what capacity cells would be in 36v 17ah battery

[Nealh]

Sorry to but(t) in (dreadful pun), but when you lot aren't having a go at each other you seem to know exponentially more than I do about batteries (though I'm quite adept at hijacking threads). So, at risk of causing more flaming, what battery would you put on a torque sensor Ed crank drive (tsdz2) that's nominally 500w but in reality up to 700 (i.e. at 36v about 18 amp)? In other words what do you thinks the best 36v 18a bottle battery out there? Many thanks

The cells inside will make a difference.
Top cells are LG MH & MJ1, Samsung 30Q & 33G.

 

[Steb]

Cos Telsa restrict the SoC -- lowest discharge / highest charge -- to between 20% & 80% -- you need their special permission/remote authorisation to charge higher than 80%.

That means they don't (cannot) top balance.

I'll leave the battery choice for someone else to help, since I tend to not buy 'off the shelf' packs. I would however recommend sourcing from within the UK if after sales service is of any importance to you.

Tesla packs achieve such a long service life due to a number of measures.

Firstly, as NJS correctly states, they only use 20-80% of battery capacity in regular daily usage (unlike most ebike batteries, their BMS deals with this practice in the correct way).

Also, they use active (fan) cooling to keep the pack at optimum temperature.

Additionally, they also use pack heating to bring the cells UP to temperature when operating in 'ludicrous mode' This is to make sure that the pack has the lowest possible internal reactance (IR), so that it can dump HUGE amounts of power with the very minimal of cell degradation.

Lastly, since they own and control the entire cell manufacturing process, they can ensure very high levels of quality control.

EDIT: Ebike batteries have the additional problem that they are much more price sensitive and as a result, compromises are often made that stress the pack far more than is ideal. Also, we often hear of packs that have been left to sit for a long time without usage, which causes issues. Electric cars don't suffer from the same usage issues as they tend to get used much more frequently. When it comes to the majority of lithium ebike batteries, it's definitely a case of "use it or lose it"!

It does appear that SOC isn't 100%, to over come this they spec and use a bms with lower balancing voltages. Also I suspect the lower discharge voltage level is higher .

Many thanks for the useful informative replies. One snag is that it doesn't seem acceptable to me to shell out £500 every 400 or so cycles on a new battery. It makes the cost per mile higher than practically any other transport. But I guess there isn't any alternative (and sincerely don't mean to reopen the debate about SOC)

 

[Steb]

The cells inside will make a difference.
Top cells are LG MH & MJ1, Samsung 30Q & 33G.

This is very useful, I'll try and find out how possible it is to build a bottle battery with these cells and a bms

 

[danielrlee]

Many thanks for the useful informative replies. One snag is that it doesn't seem acceptable to me to shell out £500 every 400 or so cycles on a new battery. It makes the cost per mile higher than practically any other transport. But I guess there isn't any alternative (and sincerely don't mean to reopen the debate about SOC)

Firstly, there's no problem in debating or discussing SOC. It's not a dirty word and a partial charge is good practice in certain scenarios. It's making blanket-wide recommendations without understanding the issues it can cause that can cause problems for others.

You are correct that a lithium battery can be expensive (although it's all relative) and the cost often makes up a significant proportion of an ebike. However, it's all about amortization. You have to pay the up-front cost of a battery whatever the case, but the more miles you travel under its power, the cheaper the cost per mile gets. I've heard of somebody getting over 20,000 miles from a garden variety 18650 pack over its whole life. I'm not sure how much they paid for the pack, but it has to figure pretty cost effective.

 

[Nealh]

The relative price of a battery can be expensive though if used regularly every week for 3 or 4 years then you can deem it as a cheap and inexpensive power source and you will have certainly got your monies worth form it.

 

[NJS]

This is why we shouldn't recommend the practice of undercharging in this context.

Which is something I agree with; and would never do. And certainly never did!

I mentioned in passing, that I prefer to undercharge. And the D8 rushed in frothing at the mouth; and you joined the bandwagon.

You cannot have a discussion with D8. You either take his edict unquestioningly; or he tries to use his forum weight to put you in your place. His way or the highway, to coin a phrase.

And as for your intervention given:

I'm pretty sure they do 'top balance' though, although I'm not totally certain.

You cannot 'Top Balance', if you are not at the top. Full stop.

Another possibility is 'active balancing' that constantly occurs throughout the charge range, although now I'm just speculating.

In order to 'balance' at a point other than the top or bottom, you need to know capacity -- per cell or per parallel group capacity -- and then coulomb count.

To know capacity, you have to start -- preferably from new -- by bottom balancing.

Only then, by having a known starting point, and by counting them in, counting them back out again, can you distinguish between cells that reach a higher voltage earlier because of a naturally higher IR -- they vary slightly even with the same batch -- and a one that is getting there early because it is loosing capacity due to any one of several forms of damage.

