Voltage Drop across Parallel Bus in Large Parallel Arrays

S

skyfridge

Member
Joined
Mar 9, 2017
Messages
149
In all of the photos that I see of DIY Powerealls, I've noticed that the parallel buses are straight down the row of parallel cells. In researching something on a related concept, I came across a design for balanced parallel wiring.


image_nwlmbc.jpg


Source:http://www.iotaengineering.com/pplib/balancedcharging.pdf[/SIZE]

Since balancing seems to be such a hot topic regarding DIY battery packs, it's surprising that I haven't seen this subject addressed, sincesuch large parallel arrays are being built.

Has anyone measured any significant voltage drop across parallel buses during charging or discharging in the typical arrangement?
 
Well that is certainly interesting, and food for thought
 
I think the CV portion of the lithium charge cycle reduces the need for this sort of cable arrangement, by the time the charge current falls nearly to zero the voltage drops have all also fallen nearly to zero. I suspect this is more important in lead acid cells where they tend to keep a charge current going more or less continuously.
 
Elmo said:
I think the CV portion of the lithium charge cycle reduces the need for this sort of cable arrangement, by the time the charge current falls nearly to zero the voltage drops have all also fallen nearly to zero. I suspect this is more important in lead acid cells where they tend to keep a charge current going more or less continuously.
Click to expand...

That's a good point, and I had considered it. Lead Acid uses the same charging profile. Only the voltages change. CV is called Absorption, and the current does fall off during absorption, just as it doeswith Lithium. For Lead Acid, absorption is usually timed. For Lithium, CV is usually stopped at a certain low current limit.

It depends on where you have chosen to call it fully charged. I'm not intimately familiar with LiPo. With LiFePO4, it has beenrecommended by some with experience, that if you charge up to 3.65VPC you stop. Charging at 3.5 you go to C/20. At 3.45, you go to C/100. That's from memory, so the exact numbers might be off. Those profIles were based on tests of capacity. If you're trying to keep it between the knee and the elbow of the discharge curve to prevent overcharging or overdischarging due to imbalance and stresses at the voltage extremes of the curve, you'd do likewise with LiPo.

Regardless,if there's a voltage drop due to unequal lead lengths among cells, some cells will be getting more current than others. Any interruption in charging might cause a cascading current rush from high to low voltage.

Iotas's theory makes sense to me, but I haven't tested it, because I opted to use balanced wiring. I wondered if anyone had measurements that would support the theory.

Experiment

Experiment
I somehow broke the link above.
 
As said. For LA this is very very important if you want all your battery packs to be equal. You can easily measure above scenario. If you have just slightly thinner cable it gets quite obvious. On Li.* Its not as important as stated but nevertheless you should still consider designing it in such a way that you try to get even draw and charge to all cell.

If you dont and you have a bit to thin cables you may stress the cells at one end far more than others. Worth noting is that most of us have rather thick busbars and cables... and that saves us some grief. My recommendation is to take all packs with same wire length to a busbar. Though im probably not going to do that due to how my design is... And up untill now i have slightly different lengths.
If i go in and measure my testpack that is "the wrong way" I can easily see 20mV difference between pack just seconds after a hard-charge. But this levels out after some time.

With running some kind of monitoring that monitors the cells voltages you can also sense the actual voltage and have the inverter/charger to set voltage to charge to based on the cells condition. In some cases this is useful though perhaps not relevant in today's discussion...
 
I'm no expert but I think a lot of people are happy to do method 1 to improve balance during charging. That is the method I will be using.
 
Yeah, I'm kinda confused with the way Method 2 is done. Not sure how it would work adding multiples of packs. And besides, Method 2 would use a LOT of extra wiring, and get confusing really fast for those of us who plan on having a LOT of packs; ie. Pete ;)
 
skyfridge said:
Since balancing seems to be such a hot topic regarding DIY battery packs, it's surprising that I haven't seen this subject addressed, sincesuch large parallel arrays are being built.
Click to expand...

Balancing is important for us once we start to put 3.7v packs in series ....but I thinkyou are raising the this point in relation to charging one 80p pack ...

This is not something we need to worry about ... the bus is so thick In comparison to the internal resistance of each cell ... andthe internal resistance of cellsvary widely .....

It's more important for LA because the internal resistance is so low , ... I think your link is in error , the currents would never vary by so much.... one battery is shown to draw 9A another 18A if this were to occur the voltage of the battery taking 18A would rise more quickly than the rest ,causing it to take less and the others to take more ... it would all balance out with the ones nearest the charger being just a few mV higher ...
 
Korishan said:
Yeah, I'm kinda confused with the way Method 2 is done. Not sure how it would work adding multiples of packs. And besides, Method 2 would use a LOT of extra wiring, and get confusing really fast for those of us who plan on having a LOT of packs; ie. Pete ;)
Click to expand...

The balanced parallel wiring would only need to be at the ends of the series of packs. No other extra wiring would be required, because the resistance and thus the voltge dropon each end would be balanced on the other end.

ozz93666 said:
skyfridge said:
Since balancing seems to be such a hot topic regarding DIY battery packs, it's surprising that I haven't seen this subject addressed, sincesuch large parallel arrays are being built.
Click to expand...

Balancing is important for us once we start to put 3.7v packs in series ....but I thinkyou are raising the this point in relation to charging one 80p pack ...

