Eric's Project #002

[fechter]

Certainly if the batteries are more discharged, the charger will get warmer charging them. I had a similar problem with ganged chargers. If the chargers are all stuck together, the ones in the middle have much less surface area to cool with. This can cause overheating. I had to separate my chargers to leave an air space between them. Even then, the one in the middle always got hotter.

If you put sort of a tube around the chargers and used a fan to blow air through, it might prevent them from overheating. Forced air cooling for chargers.

 

[Beagle123]

I Did Something Really Stupid

I was pressing to get this thing done today, and I had it all assembled at about 9pm. All I had left to do was connect the wires. I was really tired, and I plugged the negative end of a 12v, 18ah pack into its positive terminal. I didn't realize what I had done for a couple of seconds. I can't believe I did that. Well, if I didn't blow up from that, I guess I'm safe.

After this screw-up, I tested the sub pack with a voltage meter, and it was still 11.75 volts. The others were about 11.9 volts.

Did I just wreck this pack?

I finished the bike and took it for a ride. It went pretty well. I'm going to have to try it again with a speedometer. It seemed to perform fine, but I don't think it was running at 2hp. If I had to estimate, I'd say that it was running at about 20 amps, which is about 1C (18 ah pack).

Keep in mind that I'm using a 100 amp controller. If I understood you guys correctly, The batteries should be capible of delivering 2C (36 amps) easily, and the controller shouldn't flinch at that.

I'm going to have to make some kind of serious load tester to get to the bottom of this. I'm thinking that I should get a couple of 1 ohm power resistors. Together they should draw about 20 amps. That should make a good voltage drop. I can compare the pack I zapped to the others.

Does that sound about right?

I'm worried that I zapped a few of the cells, so that pack is relying on fewer cells which could cause the whole pack to underperform. Also, could this make the remaining cell overdischarge.

I hate this.

 

[vanilla ice]

Damn. If it makes you feel better I did something similar today with my escoot. I jumped two batteries together backwards. Tired is dangerous. Can't help with the questions though. My slas seem alright.

 

[Beagle123]

Screw-Update

It was eating at me, so I went to the garage to load test the battery packs with the four 10 ohm resistors I have already. Together, the four resistors should make 2.5 ohms of resistance. According to ohm's law that's 4.8 amps of current. I tested the current of each pack before and after being connected to the resistors. Here are the results:
<table border="1" cellpadding="5">
<tr><td> pack </td><td> V (before) </td><td> V (after) </td></tr>
<tr><td>1</td><td>11.65 </td><td>11.30</td></tr>
<tr><td>2</td><td>11.80 </td><td>9.7</td></tr>
<tr><td>3</td><td>11.65 </td><td>11.40</td></tr>
<tr><td>4</td><td>11.79 </td><td>10.55</td></tr>
<tr><td>5</td><td>11.68 </td><td>11.02</td></tr>


</table>

Update: I did a comparison with a SLA on my other scooter. It started at 12.8v and dropped to 12.5 under the same 2.5 ohm (5-4 amp load).

So SLAs seem to perform similarly to Lithium?!?! Or perhaps I need a bigger load.


Its strange that the battery packs with the highest voltage had the biggest drops in voltage under load.

Pack #2 dropped 2.1 volts with only about 4 amps of current. UHHHH!!

What do you guys make of this?

Another strange this is that I could have sworn that I shorted pack #3, and it performed the best of them all. It only dropped .25 voltss. Weird.

I'm guessing I"m going to have to break apart those packs and test each cell?!?!?

I guess I"m going to have to try charging them and testing again.

 

[The7]

Discharge and charge them a few time to condition the cells.

 

[xyster]

So SLAs seem to perform similarly to Lithium?!?! Or perhaps I need a bigger load.

As far as discharge rate alone, SLA's are comparable. But SLA's suffer a much greater Peukert effect. SLA's can only provide a small fraction of their rated capacity at high drain rates. The emoli's should provide 2.6ah each, out of their 3.0ah rating, at their maximum safe continuous drainrate of 12C.

The power issue could also be related to the controller. I understand that model is rated for 100 amps max, but considering it's physical size and source (cheap chinese) I'm skeptical.

 

[Beagle123]

What do you guys think of the above voltage drops? If the voltage drops 2+ volts with only a 4 amp current draw imagine if there was a 50 amp draw. It would be a total failure. I think there's something very wrong.

 

[dirty_d]

test each cell in the bad packs

 

[xyster]

What do you guys think of the above voltage drops? If the voltage drops 2+ volts with only a 4 amp current draw imagine if there was a 50 amp draw. It would be a total failure. I think there's something very wrong.

