage offers performance advantages. It does not; only more, or larger, cells does. The fact is that Spec could have dropped the voltage to 7s and used a 14 cell pack, resulting in still better performance in the same bottle cage form factor despite a lower system voltage. Also, you could actually consider the bulk of my original post(s) rather than focus on this range extender sideshow.

No we can't and you saying so doesn't mean it's true. Furthermore, "neither of us know" why "they had to go" 13s. I would wager it wasn't because of the range extender at all.

I specifically wrote 10s1p. I did not know the max current draw was 7A ( 48V). I estimated 300W which translates to 10A at cutoff voltage for a 36V system. And based on that assumption, you're right, at cutoff voltage (39V) that's 7.7A for the 48V. 6,25Ah is at 48V.

So you assume Marco is just a marketing guy that is lying, while I can't see what he has to gain with such a lie, so I believe him. I believe his statement is reason good enough to claim Spesh had to go 48V and you don't. I never said it's impossible to build a battery based on 21700 cells that will fit a bottle cage, I said Spesh tried and couldn't do it. Chances are there are obstacles or priorities we're not aware of.

But there is no need to discuss the extender further, if you didn't wanna discuss it you probably shouldn't have brought it up.

As I've said before. You're right, I forgot about the added resistance in a 13s config. And sure, they could have built a 24V system. Noone does that though, for whatever reason. I don't know if it's because of the stator windings... getting as much copper as possible in the smallest possible motor? Never looked into that. Regardless, it seems we are currently limited to 36V and 48V systems, which has been my premise all along.

 

[csj]

I specifically wrote 10s1p. I did not know the max current draw was 7A ( 48V). I estimated 300W which translates to 10A at cutoff voltage for a 36V system. And based on that assumption, you're right, at cutoff voltage (39V) that's 7.7A for the 48V. 6,25Ah is at 48V.

So we'll just gloss over that I stated the 7A vs. 9A current draw up front and you presented alternative information which was wrong. How do you not know the SL is limited to 7A? You otherwise assert yourself as quite the insider on the SL batteries.

So you assume Marco is just a marketing guy that is lying, while I can't see what he has to gain with such a lie, so I believe him. I believe his statement is reason good enough to claim Spesh had to go 48V and you don't. I never said it's impossible to build a battery based on 21700 cells that will fit a bottle cage, I said Spesh tried and couldn't do it. Chances are there are obstacles or priorities we're not aware of.

I can't even verify that Marco said that. This is private information you are claiming that no one can verify and conveniently suits your narrative. There's no need, though, for any lie to have been told. I've seen Marco speak on video and I have no reason to believe he knows anything. He's a marketing guy and they claim stuff that's wrong all the time. Their job is to push their stuff, not to inform anyone of reasons their product sucks and could be better. 21700 fits easily in water bottle form factors, end of story.

I will point out, though, that Spec is willfully dishonest when it comes to SL marketing, claiming the same range as much more powerful bikes despite achieving that by having radically different levels of assist. By that standard every unassisted bike has unlimited range. Specialized is as dishonest as it gets with SL marketing, so by all means quote them to make your case.

As I've said before. You're right, I forgot about the added resistance in a 13s config. And sure, they could have built a 24V system. Noone does that though, for whatever reason. I don't know if it's because of the stator windings... getting as much copper as possible in the smallest possible motor? Never looked into that. Regardless, it seems we are currently limited to 36V and 48V systems, which has been my premise all along.

No, you're "premise all along" is that 48V is superior in efficiency, power delivery and charge rate, as literally anyone can verify from a simple reading of your comments that motivated my initial posts. If anyone wonders why I'm not more "respectful", it's this deliberate dishonesty on your part.

There have been and are 24V Ebike systems but there's a business reason why they aren't so common. One motor with multiple voltage ratings covers more ground but makes the 24V application pretty undesirable. Don't take my word for it, go search Amazon or Google for "24V Ebike" and brace yourself. Geesh. For a custom design, there's no reason to use 24V, and also no reason not to.