This is not something we need to worry about ... the bus is so thick In comparison to the internal resistance of each cell ... andthe internal resistance of cellsvary widely .....

It's more important for LA because the internal resistance is so low , ... I think your link is in error , the currents would never vary by so much.... one battery is shown to draw 9A another 18A if this were to occur the voltage of the battery taking 18A would rise more quickly than the rest ,causing it to take less and the others to take more ... it would all balance out with the ones nearest the charger being just a few mV higher ...
Click to expand...

The examples of voltage drops in the linked document were obviously exaggerations. With LA, there is a rule of thumb about not exceeding a maximum number of batteries or strings in parallel. Though I don't recall that number, it was two or three strings. Having parallel strings ofbatteries of celast in series also complicates the ccalculus .

Now, we're talking 40p, 80p, 196p (in another thread here), etc. Any voltage drop from one end of the parallel bus to the otheraffects the next parallel modules in series. As a result, cells at one end of the module have higher current than the other.And now we're talking 3sand up.Especially when we'retalking about cells with unmatched Ri, where we'really hoping that everything averages out, we should use every trick in the book to improve balance.

It's just another detail that might improve the life cycle of our powerwalls and something that could be investigated.
 
On second thought, it might not be so exagerrated with higher currents. Some types of LA batteries have maximum charging currents of 4C (if I recall correctly).

Although, for the specific examplexamples current it is highly exagerrated .
 
Correct #n: I can't speculate that the example is an exaggeration without a circuit analysis. Without one or more figures, such as voltages, resistances, I'm finding it difficult to do so.
 
This most definitely is not an issue .... the average bus is about 10 square mm !!! that's about ONE milliohm PER METER !!!

Internal resistance of cells varies greatly 50 to 100's of milliohms per cell ...so the resistance of the bus is immaterial ...

Even with LA this is nonsense ...with a 10 sq mm connecting cable there will be (effectively) no difference in the charging ...

Only if the LA batteries were massive ... spaced far apart ... and had thin connecting cable .
 
ozz93666 said:
This most definitely is not an issue .... the average bus is about 10 square mm !!! that's about ONE milliohm PER METER !!!

Internal resistance of cells varies greatly 50 to 100's of milliohms per cell ...so the resistance of the bus is immaterial ...

Even with LA this is nonsense ...with a 10 sq mm connecting cable there will be (effectively) no difference in the charging ...

Only if the LA batteries were massive ... spaced far apart ... and had thin connecting cable .
Click to expand...

That's a little too dismissive and can't be taken seriously. Lead Acid batteries generally don't have cables welded or soldered to the terminals, which introduces electrical contact resistance. Even if the resistance of the cable is negligible, the ECR isn't when it isn't balanced with balanced wiring.

Korishan and I are looking at using battery contacts, instead of welding or soldering, so it will involve similar issues of balancing the ECR. Balanced parallel wiring offers an advantage that could mitigate imbalance.

Also, your assumptions might be relevant in the case of using cells with unmatched internal resistance. How does that correlate for someone with more closely matched cells? I have new "old stock" cells. Someone might try to match Ri of salvaged cells. Your assumptions wouldn't apply.

Also, is everyone using #7 (10mm^2) cable? How about for test beds? I have #10 on my test bed. Something else might be more appropriate for the battery assembly. With balanced wiring, it might not require #7.
 
The only way that the pictures would be true is if you were using undersized Cabling. This is yet another example of why we need to be using oversized Wiring.
The 2 things they don't provide in the pics are voltage and cable size. These are both critical assumptions. The only way one pack is getting less than the other would be due to the voltage drop that is getting to it.

For instance: the example would be correct if you were using 12ga wire at 12v and 50 amps.
It would not be an issue if you were charging 100v packs because the % voltage drop would be negligible, and likewise it wouldn't be true if you were using 2ga cables @12v. For single cell parallel packs @3.7v you may need to go up to 0 or 0/0 cables to observe the same % voltage drop.
 
Aspendell said:
The only way that the pictures would be true is if you were using undersized Cabling. This is yet another example of why we need to be using oversized Wiring.
The 2 things they don't provide in the pics are voltage and cable size. These are both critical assumptions. The only way one pack is getting less than the other would be due to the voltage drop that is getting to it.

For instance: the example would be correct if you were using 12ga wire at 12v and 50 amps.
It would not be an issue if you were charging 100v packs because the % voltage drop would be negligible, and likewise it wouldn't be true if you were using 2ga cables @12v. For single cell parallel packs @3.7v you may need to go up to 0 or 0/0 cables to observe the same % voltage drop.
Click to expand...

I asked the wrong question. There isn't an issue when you're seeing,

image_alrmwa.jpg

Source:Why is voltage constant in a parallel circuit?

The story changes when you're involving,

image_nouveb.jpg

Source:Resistor Networks

Whereas the bus bar of the common DIY battery pack doesn't truly have the same voltage at each connection to a cell, the difference is negligible.

With lead acid batteries the parallel bus commonly goes through multiple connections (ring terminals). Differences in cable length might be negligible, but differences incumulative Electrical Contact Resistance that result from different numbers of connections is not negligible.No matter how tightly you wrench down those terminals, it will add up.

I'm working with the same kind of setup. It's a completely different scenario. I didn't recognize it clearly. Oops! Never mind.

I did the math. :)
 
You must log in or register to reply here.
Back
Top