Max safe Voltage Difference when connecting two Batteries in parallel

hias9

1 kW
Joined
Jul 11, 2018
Messages
422
What would you consider the maximum safe voltage difference when connecting two 14S10P 18650 batteries in parallel?

Is it possible to calculate the maximum equalizing current? ( maybe U(difference)/(Ri1+Ri2) )
 
Calculating the flow would involve knowing the internal resistance of each battery, as well as the resistance of the interconnect.

As a guideline, I try to make sure they're within .01V per cell of eachother, so .14V in your case. Then you shouldn't have anything to worry about.
 
none will probably be the only safe answer.
such a battery will have possibly thousands of amps that can flow if the difference is big enough.
 
If you have a cycle satiator or other meter to show you what's happening during charging you can get some basic info on your pack's tolerance for voltage differences.

My 12S6p Panasonic PF pack, when about half empty, requires a delta of 0.6 volts to raise the charge current to 8.0 amps, which is C/2.

So, under similar charge state conditions for these particular packs, I am willing to parallel them when they are within about 0.5 V or less of one another.
 
That answer would depend a lot on the voltage you are talking about. You need to think about % difference.

My rule of thumb is I never worry about a 1v difference between two 48v packs I'm connecting. So this would be approximately a 2% difference. On a single 4.2v cell, a similar difference would be .1v. I see no problems with any differences that don't amount to nearly as much current flowing as when you hit your throttle. It won't stress the cells, or the wiring. I'll admit to pushing this as far as 2v, but on cells that can handle a lot of c rate. ( lipos) Low c rate cells would of course need to be closer to identical voltage, such as reclaimed laptop cell packs.

Clearly, if any wires warm up when you connect, it was too different. What's happening to cells if the wires get warm is even worse.

What you cannot ever do, is connect a 13s pack to a 14s pack. Its not that the two packs are 4v different. Its that adding another 2v to the full charge voltage of the 13s pack may damage the cells in that pack. So if you are in this situation, you have to discharge that 14s pack a little bit, before you connect up. Then,,, you still have to stop when the 14s pack cells are empty, not below it.

Better off in that case, to just never parallel the packs, they are not similar enough to do it.
 
dogman dan said:
Clearly, if any wires warm up when you connect, it was too different. What's happening to cells if the wires get warm is even worse.
I've seen people use a certain gauge & long length of wire (to get a specific resistance?)

for the purpose of safely getting an empty (smaller) pack to the same voltage as a full (larger) one.

There definitely was a temperature rise in the wire, but not a measurable one in either battery.

If I tried this, under guidance, starting with smaller V delta, using math, I'd want to measure - ideally log & graph - the Amps rate change over time, and totalize the Ah transfer.

You really think the wiring getting hot in those circumstances is a symptom of damage to the packs?

 
Yes, 1 V is for two identical 14s 18650 batteries very conservative and safe voltage difference. Knowing more, mainly cells type, but also connecting wires diameter and length, connectors, switch, BMS,….. we could calculate more precisely safe voltage difference which is surely higher than 1 V.
 
Cells are 3500mAh GA Cells which have an internal resistance of 50mOhm according to the sheet.
BMS allows 60amps continuous discharge and 50amps continuous charge.
If we ignore the wire resistance, it would be 2x70mOhms for two 14S10P packs.
So if we don't want to exceed a charging current of 1C, the current flowing from one pack to the other should not exceed 35amps, safe voltage difference would be 4.9V (.35V per cell). And in fact the current will be a bit lower because of the wire resistance.
Is this correct or did I make a mistake somewhere?
Maybe the internal resistance is lower than in the sheet, but I think until .2V per cell it should be fine.
 
Your initial thought in your first post on equalizing current calculation is generally right. We just add some points. You now also correctly concentrate on max charging current, because is lower than discharging.

DCIR depends on aging of the cells. Fresh new Sanyo GA is about 34 miliohm at 3,74 V and 25 °C.
First let calculate the internal resistance of the cells only for one battery :

(34/10) x 14 = 47,6 miliohm. For two 14s10p packs is total internal resistance 47,6 x 2 = 95,2 miliohm. We can round this number to 100 miliohm = 0,1 ohm to simplify it. Moreover fully charged cell has DCIR little bit higher.

