Help diagnosing sudden power reduction problem

[Obiwan007]

Here is the bike. It is an older Europa tres terra 3.

I have had it for about a year and it has been pretty bullet proof since I got it. Last week, however, between one battery charge and another it developed some odd symptoms. With a full charge the bike powers up but is VERY weak and the red "empty" light on the throttle mounted gauge is lighted. Almost immediately the empty light begins to flash as if the battery is near LVC and then the bike powers down. I have re-charged the battery and tested it to find it fully charged. I have gone through all of the connections and found them secure with good continuity. I am unfamiliar with the Heisman motor which I believe is a brushed motor. Here are some other pictures of the motor, battery compartment and controller.





Any help with devising a plan for diagnosing the problem would be greatly appreciated. I use this bike primarily for e-bike evangelism and having it operational allows me to take others out on rides with us. Obiwan

 

[DAND214]

You tested the battery after a charge. Have you tested it when the low battery light comes on?
Just because the battery takes a full charge doesn't mean it has any capacity left in it.
A bad cell can ruin the whole pack.

A brushed motor, even if it has bad brushes shouldn't kill the pack in minutes or less. Bad brushes will cause it to run weak but not kill the battery. If the motor causes the battery drain that fast would melt wires too.

Dan

 

[skeetab5780]

If that is the battery type i think it is. It is probably faulty. although there are only 10 cells so it could be easy to fix.

I owned a Calisto and they use 10ah Phylion Lifepo4 prismatic stainless steel cells. 4/10 of mine were noticeably puffed upon my first time opening and inspecting the new battery pack. On the Calisto model each cell has some sort of a single cell level circuit boards, Along with one larger BMS board. And both bikes have the same exact charger and battery specs so that may be what you have. My pack ended up being used as a useless bench test pack since it was fairly saggy.

On a positive note 36v(10s) is really easy to come by these days. You can buy tool packs or really anything available..
http://www.gopedelec.eu/cms/dmdocuments/PhylionGoPedelecTrainingSlidesEN.pdf
http://endless-sphere.com/forums/viewtopic.php?f=14&t=31920

 

[Obiwan007]

Thanks for the responses dan and skeet. So I will be doing a closer inspection of the charged and blinking red light voltages on the battery. But as I recall the battery was about the same voltage before as after the the cut out. I did open the battery to see if there was any obvious problem in there and the inside looks exactly like these pics I found:


This my exact charger too. The battery is a 36V Li-ion. Charger is 42V 2A. I tend to think this is not a battery problem because the problem started suddenly. Thoughts? Obiwan.

 

[Drunkskunk]

I tend to think this is not a battery problem because the problem started suddenly. Thoughts? Obiwan.

Problems starting suddenly are often a battery problem. The problem you describe is one of failed cells. the only way to know for sure is to test the voltages.

There are several ways a cell can fail. It can simply refuse to take a charge, meaning the whole battery won't reach full voltage. Or it can lose capacity, meaning the whole battery will reach full voltage, but the one cell will drop off very rapidly after being connected or used, resulting in a low voltage battery. Or a cell can suffer increased internal resistance, meaning it will keep capacity, and the whole battery will reach full voltage, but under load the voltage will drop dramatically.

And to complicate it, you can have different cells in different failure modes at the same time. each failure requires a different kind of test. But the end result is the same, regardless. The battery is probably toast.

 

[Obiwan007]

Thanks for the reply skunk. So can you describe what tests I should do on the battery? Love your bike by the way and always appreciate your informative posts. Obiwan

 

[Tench]

A good place to start would be measuring all the individual cell voltages, you can do this from the balance lead plug centre bottom in your pic. Starting from the end with the black which I presume will be the negative end of the cell string measure between the first and second terminal, this will give you the voltage of the first cell, then as the cells are in series measuring from the 2nd terminal to the 3rd will give the voltage of the 2nd cell, then measure from the 3rd to the 4th etc until you have all 10 voltages. Post these readings in this thread so we can take a look and see where to go next.

