Kalkhoff Impulse battery charging issue

[TimReynolds]

I have recently bought a 2012/2013 Kalkhoff Agattu C11 Impulse Premium. I know there are potential mechanical issues with these Impulse motors but that side of the bike seems to be fine and it rides really well with plenty of power assistance, not surprising really, as it shows only a little over 500 miles total use since new. It has been completely unused since 2019. In all respects, the bike is just like it came out of the showroom! The battery was showing 100% full and 100% capacity when I brought it home, although I have no idea when it was last charged.
I have done a few test rides, probably 10 miles with some steep hills, battery content had dropped by 1 LED so I put the battery on the charging cradle - there was a brief flicker of the battery LEDs when I connected it but the green light stayed on the charger. I left it for a couple of hours but nothing happened, the content level stayed at 4 LEDs, i.e. the battery is not being charged. Repeated this with the charger directly connected to the battery - same result.

I have checked the battery externally with a meter, the pins are un-numbered but referring to the charger cradle as pin 1 on the right, 1-2 is 36.4V, 1-3 is 2.8V, 1-4 is 2.8V and 1-5 is 0V.

I also checked the cradle with a meter; 1-2 is 0V, 1-5 is 4.8V ( there are only three pins on the cradle).

I think the charger is supposed to handshake with the BMS in the battery and then initiate charging - is that the flicker of the battery LEDs?
Looks like the handshake is failing. If so, why? Duff cell? If there is a failing cell or group of cells, wouldn't the battery performance be affected? The battery seems to be doing its job OK.
I've seen mention on this forum of a thermal fuse on the BMS board, is that a possibility? Or a duff BMS board?
What should my next troubleshooting move be? Dismantle the battery to check the cells or cell groups? Does anyone have map of the access points to test these?

Thanks for taking the trouble to read this, any help would be welcome!

 

[soundwave]

try Scott in my sig Torque-E-Cycles battery repair for can bus batt stuff as no one else can do it uk.

 

[TimReynolds]

Interesting....
Does this bike use CAN bus to communicate between processors??

 

[soundwave]

it will be can bus or uart coms but both require hand shakes from the controller and display.

the bosch bms will brick it self if power is removed but peter at ebike motor centre can service the motor if you need it as for the batts if the bms goes tits up its a bin job.

 

[TimReynolds]

Ah, it's not a Bosch motor or battery system, it's the infamous Impulse system that predated Bosch.....

 

[soundwave]

they all use a hand shake system and can bus is the worst of them all.

last resort is to use another controller for the motor and a normal dumb bms

 

[dlwest]

I have recently bought a 2012/2013 Kalkhoff Agattu C11 Impulse Premium.

Have you tried charging through the side port, or through the bottom pins using dock stand provided with some Kalkhoff bikes? Is your charger original, or an aftermarket one?
Side port pin 1 is a signal wire for the BMS to indicate the charger is connected. Pin 1 goes to 0V inside the charger's plug.
Side port pin 2 is a signal wire for the charger to indicate the load is there, it is pulled down with 10k on the battery side, right behind the socket.
Side port pin 3 is 0V
Side port pin 4 is +42V, CC/CV

These old batteries don't use any RS or CAN communication and are controlled by pull-down signals enabling charge and discharge. I have two of them, the 11Ah that works as intended, the other is 15Ah quirky one: despite having all cells perfectly balanced down to 0,01V, it shows full charge (all 5x LEDs lit) at 37V and empty at 35V. As a result, it gets fully charged in just 15 minutes and indicates zero charge left after 7 miles (not cutting off though). But if I ignore the LEDs letting the battery sit on the dock until reaching 42V, it gives full 50 miles range until the motor cuts off at ~32V, even after 0% charge was indicated.

 

[TimReynolds]

Hi dlwest, thanks for your reply!

The charger is the original one, supplied with the bike. I've tried charging through the side port and using the cradle, same result - a brief flicker of the battery LEDs then all extinguished, charger LED remains green, nothing else happens.
I'll have the battery apart and see if the pull downs, especially the 10k is O/C, that might be the problem. At the moment, the battery is almost dead flat, I should be able to check the cell voltages to see if there are any noticeably duff ones.

Thanks again!

