Torq with battery mod storms up hills

[Tiberius]

The various aspects of Lithium batteries on the eZee Torq have had some coverage here. Readers of the Battery FAQs forum in particular will have twigged that I have been working on a way to tackle this. Well, here is a brief report on some of my experiments. First the results, and then I'll explain the engineering in the next post.

My Torq has a 10 Ah Lithium battery. Before the mods, I could go for 3 or 4 miles on the flat before I got the first stage of cut outs. The first stage is the controller shutting off power, and it is restored by closing the throttle. The second, and severe, stage is a complete shut down, and requires recycling the main power switch.

With the mods, I went for a test ride to see how far I could go. I was pedalling, but basically I was using as much throttle as I could all the time. After 12 miles or so on the flat, I started seeking out hills. I went up hills that would frighten cars and pedestrians. In fact one hill was so good I did it many times. I rode it up to the top of the hill, and then I rode it back down again. And when I was up I didn't get cut outs, and when I was down I didn't get cut outs.

Eventually, after consuming 9 Ah, I did get a first stage cut out. So I went back down the hill and up again, and then at 10 Ah, and 26 miles, the first stage cut outs were frequent enough to make it unusable.

I never got it to make a complete second stage cut out.

Nick
The Grand Old Duke of Torq

 

[keithhazel]

well that was funny, didnt know what you was talking about as im technically illiterate, but it was a good poetic read.

 

[Tiberius]

Details of the Mods.

Let me say at the beginning that I don't necessarily think the Lithium batteries that eZee supply have something wrong with them, or that they are somehow less good than those from other bike manufacturers. But there are consistent reports of cutting out under load and of loss of life after a low number of charging cycles. What seems to be happening is simply that they do not like high current drains and this is damaging both their short term performance and their life. The Torq, in particular, does demand high currents. And it demands them particularly when hill climbing. On a steep hill, the speed often falls below the optimum efficiency (12 mph?), yet more current is needed and so a vicious circle develops.

So I took the view that the answer was not to change the battery type - Lithium offers the best energy/weight ratio of the types available - but to try to deal differently with the peak current demands.

I set up a system that augments the 10 Ah Lithium battery with a smaller auxiliary battery. The idea is that they will share the peak currents, but because the auxiliary is only required to contribute to the peaks, it can be a much smaller battery.

Two or more battery packs can be readily combined with Schottky diodes - this allows either or both to power the motor according to the state of charge, and prevents cross currents between the batteries. With a Schottky combiner, the main and aux batteries can be different types, so the obvious type for the aux is NiMH, which can give high currents from a small size - such as AA.

Initially I was thinking of an intelligent controller circuit that rationed the aux battery. It would be no use if the aux were used up first. I also considered making the NiMH pack 29 or 28 cells instead of 30 as a way to achieve this.

But I decided the simplest thing to try first was just to add two aux AA packs. Using 2.2 Ah AA cells, with one pack, it is the AAs that run out first, but with 2 packs there is still life when the main 10 Ah Lithium is exhausted.

The whole assembly, including mounting frame, Schottky unit, connectors, etc weighs 1.8 kg.

I have devised a way of charging the two aux packs together or individually (see Battery forum for the discussion), and this can be done on the bike or off.

I also have some photos which I will try to resize and post links to.
Photos are now on view in post #18 on page 2 of this thread.

Nick

 

[flecc]

I toyed with doing exactly this for the T bike Nick, but didn't like the charging complexities so opted for a higher voltage all-NiMh system instead.

Interesting that it's worked as predicted though, so thanks for the details and confirmation.

I had the supplementary idea of using a tilt switch and switching so that the supplementary power would only cut in on inclines over a set degree, but again dropped it due to the charging complexities.

I'm not as confident as you about Li-ion superiority, they clearly can't deliver at an adequate rate, so until one can deliver as required, I see the heavier NiMh as superior just on the basis that it's simple and it works unconditionally.
.

 

[prState]

Is this arrangement something that could be designed into the Torq from Ezee without too much trouble?

It certainly seems like it would make it more attractive from a sale's point, i.e., powering up hills while retaining the other qualities.

 

[flecc]

I couldn't see any company marketing a compound battery system of this sort, too many possible complications in use and charging. The real answer is to either sort the Li-ion battery or revert to NiM h until a better solution is found.
.

 

[Jeremy]

Good experiment, Nick. It's confirmed what I thought might have been the real cause of the problem - poor choice of specific battery type. The RC model chaps discovered a long time ago that all LiPo cells were not the same, even though they had the same notional chemistry and power/energy density. The problems they initially had (which seem to be exactly the same as those you and others have found) is that some cells had a low maximum discharge current capability. The voltage would dip above about 1.5 to 2C rate. In the case of an ebike, I strongly suspect that it's the peak current that's causing the problem, even though the battery should be OK for the mean current draw at any instant.