 

[Obiwan007]

Ok, I think I understand the process. I am on duty at the firehouse today and tomorrow so I will get this done on tues. And report back. Thanks for the help Tench, while I have built a couple of bikes now, I still have a way to go on diagnostics and such. It seems like the learning curve keeps getting steeper for someone with no electronics background. Obiwan

 

[Hugechainring]

I have a few EVG bikes that use the same components, except the battery pack you are using is an "upgrade" over the SLAs. I'm with the others. Everything you describe says bad cells/battery.
I imagine you would like to get on the road for as little money as possible. One way or another you probably will find the cells need to be replaced. If only a couple are bad now, the rest will soon follow. It is also indeterminate whether the BMS is faulty or not. If you can afford it, it might be better to replace the whole assembly with a rectangular pack by EM3EV or Ping battery. I've ordered 3 Ping packs that were excellent replacements for that frame, fit perfectly where the old battery case went. But... they are expensive. Cost me more than the bike. But the performance is excellent for the application. I could ride with human assist over 35miles on a 36V 15Ah pack and not hit LVC.
If you need more info, PM me. The specs for the Ping packs are custom for that application, and I'd be happy to provide them.

 

[Obiwan007]

I may indeed be interested in that spec huge. Let me do the cell tests firs and see where I'm at then I'll get back in touch with you. Obiwan

 

[Obiwan007]

Ok, got to do some testing. Here are some pics to start:

The battery opened up:

Closer in showing mysterious sticker with the number 44 written on it:

Closer in on the ...? cell junction plug?:

Opposite side of the plug where I probed for testing with my DMM:

And a shot of one end of a revealed cell on the left end of the board:


So, referencing this picture:
View attachment 2
I began testing the fully charged battery by probing the red wire on one end and the black on the other. Then working from right to left (since some wires are the same color) I tested each wire with the one next to it so, red and yellow, yellow and orange, and so on to get the results listed below. Note that the junction plug has two unused slots in the middle but I got consistent readings by testing the two wires adjacent to the gap. Here are the results:
EUROPA W/heinzmann motor
36V Lithium ion

1330 hrs.

Red to black = 41.3
Red to yellow = 4.15
Yellow to orange = 4.16
Orange to brown = 4.15
Brown to white = 4.13
White to blue = 4.13
Blue to green = 4.13
Green to yellow = 4.12
Yellow to orange = 4.12
Orange to brown= 4.12
Brown to black = 4.12

My research so far indicates that this pack was originally labeled as being 36V 9.5Ah. I believe I have read that Lithium ion cells have a top voltage of 4.2V. And there has been some suggestion that these were originally Sanyo cells. The charger (original) outputs 42V 2A. Based on what I can see of the cell arrangement they are arranged five wide and 10 long for a total of 50. Based on this I surmise that they are arranged 10s5p. That would mean that each cell is 1.9ah. Does this all track with you guys? I did a second test two hours later to see if any cells were self discharging and got identical readings. So where should I go from here? Obiwan

 

[Hugechainring]

OK,
So the voltages look OK based on what you are saying. Are all the cell sense wires intact? The connector block to the PCB looks a little sketchy, but perhaps I'm not looking at it right. How about the 4th wire from the right? Good condition? Something looks melty there or maybe it's just some tape residue or something.

 

[Obiwan007]

Ya, the connections look good. They used a bunch of black glue all over the place in this battery box and what you are seeing there is just some residue. I think the batter appears to be good too. What should I do next to track down this problem. This board does not appear to have problems as far as burned components or cracked solder joints. Controller next?