The voltage drop, as it's a function of resistance, should be linear with amperage. Something appears very wrong. Even a 1p string shouldn't sag nearly that much at a ~1C drainrate. That tri-subpack you said you wired backwards appears seriously damaged. From your chart, that tripack sags much more than the others, and could compromise the performance of the entire pack since it's basically a big resistor in the circuit, and as such a big heater that could get hot enough to damage nearby cells. You might try pulling out tripack 2 and operating the pack at 12S. It might perform much better. Tripack 4 also sags too much.

 

[D-Man]

I'd cycle the packs a few times and see if it gets better. (if you haven't already). Then take it apart if no improvement.

 

[xyster]

I'd cycle the packs a few times and see if it gets better. (if you haven't already). Then take it apart if no improvement.

No point in cycling lithium batteries. The resistance of that one 3s subpack is so high that the cells may overheat, causing more damage. Though ostensibly safer than Lithium cobalt, these are still high energy density lithium batteries. That tripack is seriously damaged, or suffering from poor connections. Either way, I'd yank it, fix it if I could, else replace it before any further use.

 

[Beagle123]

I tend to agree with xter. The results are so bad that the problem must be solved. I'm amazed that the bike drove as well as it did yesterday.

In the last two days I've burned-out. I went out to the garage to hook-up the bike to test it, and a couple of the andersen connectors come off (don't get excited that wasn't the original problem). I quit for the day.

However, I did motivate enough to go to RadioShack and get some 1 ohm and 8 ohm resistors to use for load tests.

Using a 1 ohm resistor, I did a load test on the one left-over lithium from this project. Using Ohms law, a 1 ohm resistor should use 4 amps of current at 4 volts, producing 16 watts.

Here are the results:

Starting Voltage = 3.95v
Voltage with 1 ohm resistor = 3.27v

This is a good test because the battery is discharging at a bit over 1C (similar to riding)

IF all the cells acted like this one, I would start at 59v, and when riding using 18 amps (882 watts) the voltage would fall to 49v.

I'm not incredibly happy with that, but it should be decent.

I don't understand how people are getting their battery packs to produce 40+ amps. I'm getting severe voltage drops on brand new cells discharging at 1C!!!!

Shouldn't these numbers be better? People who make these cells say they can discharge at 12C+. I'm not a beleiver. I"d like to see one of these cells produce like that. I'd say they can produce at about a 1.5C maximum (with big voltage drop).

This is particularly disappointing to me because I designed my bike to have the batteries built into the frame, and I barely had room to fit these packs into the battery box. I don't have places to stuff more batteries.

I'm hoping that I can dismantle my packs, and remove the bad cells to fix the problem just so I can get to the place where I'm getting 25 amps @ 48v. That's pretty similar to my bottom-of-the-line SLAs. Uhhhh.

 

[safe]

In the last two days I've burned-out. I went out to the garage to hook-up the bike to test it, and a couple of the andersen connectors come off (don't get excited that wasn't the original problem). I quit for the day.

On my Project #001 I destroyed six different versions of my transaxle before I got it right. If you never burned out on a project like this you would be abnormal... it's perfectly expected that you will sometimes want nothing more to do with the bike.

But the thing to do is to do something else for a while. Drink a beer, have some fun at something else. Then when you come back to it later it's like starting with a fresh project. This ISN'T a job... you don't have any time schedule or deadline to finish within. Also the weather is soon to get colder (in the midwest we are having highs in the upper 80's still which is way above normal) and when it gets colder you will be less thrilled about riding anyway.

At my age I also get sore fingers and generally wear out earlier than in my younger years. I'm only three (+) years from 50 years old, so you sometimes have to deal with overall energy issues. (I've been riding my #001 bike a lot and that is tiring in itself)

So relax... pick off the problems one by one and if you keep coming back to it again and again you will get there.

You just can't expect not to burn out on a long project...

(also, I find that you get a "first release" and then you keep developing it anyway, so there is no "end" until a long time passes)

Sometimes if you work on something too long you get into this "automatic movement" mode where you do stuff, but you stop really thinking hard about what you are doing. This is because keeping yourself at full attention is hard to maintain after a while. After some "vacation time" you come back to it with better mental clarity. People almost always hurt themselves when they are tired either physically or mentally... the mental exhaustion is in some ways more dangerous because people don't realize it's even happening.

Know thyself...

 

[fechter]

The extremely high current from the short circuit might have fried the conductive epoxy stuff. Possibly the cells themselves are OK.

 

[xyster]

Knowing thy batteries is more important in this case, Safe.

Beagle, check out the discharge chart below of a single emoli cell. From just looking at the chart, every 1C of additional current yields an additional voltage drop of about 0.05V. So your cells should be sagging about this same 0.05V under a 1C load compared to resting voltage.

