an ebike battery is not for life, it is for a couple of years.

 

Colin

 

Perhaps a bit gloomy Colin. I'm still using the Phylion test battery I got from Wai Won Ching of eZeebike in January 2008, so three years old. For about a year I haven't been able to use full throttle continuously, onkly short term, but that's not a problem with my powerful motor. I'm still getting an average e-bikes power output, something in the order of 400 watts max.

 

I don't know the range, but it covers what I normally do throughout the local area so would be useful to many. Of course a hard riding long distance commuter would have found it useless at two years, but the great majority of e-bikers are just utility users doing shopping and local leisure trips.

 

The enthusiasts that this forum has as members are the exception, not the rule. As I observed to someone else recently, in the four years that this forum has existed we've gained some 4000 members internationally, but in that time over 100,000 e-bikes have been sold in the UK alone and no-one knows how many millions internationally. Kind of puts things into perspective.

.

Flec,

Interesting graph. I cant read the time scale units. If in seconds this represents a 30 minute ride. So how long are those 400 watt+ spikes? At that graph scale possibly each spike represents several turns of the crank from a burst of rider pedalling. If that is the case with what time constants does the control system regonise the power is above the required 250 watt average and bring it down to that?

I don't know the time units since I didn't carry out that measurement, but the control system does not regulate the power down to 250 watts as you suggested. As I said before, the 250 watts is a notional average figure which could easily be exceeded longer term with a strong rider continuous input accompanied by a slow cadence and high power setting.

 

It just happened that the average rider who took that measurement averaged about 250 watts or a trifle below. The variations were due to the rider input changes with the hilly territory of the ride.

 

"Slugging" in the circuit has probably wiped some of the detail, since each pedal stroke gives a distinct rise and fall of power from as high as the settings and pressure allow down to a low point. Oddly enough this does tend to vary bike to bike, some showing a slight tendency for the power to run on while others can have distinctive responses to the input.

.

Edited January 22, 201115 yr by flecc

Flecc I'm struggling with your figures, perhaps I've misread them. 400w or even 500w peaks on the Panasonic system means a battery draw of 15 to 20 amp draw IE: a 1.5 to 2C draw from the battery. This is not out of the norm and all but the cheapest of batteries are capable of sustaining this amount of current draw, in fact the limit is more than likely set by the BMS. So this should not in practice have a significant effect on life.

 

In practice BLDC controllers employ PWM to control motor voltage and by consequence will limit battery current by programming to protect the battery, so I don't think the factors listed previously have a large affect (if any) on battery life.

Quite a few batteries will gain abit of capacitry over the first few cycles, if you don't cycle fully, it will probably just take more. Lead acids really don't like being run down, others are better, but have their limits, lithium about 3V, NiCad and NiMH about 1V.

 

Battery conditioning is an overcharge to even out the cells, most chargers do this automatically on a full charge now, my DeWalt drill (NiCad/NiMH) needs to be left for a while after 'full' though, probably due to it being fast charger.

 

The battery cycling thing stems from one satellite running NiCads and a badly designed BMS, cycling batteries is NOT REQUIRED for any chemistry if they are used properly, that satellite discharged the batteries exactly the same amount each day, but cut off at too high a voltage, if it had run down to 1V per cell, there would not have been a problem and none of this cycling nonsense. /rant

 

Lead acid batteries are simple, they are best kept fully charged at all times. Lithium ones last best at medium charge in storage, full charge degrades them faster, running them low possibly kills them faster so if you do run them low, charge fully, if you take a week to flatten them, charge less often, it is hard to give exact guidelines as batteries and use vary so much but I'd be inclined to charge them at 1/2 - 2/3 discharged if not hammering them. Try the manufacturers data sheet for your exact battery. I'm not entirely sure about NiCad and NiMH but autocycling chargers annoy me greatly as it reduces life...

