The whole idea is fraught with risks and full of ifs and buts and probability, plus it's a load of hassle to organise and monitor.

 

As I said before, unless you're a very high mileage rider, just charging normally, your battery will last for years, by which time there will be better ones. Batteries have been improving by about 10% per year, so in 3 or 4 years time, you'll be at a 30% to 40% disadvantage compared with a new one. When I built my bike 6 years ago, I had 11.6Ah. Now I have 17.5Ah for more or less the same weight.

 

Yes, I agree, at the moment, for most people, it's not worth the hassle.

 

However, if you are technically minded, have a pack which costs more than the motor and controller combined, if you can, doesn't it make sense to try and extend it's life where possible?

 

We live in a throwaway society and it can't carry on forever.

 

The world is full already of electronic waste. The extraction of the lithium is energy and water intensive and causes environmental and social issues in its extraction. There are also lots of unpleasant/strategic chemicals in lithium cells that have to be manufactured, creating further issues with waste products and then ultimate disposal. Most lithium cells are still not recycled.

 

Furthermore, from these kind of experiments, it becomes likely that more intelligent chargers will be manufactured to prolong the life of battery packs by charging to <42V most of the time and then occasionally doing a balance charge. For instance, if you had a Bluetooth BMS, it could transmit to the charger whether a balance charge was needed or not. Then it would all be automatic and hassle free.

The simple answer would be for all future generic BMS to be programmed so they balance at 4.1v per cell group to extend the life of the cells/battery. The throw away world can't go on for ever with resources being used up.

 

My 29E's batteries both 6 years old and only 3 or 4 digits difference between serial numbers behave the same at 41v max charge. No deterioration in mileage use or an shenanigan's in use, I simply adjust/set the San's voltage pot to the voltage required and they charge up nicely. All I do is check the battery charge voltage when the green light appears and monitor mileage usage, the longer I eke them out for my local usage then the more user friendly they are for the planet. In the bigger pics of things not going to make much difference but I'm doing my little bit for the world.

Latest update:

 

Battery pack is now up to 40 cycles of 4.1V charging and discharging to approx 3.68V per cell.

 

All banks are still within 0.01V when charged and at most 0.005V when discharged.

Up to 50 cycles now. Still nothing to report, other than all banks within 0.1V charged, 0.005V discharged to 3.68V.

Should hit 60 cycles tomorrow, at 59 at the moment.

 

After 56 cycles I measured the pack on a Fluke bench top multimeter, with 4 decimal places, as ordinary cheap multimeters are not that accurate and I wanted to be sure I was getting reliable data.

 

Discharged to 37.079V:

Bank no. ) Voltage

 

1) 3.7097

2) 3.7090

3) 3.7092

4) 3.7076

5) 3.7075

6) 3.7110

7) 3.7114

8) 3.7100

9) 3.7107

10) 3.7080

 

Biggest variation 3.9mV

 

Charged to 41.034V:

Bank no. ) Voltage

 

1) 4.1025

2) 4.1004

3) 4.0990

4) 4.0999

5) 4.1003

6) 4.1114

7) 4.1042

8) 4.1056

9) 4.1029

10) 4.1088

 

Biggest variation 12.4 mV

As promised, went through 60 cycles today.

 

Put the pack on the Fluke bench top multimeter again and had another look:

 

Discharged to 37.363V:

Bank no. ) Voltage

 

1) 3.7356

2) 3.7356

3) 3.7351

4) 3.7345

5) 3.7341

6) 3.7398

7) 3.7380

8) 3.7389

9) 3.7383

10) 3.7375

 

Biggest variation 5.7mV

 

Charged to 40.913V:

Bank no. ) Voltage

 

1) 4.0901

2) 4.0883

3) 4.0870

4) 4.0880

5) 4.0882

6) 4.0992

7) 4.0919

8) 4.0936

9) 4.0909

10) 4.0966

 

Biggest variation 12.2 mV

 

So still no significant difference.

 

For comparison, I had a look back through my logbook and I found I did actually put the pack on the Fluke back in August, after only 9 cycles of 41V charging:

 

Charged to 40.984V:

Bank no. ) Voltage

 

1) 4.0981

2) 4.0949

3) 4.0940

4) 4.0937

5) 4.0955

6) 4.1067

7) 4.0989

8) 4.0996

9) 4.0997

10) 4.1031

 

Biggest variation 13 mV

 

So no significant difference from 9 to 60 cycles.

Good work WR, I don't bother recording with my 29e's as I have no extra sense wires on the exterior to measure off. Though am not seeing any major reduction in voltage sag or range so cell groups are appearing to to be maintaining good relative balance.

Passed 71 cycles today.

 

Put the pack on the Fluke bench top multimeter again and had another look:

 

Discharged to 36.963V:

Bank no. ) Voltage

 

1) 3.6963

2) 3.6956

3) 3.6957

4) 3.6945

5) 3.6946

6) 3.6983

7) 3.6981

8) 3.6980

9) 3.6978

10) 3.6959

 

Biggest variation 3.8 mV

 

Charged to 41.041V:

Bank no. ) Voltage

 

1) 4.1029

2) 4.1008

3) 4.0995

4) 4.1006

5) 4.1009

6) 4.1121

7) 4.1048

8) 4.1063

9) 4.1034

10) 4.1094

 

Biggest variation 12.6 mV

 

So still no significant difference.