If you know the individual cell(group) starting capacities, then you can do a from of balancing called 'predictive balancing'; but that requires accurate coulomb counting and substantial MCU & memory resources.

No bike battery comes with a BMS capable of this.

To reiterate:At no point have I, nor would I, ever advocate, recommend or otherwise influence anyone to undercharge.

But be aware: Many pre-built e-bike battery pack come with BMSs that either do no balancing; or the very simplest -- and long discredited in academic papers and professional companies -- of top balancing based upon OCV leveling. (As opposed to SoC leveling.)

The problem with that is that bad cells achieve HVC first, you then drain them back and pack charge again. Again they see HCV first. Rinse & repeat. The upshot is that the worst cell in your string spend longest at HVC, and effectively received a disparately high number of charge cycles. The net result is that top balancing based solely upon OCV ends up exacerbating the weakest cell (group) and thus accelerating the demise of the entire pack.

 

[danielrlee]

Which is something I agree with; and would never do. And certainly never did!

I mentioned in passing, that I prefer to undercharge. And the D8 rushed in frothing at the mouth; and you joined the bandwagon.

You cannot have a discussion with D8. You either take his edict unquestioningly; or he tries to use his forum weight to put you in your place. His way or the highway, to coin a phrase.

And as for your intervention given:


You cannot 'Top Balance', if you are not at the top. Full stop.



In order to 'balance' at a point other than the top or bottom, you need to know capacity -- per cell or per parallel group capacity -- and then coulomb count.

To know capacity, you have to start -- preferably from new -- by bottom balancing.

Only then, by having a known starting point, and by counting them in, counting them back out again, can you distinguish between cells that reach a higher voltage earlier because of a naturally higher IR -- they vary slightly even with the same batch -- and a one that is getting there early because it is loosing capacity due to any one of several forms of damage.

If you know the individual cell(group) starting capacities, then you can do a from of balancing called 'predictive balancing'; but that requires accurate coulomb counting and substantial MCU & memory resources.

No bike battery comes with a BMS capable of this.

To reiterate:At no point have I, nor would I, ever advocate, recommend or otherwise influence anyone to undercharge.

But be aware: Many pre-built e-bike battery pack come with BMSs that either do no balancing; or the very simplest -- and long discredited in academic papers and professional companies -- of top balancing based upon OCV leveling. (As opposed to SoC leveling.)

The problem with that is that bad cells achieve HVC first, you then drain them back and pack charge again. Again they see HCV first. Rinse & repeat. The upshot is that the worst cell in your string spend longest at HVC, and effectively received a disparately high number of charge cycles. The net result is that top balancing based solely upon OCV ends up exacerbating the weakest cell (group) and thus accelerating the demise of the entire pack.

I honestly believe that there's some fundamental misunderstanding of terminology used your post (#66) that is confusing the debate, but I don't see the point in to-ing and fro-ing with you over it.

You're happy to do what you're doing and we're happy that others aren't. Shall we just leave it there?

 

[NJS]

I honestly believe that there's some fundamental misunderstanding of terminology used your post (#66) that is confusing the debate, but I don't see the point in to-ing and fro-ing with you over it.

Are you saying you don't understand my terminology?
Or that my terminology is wrong?

we're happy that others aren't.

Is that the Royal "we"? And why does it make you happy that others are needlessly shortening the life of their expensive battery packs?

Is it because you and "your friend" have a lucrative business buying up "dud" battery packs, bypassing the BMS's to "recover" them and selling them on as new(ish)?

Shall we just leave it there?

I'd be perfectly happy if you'd butt out. I had no intention of getting into this discussion in this forum in the first place.

I casually mentions my preference in post #14: My understanding is that 41.5V is the best charge voltage for a 10s li-ion battery even when the cells are new; it's where I'm charging to anyway.
Simply because the guy I was talking to happened to mention 41.5V, which as 95% SoC happens to be the same value I use.

Then the D8 wades frothing, "correcting" me, by taking me out-of-context.

So I respond very reasonably in #16 with: "Correct for a full 100% charge I know; but I've read from reliable sources that pack lifetimes can be greatly extended by charging to less than 100%.".

And like those quotes, the rest is a matter of record. D8 starts out by telling me: "Some lithium batteries can, but not yours. " despite that he has -- cannot have had -- any knowledge of MY battery pack.

IN OTHER WORDS: HE TOOK IT UPON HIMSELF TO TELL ME **** THAT HE COULD NOT POSSIBLY KNOW!

So, I'd be very happy for you (who in your last post demonstrated ZERO understand of Li-Ion battery balancing!) to simply stop talking as if you have the right or power to tell me what I should and should not do; and what I should and should not say.

If you -- individually or collectively -- don't like what I write, don't read it; but for Dog's sake do grow up and stop it with the puerile homophobic taunts.

 

[danielrlee]

Are you Smart Ebiker in disguise?

 

[NJS]

Are you Smart Ebiker in disguise?