Flecc I'm struggling with your figures, perhaps I've misread them. 400w or even 500w peaks on the Panasonic system means a battery draw of 15 to 20 amp draw IE: a 1.5 to 2C draw from the battery. This is not out of the norm and all but the cheapest of batteries are capable of sustaining this amount of current draw, in fact the limit is more than likely set by the BMS. So this should not in practice have a significant effect on life.

 

In practice BLDC controllers employ PWM to control motor voltage and by consequence will limit battery current by programming to protect the battery, so I don't think the factors listed previously have a large affect (if any) on battery life.

 

I agree, my expression "giving the battery a rough time" was relative, in fact that Agattu system is capable of over 900 watt peaks and tested at that. My reference concerned longest battery life though, and as you know, when looking for extended life from lithium batteries it's best to aim for using the central area of charge between 20% and 80% of content, and using it at the lowest C rate possible. In that sense 15 amps and more from a 10 Ah battery probably isn't commensurate with getting the longest possible life.

 

Personally I just use these batteries however it's most convenient and accept the replacement costs and frequencies. I think that's the best way and if someone doesn't feel it acceptable, it's probably best they don't use an e-bike.

.

Edited January 23, 201115 yr by flecc

 

The battery cycling thing stems from one satellite running NiCads and a badly designed BMS, cycling batteries is NOT REQUIRED for any chemistry if they are used properly, that satellite discharged the batteries exactly the same amount each day, but cut off at too high a voltage, if it had run down to 1V per cell, there would not have been a problem and none of this cycling nonsense. /rant

 

All very well for the chemistry, but bad advice for some systems with battery integral meters. The Panasonic battery integral battery meter circuit like some others requires re-zeroing from time to time by running the battery to almost empty. On the latest Panasonic e-bike batteries, which have a fast flashing LED point to indicate the meter zeroing point for empty, that's also vital for the capacity measurement function since both content and capacity measurement are very dependent on correct zeroing for accuracy. It's easy to get to a 20 to 40% reading inaccuracy otherwise, especially after a cold weather spell with lowered chemical performance.

.

Ah, OK Flecc, thanks for the clarification.

 

 

 

I agree, my expression "giving the battery a rough time" was relative, in fact that Agattu system is capable of over 900 watt peaks and tested at that. My reference concerned longest battery life though, and as you know, when looking for extended life from lithium batteries it's best to aim for using the central area of charge between 20% and 80% of content, and using it at the lowest C rate possible. In that sense 15 amps and more from a 10 Ah battery probably isn't commensurate with getting the longest possible life.

 

Personally I just use these batteries however it's most convenient and accept the replacement costs and frequencies. I think that's the best way and if someone doesn't feel it acceptable, it's probably best they don't use an e-bike.

.

All these rules of thumb (charge 20-80%, keep below 2C discharge rate) feel like they may not always be true now. We've got 2 chemistrys out there and two or three cell types out there, all with different characteristics. I'm not at all sure the same rules apply to a A123 LiFePo Cell capable of 20C vs a LiMn prismatic cell capable of 2C. And so on. I also feel that a built in BMS supplied by the manufacturer ought to enforce good practice limits, though of course commercial reality may mean that it's programmed for performance rather than life span.

 

All these rules of thumb (charge 20-80%, keep below 2C discharge rate) feel like they may not always be true now.

 

But these are really fundamental, though the optimum figures may be arguable.

 

We are dealing with chemical reactions, and the fiercer the reaction, the shorter the life, regardless of what chemical reaction is under consideration.

 

It follows that the C rate be kept low in both directions and the reaction (charge content) limits be avoided since they are in reality forced high stress and exhaustion points.

 

Regardless of the chemistry, satellite practice has shown how ten year lives are obtainable with any type using these rules, and the Toyota Prius has long put that into commercial practice with their now dated NiMh chemistry having a guaranteed 8 year life.