We don´t want exceed max charging current 0,3 C = 1,1 A for GA cells. Such capacity as 10p battery can probably keep significant equalizing current in order of minutes or more.

If we consider 1,1 A per cell which is 11 A per battery, than we get total voltage drop 11 x 0,1 = 1,1 V.
It is theoretical voltage drop on the cells only.

Now we have to add resistance of other parts like wires, transient resistance of the connectors, nickel, spot welding resistance, BMS, … For simplification let consider 100 miliohm, it is another 1,1 V drop.

Resulting safe voltage difference for this simplified example is 2,2 V.
 
I don't know exact values, but I think the resistance of the other parts could be less than 100mOhms, so it would be safer to calculate with a low resistance.

Why would you not go over 0.3C maximum charging current? The current we are calculating is only a peak value, so I think a bit more than the rated current would not cause a big damage to cycle life.
Would 0.5C be okay for quick-charging this cell or would this already have a significant impact on cycle life?
 
OK, you are right, if we don´t know exact numbers, will be better calculate instead lower value, maybe 30 – 40 miliohm. It is also safer for wiring and connectors.
Now you have better picture and you may play with the numbers.

You can be sure that voltage difference 1,1 V is absolutely OK and very probably 1,5 V is still on the safe side.

Yes, regular 0,5C CC/CV charging has significant negative impact on GA cycle life.

I have an idea to simulate this situation with two cells and measure current, so maybe evening I´ll post if got any meaningful results.
 
If your Fast Charge rate is only used when really needed, say 0.5C only 10% of the time

especially if you ensure high cell temps, say well over 25°C when fast charging

and then your normal charge rate 90% of the time is say 0.2C

IMO that won't impact longevity much.

 
I plan to quick-charge two 14S10P packs in parallel with a 1800W Flatpack S when making a break. Current would be around 0.45-0.5C.
When riding alone, charging a single pack at 1800W would mean 0.9-1C which will probably have a significant negative effect on cycle life.
Are there any numbers for these cells regarding how quick-charging affects their cycle life?
 
Here is measured current course of two LG MJ1 cells connected in parallel. Voltage difference at the start was 0,153 V peak current 1,139 A. Calculated total resistance 137 miliohm consist of 75,8 miliohm cell A + B internal resistance, 10 miliohm measuring resistance, remainder is wiring and connectors transient resistance.

Sanyo GA is not identical cell to LG MJ1, however current course will be similar.

Parallel LG MJ1.jpg
 
hias9 said:
Are there any numbers for these cells regarding how quick-charging affects their cycle life?
To many variables involved.

High temperature, even well above 30°C definitely reduces the "damage" (lost cycle lifespan)

I'd stop at 35 myself. But it's **internal** temp that matters, so big time lag from the box-air ambient reaching the target, I'd say at least an hour. So maybe start out slow then ramp up?

 
hias9 said:
I plan to quick-charge two 14S10P packs in parallel with a 1800W Flatpack S when making a break. Current would be around 0.45-0.5C.
When riding alone, charging a single pack at 1800W would mean 0.9-1C which will probably have a significant negative effect on cycle life.
Are there any numbers for these cells regarding how quick-charging affects their cycle life?

Charging GA at 0,5C is not good idea if DOD is higher than 50 % . I again recommend to read Pajda´s posts :

As I previously say that I do not like generalization, I can say that the Panasonic/Sanyo production line in 18650/21700 does have significant problem with CC continuous charge rate higher than 0.3C

https://endless-sphere.com/forums/viewtopic.php?f=14&t=102387&p=1507776#p1507776

Information about Samsung 35E and Sanyo GA cycle life under 0.5C-1C 100% DOD are correct, both suffer with 0.5C charge rate.

https://endless-sphere.com/forums/viewtopic.php?f=14&t=102387&p=1498545#p1498545

under what settings? My GA samples works fine under 0.5C-1C at 50% DoD(4.2V-3.55V) with only small (up to 5%) capacity loss after 500 cycles. This behaviour is typical for almost all cells on the market. Yes, under a high DoD GA losses capacity from the beginning very quickly, but after reaching about 500 cycles, the capacity drop is slowing down.

https://endless-sphere.com/forums/viewtopic.php?f=14&t=96079&p=1412148#p1412148

And here is Panasonic official document regarding to GA cycle characteristics :

 
The correct answer is to discharge the high pack or charge the low pack until they are within a couple volts.