As a comparison, one of my 18650 cells drops 0.12V under 1.6 amp (~3/4 C). A subpack of 15 droops 0.25V under a 33 amp, 1C load. And the entire pack of 20 subpacks sags 6V under full, 35 amp load -- some of the variance of course due to the wiring and connections.

Per Fechter's observation, you can also check the voltage drop across the epoxy. By measuring (under load) from the cap of one cell to the other cap of the same cell, and then again from the cap of that same cell to just across the epoxy junction of that same cell. Subtract the second reading from the first.

 

[safe]

It's good that you guys are addressing the technical issues he's dealing with, but I was trying to talk about the emotional experiences of a "long project". Sometimes we place our own artificial timelines and self imposed pressures on the project and if we don't live up to them then we end up hurting our own feelings.

It's like at times we can be our own worst critics...

(so that's why it's nice to come here and get other people to be even worse critics and that makes my own self criticism seem mild in comparison :lol: )

 

[Beagle123]

More Load Tests

Here are the voltages for my first load test:

<table border="1" cellpadding="5">
<tr><td> pack </td><td> V (before) </td><td> V (after) </td><td>Diff</td></tr>
<tr><td>1</td><td>11.65 </td><td>11.30</td><td>0.35</td></tr>
<tr><td>2</td><td>11.80 </td><td>9.7</td><td>2.10</td></tr>
<tr><td>3</td><td>11.65 </td><td>11.40</td><td>0.25</td></tr>
<tr><td>4</td><td>11.79 </td><td>10.55</td><td>1.24</td></tr>
<tr><td>5</td><td>11.68 </td><td>11.02</td><td>0.66</td></tr>
</table>

Here are the voltages after I charged the batteries overnight:

<table border="1" cellpadding="5">
<tr><td> pack </td><td> V (before) </td><td> V (after) </td><td>Diff</td></tr>
<tr><td>1</td><td>12.09 </td><td>11.70</td><td>0.39</td></tr>
<tr><td>2</td><td>12.24 </td><td>10.05</td><td>2.19</td></tr>
<tr><td>3</td><td>11.66 </td><td>11.42</td><td>0.24</td></tr>
<tr><td>4</td><td>11.87 </td><td>10.77</td><td>1.10</td></tr>
<tr><td>5</td><td>12.34 </td><td>11.85</td><td>0.49</td></tr>
</table>

I also verified with a ampmeter that about 4-5 amps are flowing during this test.

There are a few interesting things:

1) Pack 3 didn't change at all. Even after charging overnight it only charged to 11.66 volts. This is the pack I may have damaged. Also, the voltage under load was unchanged. The weird thing is that it seems to be producing current with no problem despite the slightly lower voltage.

2) Pack 2 is still a problem. Charging it again didn't help. It still drops more than 2 volts with only 4-5 amp current draw. Something is seriously wrong.

3) Pack 4 is still underperforming. It drops 1.17 volts with only 4-5 amps. I would assume that Packs 1 and 5 are functionsing correctly, dropping only about 0.5 volts. This is a good result. This pack is more than doubling the voltae drop of the others.

Now, I'm going to test each 4v sub-pack by connecting a load tester to the charging leads. Results to follow.

 

[xyster]

1) Pack 3 didn't change at all. Even after charging overnight it only charged to 11.66 volts. This is the pack I may have damaged. Also, the voltage under load was unchanged. The weird thing is that it seems to be producing current with no problem despite the slightly lower voltage.

If you haven't already, try using a different three chargers. I have some chargers that charge to 4.18v, and one that charges to 4.21v. However, 3.88v (11.66v/3) is way low -- that's only ~50% charged. To help ferret out the problem, you should get voltages on each of the 15 parallel subpacks individually, instead of in groups of 3. Edit: I see at the bottom you are doing this next.

2) Pack 2 is still a problem. Charging it again didn't help. It still drops more than 2 volts with only 4-5 amp current draw. Something is seriously wrong.

Definitely something wrong.

3) Pack 4 is still underperforming. It drops 1.17 volts with only 4-5 amps. I would assume that Packs 1 and 5 are functionsing correctly, dropping only about 0.5 volts. This is a good result. This pack is more than doubling the voltae drop of the others.

0.5 volts sounds good @ 5 amps. That extrapolates to about a 5 volt sag at 50 amps.

 

[Beagle123]

Mystery Solved!

I read the voltage on each sub-pack, and did a load test on each of them. All 15 had voltage drops of around 0.2 volts. That means that the 1.6 amps of current is divided by 6 cells. That's 0.26 amps per cell or 1/12th of 1C. The voltage shouldn't drop at all.

Either all 90 batteries were fried, or something seriously weird was going on.