 

I'm wary of examples and comparisons using tool and RC model batteries since their working conditions are usually so radically different from e-bike usage.

.

Edited January 23, 201115 yr by flecc

All very well for the chemistry, but bad advice for some systems with battery integral meters. The Panasonic battery integral battery meter circuit like some others requires re-zeroing from time to time by running the battery to almost empty. On the latest Panasonic e-bike batteries, which have a fast flashing LED point to indicate the meter zeroing point for empty, that's also vital for the capacity measurement function since both content and capacity measurement are very dependent on correct zeroing for accuracy. It's easy to get to a 20 to 40% reading inaccuracy otherwise, especially after a cold weather spell with lowered chemical performance.

.

Point, but if it needs cycling every charge, the pack is dying or the meter is naff, I'd rather stretch the life of batteries with a full discharge/charge cycle every 10 or 20 charges if I had a meter. Meters should reset to full on a full charge from whatever state they started, everyone I've seen does, though none were on e-bikes, a full discharge updates the actual capacity, some meters can make a very good estimate purely from voltage/current/temperature. Don't e-bike meters compensate for temperature? I build chargers that compensate for temperature, it adds a few more cycles to batteries, many don't, but it's not hard and doesn't add much to the circuits, simple compared to measuring content.

No, please don't misread this pictsidhe, as said "from time to time", and that wouldn't be anything like the 10 to 20 charges which used to be advised for NiMh some years ago. This is only for when the current Panasonic lithium battery meter has obviously drifted out of register, which will possibly be around four or so times per year.

 

It's not worth bothering to correct it during cold weather extremes, best left until temperatures normalise when the metering becomes more consistent. I've no idea on whether temperature compensation would be worth bothering with in terms of cost or complexity, but trust to Panasonic's knowledge in this battery area, particularly since Japan has extremes of temperature across the three main islands of that country that exceed ours in the UK.

 

I can assure you from experience that e-bike battery meters do not reset to correct at full charge, and the Battery University website also disagrees with you on that point, in their case regardless of battery type. Zeroing is done at full discharge, not full charge, and that is what Panasonic have designed for and which is demonstrated in practice.

.

Edited January 23, 201115 yr by flecc

But these are really fundamental, though the optimum figures may be arguable.

 

We are dealing with chemical reactions, and the fiercer the reaction, the shorter the life, regardless of what chemical reaction is under consideration.

 

It follows that the C rate be kept low in both directions and the reaction (charge content) limits be avoided since they are in reality forced high stress and exhaustion points.

 

Battery design determines the acceptable charge/discharge, most happen to be designed for similar rates though, A123 really stand out from other lithiums, they are designed for high current, it's the cost/life/performance tradeoff at work. Ohmic losses and the current density determine what reaction rates are useable with the chemistry equal, if you double the surface area of the electrodes, you can run double the current at the same current density, though ohmic losses could then be an issue, shorter cells are better for this. Some years back, 90's I think, there were some foil lead acid batteries, these had electrodes only a few thou thick and an insane rate capability, I think they took a few drag records before the company went bust.

I'm wary of examples and comparisons using tool and RC model batteries since their working conditions are usually so radically different from e-bike usage.

.

RC batteries are hammered, I wouldn't make a comparison, but could be interested in the cells for a high rate pack. Tool batteries are nearer to our playground but still want stretching a bit more, they aren't usually run for as long, so heat wil bend any comparison quite a bit, 100-200 cycles seems good for tools, working out how many more cycles you can get at a lower but more sustained rate is still going to be like throwing darts blindfold though...

Yes, full agreement there, intermittent usage can be misleading in relation to our e-bike needs. Although I can be fairly cynical about battery advances, I'm quite sure that the better e-bike manufacturers are sincere in wanting to use the best available to get marketing advantage. Wisper I know have changed battery manufacturers at least three times in pursuit of the best and work closely with them regarding present and future plans.