The fact is that from a packaging standpoint, series packs that have only factors of 2 and 3 are better than what's currently used. 5 and 7 are pretty bad options when it gets down to it, and 13 takes the cake for bad packaging. Low voltages, though, lead to higher currents and lower inductances, both of which can make controller design trickier. There are a number of factors that might drive choice of system voltage, none of which you seem to know.

 

[CjP]

I’ve learnt a lot of valuable battery information in this thread. It’s great to be able to read different opinions.
Would be awesome if people could keep their claws away so I can keep learning more.

 

[KeithR]

who cares if I think CSJ is a dick

Is that fair?

I don't get the impression that CSJ is uninformed on the subject, and he's right about something else, too: forums like this serve as a long-term repository of knowledge (or, sometimes, "knowledge", unfortunately): I agree with his assessment that it's important to remember people who might read these threads long after the current readership has moved onto other things, and make them useful repositories.

 

[KeithR]

I'd say observations & from what I've read from csj, I'd say accurate ones. If you want to communicate your expertise & be a credible voice, a friendly & respectful approach is effective. Self-righteousness - really not. Batteries become secondary when you're being a dipwad.

You're reading a "tone" into comments that really isn't there.

 

[wepn]

The ultimate respect is respect for knowledge

I absolutely agree - equal to the respect for others participating in the discussion.

something you think should "become secondary" based on your personal feelings.

I don't think it should - it just does when the tone gets personal. So I apologise for including the term 'dipwad' - whatever it means. It was unnecessary and unhelpful.

 

[wepn]

21700 snap together modules that resemble Lego™ bricks to create any size or shape battery would be very useful. There are several potential negatives when compared to a spot-welded pack - eg. larger equivalent dimensions but potentially outweighed by the ease of building in custom dimensions & shapes and especially with bad cell replacement.

With VRUZEND you can snap together the terminal caps, add the connective bars and then tighten the nuts and you can create whatever pack you want. Only for 18650 cells atm.

VRUZEND DIY Battery Kit

 

[knut7]

No, my premise was 36V and 48V, for good reason I think. I have built an ebike with 24V years ago, so I won't be bracing. Yeah, I could be lying about talking to Marco, I forgot about that, should have included that in my previous post.

I said in my first response I forgot about the increased IR when going from 10s to 13s. That affects the charge and discharge rate negatively for the 48V system.

 

[csj]

No, my premise was 36V and 48V, for good reason I think. ... I said in my first response I forgot about the increased IR when going from 10s to 13s. That affects the charge and discharge rate negatively for the 48V system.

It would be nice if you could simply acknowledge that what matters is the number, and type, of cells and not how they are wired.

Even in your acceptance of IR as an issue, you/re missing the point. When you go from 10s to 13s by adding cells, charge and discharge currents are unaffected but power is increased (because there are more cells). When you then normalize the cell count but keep the voltage difference, IR negatively effects charge and discharge in the 13s pack making the current less and power equal (because equal cell count). The only thing that matters is cell count, as I said in the beginning.

Finally, I appreciate those who have said that having the right information is what matters. That's all I care about, not who is right or wrong. People will come here and we don't want them to learn dogma, we want them to have good answers.

This is a thread about future battery technologies and there is one clear takeaway, battery chemistry improvements happen gradually. We haven't seen much in a long time and it could be years before we get anything meaningful, despite promises. That leaves us understanding that how well our bikes work is mostly dependent on how many cells our batteries have, and an important thing that some people fail to realize is that there isn't going to be any genre-busting bike with a teensy battery and big performance because low cell counts can't do that. Increasing voltage is a dream that says that we can get better performance out of fewer cells by rearranging some wires. That is false.

 

[KeithR]

If you don't like what I've put, please, just click ignore on me.

There's a disappointing amount of histrionic over-reaction to benign commentary on here, I've noticed.