My field answer is that I have always connected high rate (20C and up) large packs (>10Ah) that are off by 5V or 6V and I have never seen any fire works (or even sparks).

How the currents flow and how that degrades the internals. . . well... if you are worried about that then just match them.

A simple test can prove this either way.

Pack 1 at 44V
Pack 2 at 50V
6V of Delta
Assume 10Ah of 20C so a continuous rating of 200A with much higher burst capability

Go to the store and buy a 1 ohm 50W resistor

OhmsLaw.png

Wire that between the packs
Attach them together

...

Measure the voltage across the 1ohm resistor
It will probably be a few volts
Use ohms law
Prove that only a few amps are flowing

...

Now scale that down for 100mOhms

-methods
 
Maximum possible voltage could be:

50V - 44V = 6V

Minimum possible inline resistance could be:

100mOhms + Wiring + other

I = V / R

V = 6V
R = 0.1Ohms
I = 60A

So...
In that case...
The most you could possibly see would be 60A
(and you wont)

60A / 10Ah = 6A per 1Ah cell
That would be 6C

Packs are rated these days for 5C charge. . . so . . .

Right on the money

... Measure your current with a shunt, it does not lie. I have shown previously how to do this with a CA.

1) Power your CA seperately
2) Wire your CA shunt inline as described above
3) Read the DC Current

You are not worried about AC or Inrush or Spike
Just steady state DC

The average CA has a 0.001ohm shunt... so... in this case it will not affect things all that much.

...
Do it with 1 ohm
Do it with 100mOhm
Do it with 1mOhm

Graph that in EXCEL, get a straight line?

...

Clamp on a DC Meter -
Non-inline
FLUKE
Calibrate it

Directly measure the actual current with no shunt inline

...

Those are solutions, not speculations.
Let us know how it works out.

-methods
 
methods said:
Go to the store and buy a 1 ohm 50W resistor
Wire that between the packs
Attach them together
Measure the voltage across the 1ohm resistor
It will probably be a few volts
Use ohms law
Prove that only a few amps are flowing
Now scale that down for 100mOhms

dont do this.

chinese battery makers that put 2 batteries parrallel in a scooter also do this crap. this is the result even if you glue the resistor against the metal frame:


jHSvEyal.jpg


this battery is 3 months old.
 
Proof that the resistor in the picture does get too hot for the foam. The foam can't take much heat, that is certain sure.

What's safe in china, aint safe in my garage, as I found out one night.
 
There is a big difference between measuring / attending / conducting experiments, and putting the results into production once you've determined what is safe.
 
flippy is in straight up "Vendor Mode"

I clearly described an experiment that would put to rest any questions about how much current flows between two packs when you hook them in parallel.

I described it using a large resistor for easy measurement with any DMM in a low risk configuration
I indicated that this was not representative and provided guidance on a smaller shunt
I suggested that one graph this to see the linear relationship
I then suggested a method to directly measure the currents without using a shunt to get the actual pack to pack current.

...

I would never wire packs in parallel with diodes or resistors in a running configuration...

and

I know how to calculate power across a resistor in Air Cooling, packed into foam, or up against a heat sink. You use a factor of 4 for air cooling and a factor of 10 for potted*

....

Oh well
If he sells good packs then so be it.

-methods
 
It would greatly depend on the cell but in my experience throwing a 30q cell at 4v in parallel with a 3v cell I can’t say for sure if there was even a noticeable increase in temp. I forget the details on the cell resistances and subsequent flow of current but the charge current was way beyond the stated max charge rate for the cell and likely would’ve been more heat n damage if at a higher voltage

If you’d rather get the pack voltages closer the highest wattage discharger I’ve found around the house is a hair dryer. Still far from fast.
 
methods said:
flippy is in straight up "Vendor Mode

no, i am in "dont burn your house down" mode.

dispite me being a battery vendor has so far never resulted in any financial gains from this forum. i am not here to shill services, i am here to help people. try and get that in your head instead and stop doing this condecending behaviour.

using resistors to limit current is bad engeneering and gives unskilled people that read a single topic or posting the idea it might be fine to do.
 
Back
Top