 

Equally eZee Kinetics have paid an extraordinary amount of attention to batteries, changing suppliers four times and currently giving an excellent choice of sizes, mountings and two chemistries.

 

Panasonic of course are a world leader in rechargeable batteries and since taking over Sanyo have a probably unique aggregation of battery knowledge in one place.

 

For these reasons I'm reluctant to consider that there are better ways than present practice as many members seem to feel.

.

I can assure you from experience that e-bike battery meters do not reset to correct at full charge, and the Battery University website also disagrees with you on that point, in their case regardless of battery type. Zeroing is done at full discharge, not full charge, and that is what Panasonic have designed for and which is demonstrated in practice.

.

I was having trouble finding fuel gauge algorithms on the web, so had a look at the Battery University, interesting website, it agrees with me:

While SoC information displayed on a battery or computer screen is helpful, the fuel gauge resets to 100% each time the battery is recharged, regardless of the battery's SoH.

Maybe I wasn't clear, when I said resets to full, I meant resets to as full as the battery can be, not maximum as new capacity, if it only holds 60% of it's capacity, it will still show 100% on full charge, and reset the fully charged point there, most gauges will only reset the fully discharged point at full discharge as it is very tricky to do otherwise, the military may do otherwise, but if the battery isn't dying rapidly, only reseting one point isn't going to cause much error over the next 10-20 cycles unless something bad/weird is going on, 100 cycles probably will put it out of whack am I making sense yet? I don't know what the Panasonic ones do but I also don't doubt their ability to make quality functional products. Maybe e-bike meters are different, but the resets to full on full charge and resets to empty on empty is standard everywhere else I know of.

Understood, we may have been at cross purposes, referring to somewhat different factors it seems. The Panasonic batteries fail to zero their meter for capacity or content unless taken down the fullest discharge as indicated by the fast flashing LED indication when close to the low discharge voltage cutout. This is preceded by a slow flashing indication commencing when there's about 10% left, but discharging only to this point fails to give a correct meter reset.

 

The bold letter highlight that I've added is what I was referring to from the Battery University website:

 

Short discharges with subsequent recharges do not provide the periodic calibration needed to synchronize the fuel gauge with the battery's state-of-charge. A deliberate full discharge and recharge every 30 charges corrects this problem. Letting the battery run down to the cut-off point in the equipment will do this. If ignored, the fuel gauge will become increasingly less accurate.

.

The better manufacturers are indeed close to optimum, although there is always the possibilty someone will come up with something radically different, but the current knowledge applied properly milks the most from current cells. The $areyoujoking military cells will blow most if not all e-bike cells away (look at the Saft Li-ions and LiFePOs, try not to drool too much), but most of us don't have money to burn or the necessary security clearance. The technology trickle down does help us though. The Saft cells are also interesting as they have quite a range of rates, what balance of performance do you want? You pays yer money and takes yer choice.

 

You pays yer money and takes yer choice.

 

Exactly!

 

I couldn't see any of today's e-bikers willing to pay more for their batteries, they are already in a state of near revolt against present prices.

.

Recalibrating every 30 charges sounds good to me! From that, I'd assume only pack degradation needs accounting for, I'd guess the meter would only 5% out at the most by 30 nearly cycles, testing one would give a better idea, which people tend to do anyway.

Yes, that's why I'd advise not discharging so often on lithium types since such a small error isn't worth correcting at the cost of shortening battery life.

.

Exactly!

 

I couldn't see any of today's e-bikers willing to pay more for their batteries, they are already in a state of near revolt against present prices.

.

Yes, the prices are pretty painful, and the reason there seem to be a lot of substandard budget batteries, I found what looked like some good cells listed on fleabay, a few google searches later I found someone who'd bought a few and tested them to death, in 30 cycles, starting at under 1/2 the claimed capacity...

Ouch! That's even worse that the worst I'd previously found. Where cells are concerned there really is some worthless junk on the market.

.