Charging pack to 41V experiment

[WheezyRider]

I haven 't got a theory. All I want to know is how much my customers would lose in capacity with a charger set to 41V instead of 42V. If it's 3% or less, it's worthwile, if it's 7% or more, I'll pass.

It's around 7 to 8%. It is noticeable, but it's not a serious loss in my case as I wouldn't start off with buying a pack size of less than 15Ah. I doubt it would be an issue for most users unless they have a small pack and need to get every last ounce of range out of it, or they are long distance riders and need 40 miles plus on a charge. It does prolong life, perhaps as much as doubling it.

However, if you go from 400 cycles to 800, will the average user get much benefit out of that? Say the average person recharges 2 times a week, that would be 4 years before reaching 400 cycles and even then, the capacity is still expected to be 70% or so of the original, so probably still giving useful service beyond that. Do you get many people coming back and telling you the life of the pack had not been as long as they were expecting? How long do people tend to keep bikes before replacing them?

For most people, it's probably not worth the bother. But you may have high use customers who will get through 400 cycles in a year and still want good capacity left after that, who would appreciate longer pack life.

 

[Nealh]

40 miles is short ride to me.

 

[StuartsProjects]

However, if you go from 400 cycles to 800, will the average user get much benefit out of that?

Reducing the charge termination voltage does reduce the degradation or reduction in capacity of the lithium cells, heaps of evidence for that.

So if you charge to 41V, which whilst initially results in the battery having a slightly reduced capacity, its possible that after 4 years and 400 charges that battery will have more capacity than one that was charged to 42V.

 

[saneagle]

Samsung 30Q did 3090 mah down to 30.0v at 1A discharge rate. That was a bit more than I expected. 98/3090 = 3.17%.

 

[WheezyRider]

Reducing the charge termination voltage does reduce the degradation or reduction in capacity of the lithium cells, heaps of evidence for that.

So if you charge to 41V, which whilst initially results in the battery having a slightly reduced capacity, its possible that after 4 years and 400 charges that battery will have more capacity than one that was charged to 42V.

I totally agree, and that is what I've seen in my experiments. But in our consumerist world most people don't seem capable of looking much further ahead than next week. Tell them they will have to recharge a little more frequently and get a little less range so that in 3 or 4 years their pack will still be in tip top shape and they will probably walk away and look to the brand that exaggerates range etc and ignores longevity.

I wouldn't be surprised if a lot of customers wouldn't even think of keeping their bike beyond 3 or 4 years.

Also, there is the broader issue of having to explain to a customer about cell balancing, which is a potential problem (although very small for decent, matched cells). This is something that may put off the average customer who is clueless about such technical matters.

As a retailer, you have to make the decision according to the customer base and what you think they can handle.

 

[WheezyRider]

40 miles is short ride to me.

You would certainly be in the "high use" category

Also you would be an informed user who would understand all the nuances involved in prolonging battery life.

 

[Sturmey]

Reducing the charge termination voltage does reduce the degradation or reduction in capacity of the lithium cells, heaps of evidence for that.

Lygte has some extra data on effect of charger termination current (which also has to be taken into account) below on a sample cell to overall capacity.

Battery charge voltage and current

 

[AntonyC]

Below is a test by Lygte-info. The result is the same. The lowest discharge rate curve crosses 4.1v at 0.1Ah - exactly the result I got. 0.1/ 2.8 = 3.5%

Maybe that steep slope above 4.1V is unique to the 30Q cell, that might account for it losing 3% while other cells lose nearer 10%.

Charge voltage experiments with lithium ion batteries showing how capacity varies with charge voltage and higher cycle live with lower charge voltage

Lithium battery charging data (voltage versus capacity) for equipment designers from PowerStream

www.powerstream.com

 

[WheezyRider]

When I charge to 41V, I usually let the termination current drop to almost zero. I feel quite comfortable doing that at 41V, but I would not like to be leaving the battery sitting at 42V for long periods waiting for saturation to occur.