 

[Hugechainring]

Could be the controller. However I went back to your original symptom description, which mentioned the battery level indicator dropping off nearly immediately even though it seems like your cells are charged. I'm still going for either the battery or BMS. I had a few SLAs that acted like that. I'd plug them into the charger, it would indicate full charge. Measured the voltage, showed 37V, put a load on it, it fell through the floor. Point is, it is not possible to evaluate a battery based on voltage at rest. You need to put some stress on it. It's like the battery testers the automotive shops use. All it is is a load bank with a voltmeter. They put a 500A load on the battery and watch the voltage drop. Good batteries should hold say 10V for 30 seconds.

I would suggest attaching a couple of small insulated leads to the main pack terminals, where you should be reading about 41V. This is awkward, but then reassemble the battery case and put it in the bike, with access to your leads. Insulate the leads with tape or something until you are ready to use them. Hook up your DVM, turn on the key switch while reading the pack voltage.
If the voltage maintains at 40-41V while the charge indicator goes to red, there is something wrong further downstream. If the voltage takes a cliff dive, there is something wrong with the pack.

 

[DAND214]

Have you checked the cells after the Empty light comes on?
Not the whole battery voltage the cells themself. All it takes is one cell to crap out and the whole thing dies.

You can wire a couole househild 100watt bulbs to the output and test the cells as it discharges.

the way you discribe how it started and how fast it dies, still leads to the battery.

Dan

 

[Obiwan007]

OK, I'll do a load test of some sort in the next couple of days. I like the idea of ruling out the battery before moving on anyway. Stay tuned, I'll report back... Obiwan.

 

[Obiwan007]

Here is the report on my load test. I set it up like this:

The power cord was long enough to reach out to where the battery sits:

Had to put the kickstand up on a block and weight down the front wheel to keep the rear off the ground:


So I started by testing the full pack voltage and found it unchanged from two days previous. Same with the individual cell voltages so those listed in the previous post represent the starting point. I then ran the motor and rechecked each cell voltage while the load was on. Then, without changing the throttle speed, I applied the rear brake enough to put some additional load on the motor and rechecked each cell voltage. The result was that with the wheel turning but no brake each cell dropped an average of .05V. With the brake applied the average drop was .08V. There did not appear to be any individual "bad actors" among the cells and the readings were consistent within .03V from cell to cell. At the end of the test the total pack voltage had dropped by .3V which also seemed consistent to me since the test took longer than you might think since it took two guys coordinating together and me trying to probe tiny components. finally, the strangest thing of all; after the test I realized that the original symptom I had been describing should have caused the Red light to blink and then the motor to cut off. But it did not. Looking at the charge condition lights I find the green light shining. WTF! The bike is now operating perfectly. Same as it always has. I have no idea what I did to fix it. I realized that after charging it fully before beginning all of these tests I never connected it back to the bike to check its operation. That should not make any difference though since I charged it and checked several times prior to opening the battery. I hate to look a gift horse in the mouth but I would like to understand and really don't want it to crap out again halfway through a ride and have to pedal it home. Any ideas? Obiwan

 

[Hugechainring]

I'm liking the gift horse. The one variable we haven't really touched on is the pressure connector with the two spring loaded pins that connects the battery when you install it in the frame. If these pins are all burnt and pock marked you can get an intermittent bad connection with the symptoms you describe. I believe they are silver plated , so you can't just sand them. They will degrade very quickly.

You might also check the controller power leads that screw onto the frame mounted connector. If you need to replace the connector (both parts) let me know. I have a few of those. Otherwise, the best you do is go through everything one more time, making sure all the connections are secure. If the problem comes back, maybe some new evidence will come with it.

 

[Obiwan007]

Thanks Huge. This battery connects to the bike via a power plug. One of those 6 sided computer 110V looking things. Not the spring loaded type like you mention (I have some of those on my e-moto). The contact on this one is solid and secure so I don't think that is the problem. I did find a connector hidden behind some sheet metal in the back of the battery compartment that I plan to check out carefully but a cursory inspection of it looked like it was just fine. I'll run this thing for a while and see how it does. Thanks to all for the input from everyone. Problems are a pain but I learn something each time for the data banks. Obiwan