After a break following Christmas and then an accident on the ice in January, I've started riding again this week and collecting data.

 

82 cycles now. After charging:

 

Charged to 41.013V:

Bank no. ) Voltage

 

1) 4.1013

2) 4.0998

3) 4.0982

4) 4.0967

5) 4.0982

6) 4.1092

7) 4.1018

8) 4.1038

9) 4.1007

10) 4.1067

 

Biggest variation 12.5 mV

 

Still no significant change...

Today the pack finally passed the 100 Cycle mark!

 

Charged to 41.001V:

Bank no. ) Voltage

 

1) 4.0986

2) 4.0972

3) 4.0955

4) 4.0972

5) 4.0966

6) 4.1079

7) 4.1006

8) 4.1027

9) 4.0995

10) 4.1055

 

Biggest variation 12.4 mV

 

So 100 cycles, charging to 41V and no significant change in balance.

 

Bank 3 seems to be consistently the lowest voltage after charging and bank 6 the highest. However, bank 3 seems to maintain voltage better than some other banks when discharged, although these changes are tiny really.

 

Either this is an example of a well behaved pack, or the BMS is still balancing, even though it is only being charged to 41V and the balance feature is not supposed to activate until 4.18V is reached.

Have you checked with a meter when the bleed resistors start bleeding? Your whole test would be a waste of time if they're opening.

Have you checked with a meter when the bleed resistors start bleeding? Your whole test would be a waste of time if they're opening.

 

Unfortunately the BMS is a sealed unit, so I can't get to the bleed resistors.

Unfortunately the BMS is a sealed unit, so I can't get to the bleed resistors.

 

However, even if the BMS is balancing at 41V, it's still a win, as without this experiment, I wouldn't know that it was.

 

I have another identical BMS and at some point I'll rig it up with some cells that are out of balance, charge at 41V and see what happens.

Up to 150 cycles now.

 

Pack voltage 41.044V:

Bank no. ) Voltage

 

1) 4.1029

2) 4.1016

3) 4.0999

4) 4.1020

5) 4.1009

6) 4.1121

7) 4.1047

8) 4.1071

9) 4.1037

10) 4.1097

 

Maximum voltage difference 12.2 mV, so still no sign of the pack going out of balance.

So after a year I finally got the pack back after lending it to someone with a charger set to 41V. It looks a little worse for wear now as it was attached with cable ties and gaffer tape to their bike! But it is still functioning fine.

 

 

 

It's been 3 years now of charging to 41V and hundreds of cycles.

 

In terms of the state of balance:

 

Pack voltage 40.882 V

 

Bank no. ) Voltage

 

1) 4.0856

2) 4.0872

3) 4.0830

4) 4.0883

5) 4.0835

6) 4.0957

7) 4.0872

8) 4.0920

9) 4.0861

10) 4.0939

 

So even after 3 years, the biggest variation between cells is still only 12.7 mV.

Are you going to do a capacity check ?

Are you going to do a capacity check ?

 

Yep, will do. Not had a chance to do much with the pack so far since getting it back. Hopefully next week.

Can you make a discharge curve like saneagle's in the other thread?

Can you make a discharge curve like saneagle's in the other thread?

 

Can you link the thread?

This thread, post #63 maybe;

 

https://www.pedelecs.co.uk/forum/threads/batteries-charging-routine.45703/page-4

This thread, post #63 maybe;

 

https://www.pedelecs.co.uk/forum/threads/batteries-charging-routine.45703/page-4

 

Ok, but I'll probably use a higher current than 0.5A, otherwise it'll take too long.

This graph from saneagle is very easy to understand

https://www.pedelecs.co.uk/forum/attachments/battery-discharge-jpg.52902/

 

Yes, but it is a check from 4.2V to 4.1V per cell, so for accuracy, done at 0.5 A.

 

I would do a capacity check from 41V to 35V at say a couple of amps.

I think there is strong interest in charging at 41v in order to prolong life of the battery and reduce fire risk during charging. However, the cost needs to be quantified, how much loss in capacity? I can get 41v chargers made.

I think there is strong interest in charging at 41v in order to prolong life of the battery and reduce fire risk during charging. However, the cost needs to be quantified, how much loss in capacity?

 

In my experiments around 7 or 8% and most of that is between 41 V and 41.5 V. Above 41.5 V the added energy becomes very small and then when the balancing trips in at around 41.8 V, it becomes hard to tell how much energy is going into charging and what is being consumed by bleed off in the balancing process. It might be better doing such capacity experiments on a single cell, where there is no balancing.

Yes, but it is a check from 4.2V to 4.1V per cell, so for accuracy, done at 0.5 A.

 

That graph gives you the capacity loss when drained from 4.2V to 4.1V per cell. Which does not tell you what the capacity would be when charged to 4.1V and drained from there.

 

I used a practical drain load of 2A on my 42V 5Ahr battery as that was the current used when on a typical ride, assist level 2 at 14mph or so.