Interestingly, this was quickly overcome by the RC model chaps by selecting cells designed for high rate discharge. They commonly use LiPo cells with a rated discharge capability of 10 to 15C now. The other thing that the modellers seem to have discovered is that there is a marked difference in voltage characteristic under load between machine wound cells and hand wound cells. Although machine wound cells seem to last longer, they have a near-immediate voltage drop under load, sustaining this reduced voltage for a long time. Hand wound cells seem to maintain their terminal voltage better, but don't last as long when subjected to repeated very heavy discharge.

Overall, I'm not convinced that LiPo cells are the best for an ebike. The very best bet by far seems to be the newer LiFePo4 cells, which now have a very good track record and seem extremely robust. The snag with them is their very limited availability and, as always, the price (at least for the two well-established cell types).

With luck, I should have my 36V, 7.4Ah NiMH pack finished by the weekend (I'm just waiting for some schottkys that are in the post) so will see how that goes.

Jeremy

 

[flecc]

I agree that Lipos aren't ideal for e-bikes, and eZee found the same on theirs when experimenting about a year ago. Their battery isn't Lipo therefore, it's Li-ion manganese, also with somewhat inferior discharge characteristics to the older Li-ion cobalt batteries.
.

 

[Tiberius]

Is this arrangement something that could be designed into the Torq from Ezee without too much trouble?

It certainly seems like it would make it more attractive from a sale's point, i.e., powering up hills while retaining the other qualities.

Well, yes and no. If you try to combine it into one battery assembly, then why not just make a better (or more suited) battery in the first place?

On the other hand designing a system that allowed two similar batteries instead of one to be fitted would be good.

I'm starting with a Lithium battery. But it always keep coming back to the same question - why not just start with a NiMH in the first place?

Nick

 

[Tiberius]

I couldn't see any company marketing a compound battery system of this sort, too many possible complications in use and charging. The real answer is to either sort the Li-ion battery or revert to NiM h until a better solution is found.
.

Hi flecc,

I agree with your second sentence. But I think the use and charging could be make simple enough for everyday consumer use. It would be perfectly possible to connect up a single power source and let some electronics manage the charging of different types. I did actually consider that, but settled on the simpler approach first.

Nick

 

[Tiberius]

Good experiment, Nick. It's confirmed what I thought might have been the real cause of the problem - poor choice of specific battery type. ..... In the case of an ebike, I strongly suspect that it's the peak current that's causing the problem, even though the battery should be OK for the mean current draw at any instant.

Hi Jeremy,

I think that's the case, and it is one of the things I was trying to test. It is clear that in my new system the NiMH's are assisting rather than replacing the Li. I would really need to have discharge curves of all the batteries to put better numbers on it, though. Reports on mileage achieved can be subjective, so I am trying to work in Ah's instead.

It would also be interesting to know what the mod does to the Li battery lifetime, but that could be difficult to determine properly.

I'm afraid I don't know enough about the different Li technologies to comment on the differences between them.

Nick

 

[flecc]

It would be perfectly possible to connect up a single power source and let some electronics manage the charging of different types. I did actually consider that, but settled on the simpler approach first.

Nick

Of course, but I'd sort of ruled out the possibility from the cost aspect, and like you, thought in terms of simpler first. I'm happy with my 39.6 volt NiMh solution, though it does seem to weigh a heck of a lot when I pick it up. The battery mounting has survived so far!

It's interesting that the 39.6 volts radically improves the overgeared T bike/Torq's hill climbing, up from 6% to 10% unassisted, but has virtually no effect on the same motor/controller in the normally geared Q bike/Quando which feels no different in speed or climbing.
.

 

[Tiberius]

It's interesting that the 39.6 volts radically improves the overgeared T bike/Torq's hill climbing, up from 6% to 10% unassisted, but has virtually no effect on the same motor/controller in the normally geared Q bike/Quando which feels no different in speed or climbing.
.

That is interesting; you'd expect the extra voltage to have its main effect on top speed. It suggests perhaps that one of the key things with lower speed hill climbing is the battery internal "resistance" affecting the current available. (I use " " marks because I'm sure there are chemical reaction effects as well as Ohmic resistance.)

If that's right then you have overcome it with more voltage, and I have overcome it with parallel batteries.

Nick

 

[flecc]

Yes, I'd done it to cancel some of the voltage drop under load to improve hill climbing and expected just a bit of improvement in that plus a proportional increase in speed. But as you see on the Torq, the hill climbing improved more than a bit, with just a slight speed gain.

