Mistake busbars ?

wim

Well-known member
Joined
Apr 6, 2017
Messages
560
Hi all,

Did i made a (big)mistake in the design of the busbars in my setup ? :huh:
Afew members made a correct observation in my build thread, so just a bit unsure
I know the general idea of the busbarson the packs is to dothe terminals of thepositive side opositto the negative side. This to distribute theload moreeven over all the cells in the pack... I do understand this, and do believe it is the best way... no discussion there.
I had to do it otherwise due to design reasons.

I want to know how bad it is in our type of systems to do it my way, withboth terminals on the same side ?

Are the cells charged/discharged uneven? are they getting (temporarily) out of balance in one and the same pack? hard to believe...
Are the cells in the front of the pack more stressed?
Is a setup with a (even a small)difference in lenghts of fuse wire not worse?
Internal resistance beween cellsin a pack, same problems?

In my setup busbar resistance is negligible ( 5 x 32cmcopperwires of 6mm2 in "parrallel"is0,0002 Ohm, even less with the "middle"busbar connection)
Currents more than 150A (75A / string)arerare,there are90 cells in a pack.:
More specs here: http://secondlifestorage.com/t-Instant-powerwall

Is there a way to do some tests?

My way:

image_ztwezx.jpg


Other (right) way (thanks Peter)

image_qfwujr.jpg
 

Korishan

Moderator
Staff member
Joined
Jan 7, 2017
Messages
6,325
It's not a "big" mistake, and possibly not even a "mistake", per se.
But I would venture into the area of possibly not recommended.

You are correct as you will have a little bit of imbalance. However, this isn't as much of a big deal as imbalance in serial cells. Parallel connections can get imbalanced for a little while with little to no issues. I doubt you'd have more than 100mA/mAh difference between cells close to the busbar, and those at the other end. Especially after resting for a few minutes.

If you think it's a big issue, you could always just move the one bar to the other end. Your design makes doing that a big easy to do. Albeit, the bar moved to the other end just won't have solid copper to the bar from the pack (unless you resoldered those runners, which, might not be feasible as you have a bunch of fuses)

I personally don't believe you'd have a lot of issues with it.
 

dougal

Member
Joined
Sep 13, 2017
Messages
43
Surely it is a total non-issue with *adequate* bus bars?

What matters is the difference in resistance between the pathways to the different cells. I can't see there being a significant difference with plenty copper in the barsalong the pack. Sure more is better, but even the manufacturing differences between your (all new) cells are going to be more significant.
Nickel has a much higher resistance than copper, and some of the nickel strips being used are quite thin. There probably is the potential for such an effect with nickel strips, but not with a fewsquaremm cross-section copper, and you have 30mm2...

As a counsel of perfection, certainly opposite ends, equalising the path through each cell is ideal. But not needed if your bus bar resistance is insignificant,

The opposite ends design is less convenient for racking, but does have a safety advantage in keeping the terminals as far apart as possible!
 

jdeadman

Well-known member
Joined
Mar 28, 2017
Messages
926
The Variance would only be evident is Really high Amp draw conditions. We are talking about max discharge so around 1A per cell and in an 80p packs that's 80 Amps. Not sure what your max load is going to be but even with 80Amp draw the difference we are talking about is .01V from end to end. and if you are leaving .1v wiggle room at the top and bottom you still have a lot of wiggle room. For example i run my Wall from 4.1 to 3.6V.

I did do a test when i was first looking at pack design (and I'm still making changes as I go lol) and took 8 cells hooked them together in parallel with the power leads off cell 1 only. Cycled the pack a few times then pulled a high drain (8Amps ) from the pack and then removed the batteries from the holder right after it was done and I only saw 2 batteries that were different and they were in the middle so I think that was just a happenstance of the cell being not exactly the same as the others in capacity or max draw

So personally I would not sweat it at all.
 

Hanssing

New member
Joined
Jun 6, 2017
Messages
15
You variance due to difference in individual cells are easily 10mOhm or more ofr used cells.

Therefor your design is fine, since you have bus-bars with resistance which is a lot less than IR-variance.

Secondly, is only relevant at extreme currentloads.

And even with small busbars, the problem is stressing of front cells compared to back cells, and some extra charge-discharge cycles locally in the packs whilst it eaqualizes - so some coulomb-losses (<5% of the imbalance-charge), and some wear on the cells - not optimal for teh pack, but possibly optimal in other regards such as access to maintenance. You get some and you loose some.

Your current* design is fine, dont sweat it.