No idea who that is; so emphatically: NO!

 

[danielrlee]

No idea who that is; so emphatically: NO!

Hmmm.... That's what Smart Ebiker would have said.

*backs away slowly*

 

[NJS]

Hmmm.... That's what Smart Ebiker would have said.

Gosh darn! What a decisive put down. Not!

Puerility upon puerility.

*backs away slowly*

If only. You really don't know when to STFU do you.

Here ya go. Since, you want the last word, this'll be my last here, so take it.

Hopefully, it will be "sorry" to everyone who's time you've wasted; but I doubt it.

 

[danielrlee]

Gosh darn! What a decisive put down. Not!

Puerility upon puerility.



If only. You really don't know when to STFU do you.

Here ya go. Since, you want the last word, this'll be my last here, so take it.

Hopefully, it will be "sorry" to everyone who's time you've wasted; but I doubt it.

Do you usually get this worked up over an internet forum?

 

[NJS]

You seem far more "worked up" than I.

 

[footpump]

i tend to agree with steb, I would consider my ebike batteries to have a somewhat pampered life.
they are kept inside, not used in very cold weather but still
kudos 10.4ah 4000 miles, range now at about 15 miles flat terrain nownear 500 charges.
ebc 13.4 about 900 miles, range seems to be dropping about 25miles very few full charges.
yosepower 10.4 about 2000 miles about 30 miles range would be 66 full charges.

I have always looked at capacity a 2.6 or 2.9 3.ah or higher not ever thought about any other battery factor.
but if for instance one had a 13ah battery with lets say 30q cells at £320 or 2 10.4ah inferior celled ones for the same price ,
which would last the longest , I think its hard to know unless you have been able to try both options.
or inferior battery recelled when its become useless

 

[Deleted member 4366]

The older Kudos bikes, like the Safari had/have LiFePo4 batteries, which seem to be good for 2000 cycles. They definitely last a lot longer than most Li-ion ones.

My 48v 11.6 Ah Li-ion one with Panasonice cells is approaching 5000 miles and cost £250. It's definitely getting tired, but still has plenty of range. It's become quite saggy now. That works out at something like £80 per year of 5p per mile. Either way, that's fairly insignificant compared with the deprecion on buying a new ready-made bike.

Batteries are improving every year. Its probably worth buying a new battery in three years time to get the improved properties, like reduced weight or bigger capacity, never mind the deterioration of any that you have today.

 

[anotherkiwi]

My Samsung 26F 10.4 Ah bottle cost so far about 0.06 €/km which is fine by me, much cheaper than taking the bus.

In the coming year I am going to get a welder and start to make my own batteries. Using HK Graphene LiPo was my first choice but they never have the size I need in stock in the EU store. I will rebuild my bottle battery with Samsung 30Q cells so 36v 12 Ah and no sag as a goal.

I want to use the Sanyo 20700 cells on my trike, these are also 15 A constant discharge and +4100 mAh. I am thinking about making 5P6S batteries which I will balance charge with my HK Reaktor chargers (in Li-Ion mode) and use in series so a 44.4v >20 Ah total battery. LVC using LiPo alarms set to 3.3v and the controller cut off set to its minimum 38v (3.16v per cell) as a fail safe. For 500 full charge cycles the cost per km works out to about 0.02 €, so being pessimistic let's say less than 4 cents.

 

[Nealh]

Samsung have the 20700/21700 48G 4800mah 9a rated cell but have not seen any for sale, though some one on ES has acquired some.

 

[anotherkiwi]

We have some nice roads on the other side of the border which have +20% slopes so even a small (6 Amps difference is actually more than my 26F is rated for!) difference in C rate is more important to me than 3 or 4 Ah extra capacity - 15 Amps continuous discharge wins hands down.

What price did they pay for the Samsung.

 

[Steb]

My Samsung 26F 10.4 Ah bottle cost so far about 0.06 €/km which is fine by me, much cheaper than taking the bus.

In the coming year I am going to get a welder and start to make my own batteries. Using HK Graphene LiPo was my first choice but they never have the size I need in stock in the EU store. I will rebuild my bottle battery with Samsung 30Q cells so 36v 12 Ah and no sag as a goal.

I want to use the Sanyo 20700 cells on my trike, these are also 15 A constant discharge and +4100 mAh. I am thinking about making 5P6S batteries which I will balance charge with my HK Reaktor chargers (in Li-Ion mode) and use in series so a 44.4v >20 Ah total battery. LVC using LiPo alarms set to 3.3v and the controller cut off set to its minimum 38v (3.16v per cell) as a fail safe. For 500 full charge cycles the cost per km works out to about 0.02 €, so being pessimistic let's say less than 4 cents.

I'm about to do the same. But if one Google spot welding batteries many recommend soldering for creating better connections and being safer and cheaper(if taking a bit more space). So left field as it may sound, why not solder?