 

[dlwest]

Please check charger output voltage first: it must output 42V briefly when you pull down the "R" pin on the cradle with 10k. If no other load is connected, 42V will disappear and the charger will lit green.

 

[TimReynolds]

OK, have done the test. With the R pin strapped to 0V, 42V appears for about a second, then disappears. Charger LED goes briefly red, then green when 42V has gone. Same test done at charger XLR connector, same result.
I then tried to mock up a load with three 12V, 21W bulbs, connected in series. My calculations say that these should draw about 2A from the supply. The result was that the bulbs flash, the 42V appearing and disappearing in rhythm and the charger LED flashing red and green.
I've opened the battery and checked the direct charge port 10k resistance is there - it is.
It looks to me like the charger is probably working, just confused about a static, resistive load.
I think the battery or BMS is SNAFU. Any idea how the cells are organised or where to measure voltages?

 

[Ghost1951]

No expert - but have you checked the voltages of the individual banks of cells?

It might just be that one or more of the parallel groups of cells are at a low voltage and the BMS is refusing to charge them. If they are not too far down, you could raise their voltage by charging individual banks of cells to about 3.3 volts so that the BMS will allow them to charge further by the usual method. Quite a lot of batteries which have been left alone for a good while meet with this problem.

I am sure you can find an inexpensive lithium charger which will charge at up to 4.2 volts and this can be used to raise the unbalanced cell groups up to the sort of level the rest of the cell groups in the battery are at. At a push and WITH CAREFUL SUPERVISION so it doesn't go over voltage, you could use a USB charger to lift the level, BUT DON'T go off and leave it alone because it will over charge the cells eventually because it will ultimately go to 5 volts. Better and safer by far with a proper small lithium charger.

If you find on measuring that some of the cell groups are very low - say below about 2.3 volts, there is nothing you can do to recover them safely.

 

[dlwest]

say below about 2.3 volts

Actually, I have two Kalkhoff batteries successfully recovered exactly from that state. Both were forgotten in storage for some time until the BMS was no longer able to take charge. Recovered by trickle charging using a bench DC power supply. Don't know if any capacity was lost or not, but they still do few days of commuting between the charges.

 

[TimReynolds]

My thoughts entirely. I mapped out the layout of the batteries and how the 50 cells were strapped (a string of 10 sets of 5 cells in parallel). I read the voltages of each set of 5, they were within 0.1V, between 2.8 and 2.9V. Just to be sure, I used my bench power supply to bring them all up to exactly the same voltage. Put them on the charger - nothing.
The battery was fully charged when I got the bike, but has never recharged since I've had it. The vendor must have had the charger working though.....
Sadly, my bench power supply only goes up to 32V, so I can't directly trickle charge the batteries - unless I split them in two.
So what's the betting?
Duff charger or duff BMS board? Both seem to be working individually, just not together. Charger is £69, battery is 333 Euros from Germany.

 

[Ghost1951]

if your bench supply managed to get them to 32 volts, they might still be just under the point at which the BMS will charge them. Can you re-check the individual parallel group voltages? If any one was just below the cut off point you might be stuck with the BMS refusing to 'play ball'.

If that is the case, I am thinking it might still be recoverable if you could get the cells a little bit higher.

EDIT: this is just an opinion and other views may differ.

 

[dlwest]

I'd connect the bench supply to the whole string, with current limited to 100 mA. It will reach 32V in a few hours. After that, the BMS should be able to take charge in the normal way.

 

[dlwest]

Actually, even if the cells look to be out of balance with 0.1V difference, they still may have the same SoC because of steep V curve at low charge levels. I would be charging the whole string to see if the individual cell voltages will level up at above 3.3V.

 

[TimReynolds]

I would agree but when I first took the bike on, earlier this year, the battery was fully charged, right to the top. I didn't check the internal battery voltages at that point but they must have been ~3.5 to ~3.6V. However, the battery didn't charge then, when put on the charger. I ran it down to 4 LED's, thinking it needed to be a bit discharged to take any current but nada.
Now, 9 months later, the battery's dead flat and still won't charge, hence I think the problem is with the electronics somewhere rather than the cell chain......
I can try charging the whole chain of cells but I fear it won't make much difference.