* pun intended
 

daromer

Moderator
Staff member
Joined
Oct 8, 2016
Messages
5,391
As said its not a big deal as long as you keep your max current low and you have thick busbars. I have packs like that too and still no issues 1 year later... Not ideal no but it works.
 

wim

Well-known member
Joined
Apr 6, 2017
Messages
560
A little test, with 2 V-meters, one in the front of the pack and one in the back, both on the same bus-bar in the middle, this is te one with 2 rows of cells.


image_niuvio.jpg


In the cupboard,left top corner, without load:

image_ogwsia.jpg

image_zditjc.jpg


With a 10minload of 130A, still the same voltage :

image_zlnnav.jpg

image_fiwntt.jpg


And after 45min "medium"load of+/- 60A, nodifference between the voltages.

image_ainixk.jpg


Ialso measuredthe voltage between the common-pack negative and the positive on a random cell in the pack (as Daromer suggested ) with a handheld V meter, and there is no difference between the cells ( only a 0,01V precisionmeter )
So, maybe some unbalance in the 0,001V range on aheavy load, but i think we can put this to rest and carry on happy :)

I will leave the Volt meters in there for a while...

suggestions arestill welcome.

My build: http://secondlifestorage.com/t-Instant-powerwall
 

Korishan

Moderator
Staff member
Joined
Jan 7, 2017
Messages
6,325
The volt meter wont show discrepancies while they are hooked up. you would need to disconnect the cells to be tested immediately after drain, if not while dain so you can see it before and during bounceback
 
  • Like
Reactions: wim

wim

Well-known member
Joined
Apr 6, 2017
Messages
560
Korishan said:
The volt meter wont show discrepancies while they are hooked up. you would need to disconnect the cells to be tested immediately after drain, if not while dain so you can see it before and during bounceback

Very much true, but disconnect the cells immediately after drain is a bridge to far for me, before they "self-balance"i aim not fast enoug for that... sorry ;)

I believe the un-balance is going to betoo small to messure with my voltmeters, maybe when discharged very fast ( way beyond normal use ).

This kind of testonly shows, ashoped,there is no significant voltage drop in the lengt of the bus-bars under load, no voltage drop means the cells in front of the pack will not ge(much)more stressed then those in the back.

If there is a way to test without having to dismantle the pack, i would do it... i can not think of one....
At the end, for suretime will tell i guess....


dougal said:
wim said:
A little test, with 2 V-meters, one in the front of the pack and one in the back, both on the same bus-bar in the middle, this is te one with 2 rows of cells.
...

My build: http://secondlifestorage.com/t-Instant-powerwall

I think your test demonstrates that your mini-meters are consistently calibrated!
Yep,they even blink between thenumbers at the same time... :Dwas Lucky ... at lastsome great chinese stuff ;)
 

Korishan

Moderator
Staff member
Joined
Jan 7, 2017
Messages
6,325
What you might be able to try is to take the Neg of both meters and connect them at the same point on one end. Then take the Pos of one and connect it to the same end and the Pos of the other and connect at the other end. This "may" give you a voltage difference. It's just that the two meters must have a common starting point. Connecting them completely at either end won't yield any differences as they can't show you the true load across the pack.

But, then again, if your buss bars large enough, they won't show any differences anyways as the voltage drop is too low, or non-existent. You could test your voltage drop across the buss bar by using it as a shunt type of connection. Put a load across the buss bar and see what the voltage drop is at either end that way. This would allow you to use your multimeter in .0001 mode instead of the .001 or .01. We don't care about full voltages, we care about the milli-volts or micro-volts.
 

wim

Well-known member
Joined
Apr 6, 2017
Messages
560
Kind of did that with my multimeter ( common negative rail in the front and on the positive side on the cells in the back and front )
my meter is not sensitive enough (millivolts only) to detect a difference.

Microvolts...do you think that low is of any importance?
 

Korishan

Moderator
Staff member
Joined
Jan 7, 2017
Messages
6,325
Depends on how often you pull on your pack that hard. Wear and tear is cumulative. So, there's not much difference between 1000 hits of .0001v, or 100 hits of .001v. It still kinda does the same thing, kinda. It's probably more like 2000 hits of .0001v is equal to 100 hits of .001v. Just theory at this point :p

Actually, I guess I mis-typed that. When the MM has no decimal reading, it is reading millivolts. Microvolts would be 5 decimals. So, really, what I meant/should of said was reading in decivolts (0.01V), centivolts (0.001V) or millivolts (0.0001V). Most MM's won't read in centivolts, though. You jump from decivolt to millivolt. The problem is when reading in millivolt mode, you can't go over a few thousand or the meter goes wonkers. Which is odd, cuz you'd think it should be able to read up to 9999 millivolts, or close to. At least read up to 5000mV or so.