I dismantled the worst pack down to the cells, and load tested them. They were perfect! It turns out that the "conductive" epoxy doesn't conduct.

There was epoxy residue on each side of each battery. When I touched the volt-meter to the epoxy, the voltage jumped around. I had to touch it to the bare metal to get a reading.

I went from being pissed that all my batteries were fried to pissed that they sold me that crap.

I'm going to have to start over. I tearing down each pack to extract the cells.

 

[xyster]

Well good, glad it was the epoxy instead of the batteries. That stuff sounded too good to be true in that if it worked well, it should be more common than solder. Another option is to send your cells to Mike (mcharles here) at http://www.bigerc.com -- he has a spot welder.

 

[Beagle123]

Thanks xter. I'll keep that in mind. I'm going to have a real challenge getting them into that battery box. Its really tight. Its even harder to connect the wires. My first attempt was a decent try, but I may have to put some into the neck.

My first goal is to make a really good battery tester. I want to know exactly how good each battery is. I also have to make a manual charger so I can charge them equally. Then I'd like to make a system where the batteries aren't soldered or welded into place, so if anything else happens I can remove the bad cells.

My ideal solution is plastic boxes with the batteries held in place with lots of friction.

I'll keep you posted.

 

[dirty_d]

i think you should solder them, its easy enough to de-solder them if you need to remove bad cells.

 

[Beagle123]

Batteries Still Aren't Performing Properly

I've been load testing batteries from all my packs, and they don't seem to be able to handle larger loads. The chart below that xter posted doesn't represent the results that I'm getting from these batteries. I think this may be propaganda put out by the company.

These batteries seem to respond pretty well to small loads, but when you run 1C - 1.8C loads, the voltage drops severly. I tested three batteries:

Battery X = battery from Pack #5 (should be perfect)
Battery Y = battery that <u>was never installed</u>
Battery Z = battery from Pack 2 (may have had a short)

They're all about the same:


<table border="1" cellpadding="5">
<tr><td> Battery </td><td> V (before) </td><td> 2.5 ohms (1.5 amps) 0.5C </td><td> 1 ohm (3.2 amps) 1.1C</td><td>0.5 ohms (5.4 amps ) 1.8C</tr>
<tr><td>X</td><td>4.01v </td><td>3.72v</td><td>3.25v</td><td>2.60v</td></tr>
<tr><td>Y</td><td>3.95 </td><td>3.65</td><td>3.25</td><td>2.53</td></tr>
<tr><td>Z</td><td>3.95 </td><td>3.66</td><td>3.20</td><td>2.53</td></tr>

</table>

The chart below is clearly wrong. All my batteries are testing the same.

As a comparison, one of my 18650 cells drops 0.12V under 1.6 amp (~3/4 C). A subpack of 15 droops 0.25V under a 33 amp, 1C load. And the entire pack of 20 subpacks sags 6V under full, 35 amp load -- some of the variance of course due to the wiring and connections.

Xter, did you measure the amps with a multi-meter or did you compute them with math? I know you're using four 10 ohm resistors together, which should produce 2.5 ohms, but maybe you're getting more resistance than you think?

My batteriess that are sagging 1.4 volts are measuring similarly to yours (mine sag 0.19volts @ 2.5 ohms whereas yours sag 1.2v) So your batteries must be better at higher loads than mine.

I don't think there can be anything wrong with my test because I'm using 2 multi testers to test current and voltage, and all my batteries seem to be testing the same.

I'm going to try charging some manually, and test those with this same test.

But its looking like these batteries are only good for about 1C or so.

 

[safe]

Conductive Epoxy carries with it the idea of the material "doing something" like holding the batteries in connection with each other. That's the real problem I suspect, that in trying to replace solder as a connective medium you've basically recreated all the old problems again.

Instead of a rigid connection (of any type) you need to migrate towards something like Xysters battery with the built in flexibility in the design.

I'm going to experiment with the silver paste (or nothing at all) but I'm also going to physically provide mechanical force (a spring) to press the batteries together.

The moral of the story seems to be that rigid connection is bad, not so much that conductivity can't be achieved...

 

[fechter]

Something is wrong somewhere.

When you do your discharge test, make sure you are measuring the voltage directly on the battery with a separate set of wires. You don't want to measure the voltage drop in the wires going to the load.

I really didn't think the epoxy stuff was going to work, but nobody ever tried it. Thanks for learning the hard way for the rest of us.

Spot welding tabs or soldering to existing tabs is the best way to go.

You could possibly try stacking them in tubes and using some kind of spring to maintain compression. Sort of like a flashlight. This arrangement is prone to oxidation on the ends of the cells over time, but it's not that hard to take apart and clean once in a while. You also need to compensate for thermal expansion of the cells, which could be considerable.