 

[WheezyRider]

I totally agree, and that is what I've seen in my experiments. But in our consumerist world most people don't seem capable of looking much further ahead than next week. Tell them they will have to recharge a little more frequently and get a little less range so that in 3 or 4 years their pack will still be in tip top shape and they will probably walk away and look to the brand that exaggerates range etc and ignores longevity.

I wouldn't be surprised if a lot of customers wouldn't even think of keeping their bike beyond 3 or 4 years.

Also, there is the broader issue of having to explain to a customer about cell balancing, which is a potential problem (although very small for decent, matched cells). This is something that may put off the average customer who is clueless about such technical matters.

As a retailer, you have to make the decision according to the customer base and what you think they can handle.

One thing I think is very worthwhile to do as a retailer is to check the output of the chargers you sell do not exceed 42 V - using a calibrated DVM. Plus, if some supplies are outputting below 42V, are they so low that balancing could not occur?

 

[Woosh]

Balancing is automatically performed when the voltage is within 5% to 10% of the maximum, eg 41V-41.5V, depending on the balancing circuit trigger setting of that particular battery and stops when the delta value is less than 10mV.

 

[WheezyRider]

Balancing is automatically performed when the voltage is within 5% to 10% of the maximum, eg 41V-41.5V, depending on the balancing circuit trigger setting of that particular battery. And stops when the delta value is less than 10mV.

Depends on the specs of the BMS you have in your packs. In my experience, most don't start balancing until 41.8V. Do you have a spec sheet for them?

 

[StuartsProjects]

All I want to know is how much my customers would lose in capacity with a charger set to 41V instead of 42V. If it's 3% or less, it's worthwile, if it's 7% or more, I'll pass.

If the capacity loss were 7% or more, I can see why customers would not be happy, 7% less range etc.

I would suggest you check your batteries for yourself, using a testing method that reflects how users would charge and use their batteries.

I have just been testing one of those electronic load testers, it cost £28.




The display is a bit dim, but there is a USB connection and a Windows application which allows you to see whats happening on screen, saves a CSV file of the data too;



You set the discharge current, say 2A for a typical use scenario, and a cutoff voltage and start it running.

The test method that represent how the comparison between 42V or 41V chargers is achieved should be clear;

Charge battery with 42V charger.
Rest battery a bit
Run discharge capacity test
Rest battery a bit
Charge battery with 41V charger.
Rest battery a bit
Run discharge capacity test
Compare capacities

Now, I dont have an actual 41V charger, so I put a silicon and schottky diode in series with the output of my 42V charger, this allowed the charger to complete the constant volt stage of the charge and at the end of this the battery voltage was 41.04V.

The process used shows how easy it is to test the capacity of a battery that has actually been charged with a 41V charger as apposed to some attempt at a simulation using a 42V charger.

 

[saneagle]

Maybe that steep slope above 4.1V is unique to the 30Q cell, that might account for it losing 3% while other cells lose nearer 10%.

Charge voltage experiments with lithium ion batteries showing how capacity varies with charge voltage and higher cycle live with lower charge voltage

Lithium battery charging data (voltage versus capacity) for equipment designers from PowerStream

www.powerstream.com

Every lithium cell I've ever seen has that characteristic flip up in the discharge curve at the start of the cycle, the same as it goes steeper at the end..

 

[Woosh]

Every lithium cell I've ever seen has that characteristic flip up in the discharge curve at the start of the cycle, the same as it goes steeper at the end..

may it be beneficial to avoid that characteristic 'flipup' ?

 

[saneagle]

may it be beneficial to avoid that characteristic 'flipup' ?

About as much as it is for you to avoid a spring flip up in sales.

 

[WheezyRider]

may it be beneficial to avoid that characteristic 'flipup' ?