Perhaps not surprising the Quando doesn't benefit much since it's got such prodigious climb ability anyway and geared as it is, it easily runs out to maximum revs anyway.
,

 

[Tiberius]

I feel I should add a comment to all this.

Flecc points out that a solution to the Li problem is to change back to NiMH. He's quite right, that is a solution to the existing problem of cut outs and battery lifetime, and it is probably the best and most pragmatic if you can get hold of the NiMH batteries.

I'm also trying to address a wider question. Why is the Li battery on the Torq a failure, when Li seem to be successful elsewhere? If my understanding is right, then there are other solutions worth considering. Such as using a much larger capacity Li battery, or paralleling up two Li batteries. That may not be an immediate pragmatic solution for most people, but it could lead to designs in the future with much better overall performance.

Nick

 

[flecc]

If my understanding is right, then there are other solutions worth considering. Such as using a much larger capacity Li battery, or paralleling up two Li batteries. That may not be an immediate pragmatic solution for most people, but it could lead to much better overall performance.

Nick

I'm sure you're right Nick, and I've been saying this too. That's why I applauded Wisper in introducing a near 14 Ah Li-ion for the 905se. Their battery supplier Lishen also uses a compound manganese-cobalt cathode, recovering some of the better performance of cobalt, while presumably still safe against the fire problems of the cobalt only versions of old (One hopes!).

Despite this, David Miall of Wisper says they still don't think this the final solution, presumably with an eye on lithium iron phosphate.
.

 

[Jeremy]

David Miall of Wisper says they still don't think this the final solution, presumably with an eye on lithium iron phosphate.

I think he's spot on, at least for now. I've spent a few hours reading up on all the various permutations of lithium chemistry and it seem that LiFePo4 really does offer the best compromise in terms of energy/power density, life and discharge performance.

The problem is one of supply and in many ways is similar to the non-availability of high capacity NiMH cells problem. However, rather than being a function of licensing restrictions prohibiting manufacture of high capacity cells (which is the case with NiMH, to block their use in electric vehicles), it looks as if the two major LiFePo4 cell manufacturers are in near exclusive deals with two power tool companies, DeWalt and Milwaukee that is similarly restrictive. You can obtain these cells for free though, if you're cunning. Just find your nearest DeWalt power tool service centre and go and ask for some dead packs. They will often have a few packs around and very often only one cell will have failed, the rest will be OK.

At least two other companies seem to be making LiFePo4 cells, the "Ocean" brand ones that seem to be on sale from a couple of eBay vendors and the very expensive LiFeBatt cells. We have yet to see how well these perform though; unlike the eMoli and A123 cells they are relatively unproven. I hope that work out to be good, as I have a potential project that could make very good use of cells like this, if the price was right.

Jeremy

 

[Tiberius]

Here come the photos at last

36 V AA Cell NiMH Battery Pack

Installation on bike with two auxiliary battery packs

Detail of the mounting frame, with Schottky diode combiner

Batteries and cover in the mounting frame

Total weight with two auxiliary batteries: 1.8 kg.

 

[flecc]

Very neat Nick. Keep us informed of it's ongoing performance.
.

 

[JamesC]

Details of the Mods.

Let me say at the beginning that I don't necessarily think the Lithium batteries that eZee supply have something wrong with them, or that they are somehow less good than those from other bike manufacturers. But there are consistent reports of cutting out under load and of loss of life after a low number of charging cycles. What seems to be happening is simply that they do not like high current drains and this is damaging both their short term performance and their life. The Torq, in particular, does demand high currents. And it demands them particularly when hill climbing. On a steep hill, the speed often falls below the optimum efficiency (12 mph?), yet more current is needed and so a vicious circle develops.

Nick

Nick
This is an excellent approach to alleviating the problems of high current demand on the Li-ion battery. Thank you very much for sharing it.

If I have understood correctly:
The standard battery that you are using in the trials had already reached the point where it was yielding up very little of its remaining capacity (3 to 4 miles on the flat compared with say 28 miles when new ?).

In normal configuration, the high currents damage the pathway out of the battery, causing the pathways to become narrower with usage, and therefore less able to empty the battery on the next ride, before everything clogs up.

Using the small additional NiMH battery to absorb the current spikes, you have been able to make use of some 20 miles of capacity in that old Li-ion battery that had become inaccessible due to the "narrowing of the arteries" (which I can understand).

So now you are assessing how much performance and longevity could be achieved with the Ezee Li-ion battery if the "peak absorber" were fitted from the outset with a new battery.

It has the hallmarks of being excellent, doesn't it ?


If you are successful in restoring the number of charge cycles of the main Li-ion battery, would you expect the small AA NiMH packs to have a similar life ?

James