Ok, I hope i didn't confuse anyone with my rambling set of corrections of myself :p
And please feel free to correct my corrections that may not be correct, or partially correct, or somewhat correct.
 

wim

Well-known member
Joined
Apr 6, 2017
Messages
560
One correction on the corrections :) you are a digit off.

1,0 = 1 Volt
0,1 = 1 "Decivolt" or 100 millivolts
0,01 = 1 "Centivolt" or 10 millivolts
0,001 = 1 millivolt
...

i do not think decivolts and centivoltsis a real thing ;) at least not here.

So my MM is accuratup to10 millivolts, as most "cheap" ones are...

Wear and tear cumulative ? makes sense... dont know...
It is never going to be easy to detect/messure, so it stays a theory i am afraid....
 

Korishan

Moderator
Staff member
Joined
Jan 7, 2017
Messages
6,325
@wim: thanks for the correction on my correction that needed a little more correcting. ;)

Well, I wouldn't quite say deci/centi-volts aren't real. Most MM's are either xxxx, xx.xx, xxxx. Or, at least the few different ones I've used only show those 3. So we readily use, or have access at least, to all 3. On packs that are over 4 volts, I'm sure we all use the deci-volt mode as we only need to see the tenth of a volt. I'm not sure how far the MM will go in centi-volt mode. I know they have kick offs at certain voltages and I don't have a variable PS to test mine (yet).

Hrmm, actually, there are 5 settings. HV (High Voltage; not sure what it's range/max is), 200 (00.0), 20 (0.00), 2000m (000), and 200m (00.0 ?? I'm guessing this is fractional millivolts, or, pico-volts. it's sensitive enough that moving the wires slightly generates a voltage).

But we are really interested in the main 3. However, you could use the 2000m setting to detect voltage drift across the pack as it's that sensitive. I'm guessing you could use 200m as well, but you'd need to not move the wires afterwards to make sure you get an accurate reading.
 

SimonW

Member
Joined
Aug 26, 2017
Messages
89
Haha, I've been involved in electronics in one way or another(hobby/professional) for over 35 years and this is the first time I've heard decivolts and centivolts.
Sounds really weird.
 

Korishan

Moderator
Staff member
Joined
Jan 7, 2017
Messages
6,325
Yeah, we actually don't use that "term", but we use the numbers, though ;) I think most ppl would say 10 millivolts instead of 1 centivolt, and 1 tenth volt rather than 1 decivolt.

When i was writing it up, I was basing it on the fact that mill is metric, so I went with metric on the other decimals. I guess some things just don't translate well :p
 

wim

Well-known member
Joined
Apr 6, 2017
Messages
560
Yep, from Volt straight to millivolts... thatswatwe do. :cool:

Then you say:

Hrmm, actually, there are 5 settings. HV (High Voltage; not sure what it's range/max is), 200 (00.0), 20 (0.00), 2000m (000), and 200m (00.0 ?? I'm guessing this is fractional millivolts, or, pico-volts. it's sensitive enough that moving the wires slightly generates a voltage).

But what you realy want to say is : "Go andbuy yourselvea decent multimeter"is it not? :D :D

My multimeter starts at 2V as a lowest setting, so you are probably right... :rolleyes:
 

Korishan

Moderator
Staff member
Joined
Jan 7, 2017
Messages
6,325
Well, mine are the "cheapo" harbor freight ones that we got as freebies with their coupons. However, the board model number on the casing is the same model number that's in the Lowe's electrical section that sells for >$30. So, just because the case is cheapo, doesn't mean the guts aren't :p Of course, I guess that could go the other way, too. Just cuz the outside piece is expensive, doesn't mean the innards are worth crap ;)

But I've checked my 3 MM's and calibrated them best I could. Fairly easy to do. It does everything I need it to do for now. I did have to buy a separate amp-meter, but I think most of would have to do that anyways if we wanted that feature.
 

SimonW

Member
Joined
Aug 26, 2017
Messages
89
Can get some decent meters for next to nothing these days.
Mine hurt, it cost me about $600 about 20 years ago... though it's still pretty decent even by today's standards, was better than the competing flagship Fluke of the time. (Autoranging 40,000 count and 0.06% basic accuracy (DCV)
 
Top