Yes, the higher the voltage, the more the strain on the electrode/electrolyte interfaces. When the voltage increases suddenly it means very little capacity is being added and energy is instead beginning to drive different electrochemical reactions at the interfaces, which are detrimental to the electrode. The electrode interfaces are protected by a passivating layer, which is only atoms thick. Once the applied potential goes much above 4.1 V/cell the passivating layer starts to be broken down.

Ultimately, if the potential is allowed continue to rise, this will lead to the plating of lithium on the electrodes if the applied potential gets as high as 4.35 V /cell. Then fire is quite likely.

Another thing to bear in mind is that for cells in a string, if they are not so well balanced, some cells may hit 4.25V, even if the BMS is doing it's job properly, while the others catch up.

 

[WheezyRider]

If the capacity loss were 7% or more, I can see why customers would not be happy, 7% less range etc.

I would suggest you check your batteries for yourself, using a testing method that reflects how users would charge and use their batteries.

I have just been testing one of those electronic load testers, it cost £28.

View attachment 53439


The display is a bit dim, but there is a USB connection and a Windows application which allows you to see whats happening on screen, saves a CSV file of the data too;

View attachment 53440

You set the discharge current, say 2A for a typical use scenario, and a cutoff voltage and start it running.

The test method that represent how the comparison between 42V or 41V chargers is achieved should be clear;

Charge battery with 42V charger.
Rest battery a bit
Run discharge capacity test
Rest battery a bit
Charge battery with 41V charger.
Rest battery a bit
Run discharge capacity test
Compare capacities

Now, I dont have an actual 41V charger, so I put a silicon and schottky diode in series with the output of my 42V charger, this allowed the charger to complete the constant volt stage of the charge and at the end of this the battery voltage was 41.04V.

The process used shows how easy it is to test the capacity of a battery that has actually been charged with a 41V charger as apposed to some attempt at a simulation using a 42V charger.

Trouble is, you still need to know what output you are getting from your charger before you add the dropping diode - are you getting 41.6 V from the charger or 42.4? If you add some dropping diodes to bleed off a volt, are you then getting 40.6 or 41.4 V? If you then use another charger, is the output from that similar? What is the spread you get from one charger to the next?

Hence you need a calibrated DVM.

Also, make sure you chose a diode that can handle the charging current and if it is dissipating a lot of heat, make sure it has a heat sink. The temperature of the diode will affect the voltage you see at the output.

 

[StuartsProjects]

Trouble is, you still need to know what output you are getting from your charger before you add the dropping diode - are you getting 41.6 V from the charger or 42.4? If you add some dropping diodes to bleed off a volt, are you then getting 40.6 or 41.4 V? If you then use another charger, is the output from that similar? What is the spread you get from one charger to the next?

If you measure the charger output voltages, as in direct across the battery, and your multimeter shows one charger as 42V and the other as 41V does it really matter if the multimeter has an accuracy of 'only' 1%, the difference in the charger voltages is going to be very close to 1V ?

Remember that what is being tested is the relative change in capacity with a 1V change in charge voltage and you dont actually need to know the absolute capacities, since the same discharger is used for both tests so the relative % change will be accurate.

Keep it simple.

 

[WheezyRider]

If you measure the charger output voltages, as in direct across the battery, and your multimeter shows one charger as 42V and the other as 41V does it really matter if the multimeter has an accuracy of 'only' 1%, the difference in the charger voltages is going to be very close to 1V ?

Remember that what is being tested is the relative change in capacity with a 1V change in charge voltage and you dont actually need to know the absolute capacities, since the same discharger is used for both tests so the relative % change will be accurate.

Keep it simple.

Yes, it does matter. There is a big difference if you are measuring discharge capacity from 42 V to 41 V compared to 41 V to 40 V.

All likely to be good for relative measurement with a non calibrated DVM from one charger to the next to give an idea of output variability, but it does not give the absolute value unless properly calibrated.