In paralell connections, different chemistry OK, different ah size OK but best if not hugely different. Different voltage, Not OK. Voltage can be a little off, like the difference between 24v sla and 24v lifepo4.
Series connections, Idealy eveything the same except for voltage. Some have gotten away with different chemistries but whether it works or not depends on how similar the internal resistance of each one is. Different voltages are ok, like a 24v with a 36 v.
In general, paralell connecting is not too fussy, but series is. In paralell, a weaker battery will not put out 50% of the current. The stronger battery will pull more amps and the weak one less, till they both run out at nearly the same time. Some chemistries like you to stop sooner that others though, so for example, a nicad and an sla together is likely to hurt the sla more if the whole pack is discharged 100%. Or a lifepo4 could have a higher bms cutout, and leave the paralelled sla carrying it all at the end.
Perhaps you could charge one pack to just only 4500mah, you would lose 500mah, but at least they would be more similar… I think someone mentioned it would be great to stay between 200mah difference in capacity and not go higher.
I agree with no voltage difference. That should NOT be used. So li ion lipo setting would be like combining space cell 36v which is known for 10s4p li ion with a 9s1p lipo.
Seriously changing liion and lipo on the go wouldnt matter so much like a turbo. Doesntworklikethis.com
Provide each battery setup with higher amps discharge with more parallel connections would do the trick. Electric drive is not a combustion that can benefit from turbo. The turbo of electric drive is a big capacitor, as they can discharge so high in short amount of time. The bigger the caps, the more boost youll have. My caps can hold so much power that enable 2seconds rapid acceleration. So you can feel the power climbing short steep road incline or burst power during stating off.
Two components in parallel have the same voltage. So if the smaller battery pack is drained it will drop voltage which makes the larger battery pack provide current that will charge / re-balance the smaller pack.
(I’ve stated that I neglected the voltage drop created by the internal resistance which has an effect under load. When the load lessens the balancing will start)
You can’t overdrain one battery in parallel while the other still has juice. In series this is a different story.
One thing is clear : if you go with a setup as described by the OP you have to know what you are doing and you have to monitor your batteries.
It will work though similar to a capacitor. The lipo providing the amps for aggressive riding, the lions providing the range. I’m fully in line here with @laurnts.
From my looking into capacitors supplying burst power, caps do behave like that I general but I found a lot of evidence stating you can’t get more instant power from bigger caps. Scorpion sells their power bank capa but I hear it’s marketing bunk
With the parallel connection of dissimilar cells it seems u want the weaker cells to have more juice to compensate, I think maybe the same helps for series. The weakest series linked cell I guess takes the most beating? How does that work?
Not entirely sure. What comes clear now is that it is hard to speculate about any of this any further and some tests should be done. Will see how it goes when I get the chance to test and assemble my liion pack. Though, it might still take a month till I get any sort of results.
I think we could replicate this on smaller scale and see what happens.
Do you know anyone who already used different capacity batteries in parallel? I do believe this should work ok, when there are little difference between capacities… but what about when one pack has 5000mah and the second one 8000mah? Will one battery be able to charge the second one while also supplying current for the load? If someone can explain this process and how it should work, then that would shed some clarity on this, whenever it would balance/charge the second battery or not.
Post author: magellan
_ / 09-07-2015, 11:49 PM #6 / _
Knowledge Credit goes to: ‘‘mattheww50 and mdocod’’ from candlelightforums.com
Default Re: Mismatching chemistries question
If the cells are in series in that flashlight, then you need some sort of protection to prevent any one or more cells from being completely discharged, or reverse charged by the others.
This applies whether or not the cells are the same or not, but obviously, with a cell mismatch, global cell-by-cell protection (or every cell in the bunch being individually protected) is even more important than it otherwise would be. As long as all the cells in series are rated to handle the current that the flashlight draws, and each cell has its own protection (provided either by a global or built in device) to kill the circuit when discharged, then mismatched cells wouldn’t be too big a deal. I don’t really know what Mathew is talking about above there, there is no “forcing” issue here. The resistance of all of the components in the circuit adds up, including all of the cells in series, and the final flow of electrons from the battery is voltage/resistance=current. The current flowing will be the same across all the cells in series. The cells with higher resistance will drop more voltage then those with lower resistance, but this is only a problem if the rate at which energy is dissipated in the cell causes a thermal problem, which, shouldn’t be a concern if all of the cells in the pack are indeed a good match to the load in the first place.
That’s all pretty academic at this point. If your flashlight employs 18650s in series, then global-cell-by-cell protection or individual cell protection is highly recommended regardless of whether the cells are the same or not, and the mismatch just creates a much greater demand for individual cell protection. I would not want a flashlight that runs a bunch of any lithium chemistry cells in series without over-discharge protection on every cell.
If the cells in the light are all in parallel, you may have more wiggle room to get by with mixed cells without concern, but I would still have some concerns…
The current flowing from each cell in parallel, will NOT be fixed like it is when they are in series. Cells in parallel with different internal resistance and capacity, will always remain nearly “balanced” in terms of voltage, but that doesn’t always mean they will be balanced in terms of state-of-charge, and doesn’t always mean they will deliver the same current under a load. Though most of the discharge, the minor variations in current flow from mismatched cells in parallel would likely be a non-issue between “similar” cells. The problem is towards the end of the discharge, especially if the flashlight has boost regulation to maintain constant brightness. If it does, then the highest current load, could wind up coinciding with a condition where some of the cells in the pack are effectively drained, while others still have lots of energy left in them, and those cells which had higher capacity, could wind up forking over 2-3X the current expected towards the end of a discharge as a result of this mis-match in discharge characteristics between cells…
Lets assume these 2 cells were in parallel, discharging together at ~6A total, or ~3A per cell. Things would be fine from ~4V down to ~3.3V under a load, but below that, what happens? Suddenly, the NCR18650B has all this capacity still available below 3.3V, while the other cell is “dead,” thus, the NCR18650B winds up delivering all 6 amps for the remainder of the discharge.
Now, if the cells being “paired” in parallel (or in your case, “quaded”), have similar discharge profiles, and plenty of headroom between the load and max current rating for the cells, then mixing cells in parallel is far less of a concern, even if they are a bit different. If the flashlight in question has a boost regulator, then mismatched cells are going to see more “abuse” towards the end of a discharge than matched cells. Something to keep in mind.
-Eric
Default Re: Mismatching chemistries question
Quote Originally Posted by RetroTechie View Post
Oooohhh… then there is no problem, right?
With mixed cells in parallel, there is no chance of cells driving one another into polarity reversal, all cells produce the same voltage, and higher impedance cells just deliver a smaller % than their “fair share” of total current. And with the smaller current delivered, self-heating drops too. The cells that take the bigger share of total current, are precisely those cells with the lowest impedance (lowest self-heating @ same current).
The cell that will wind up delivering the highest average current through a discharge in a bank of mismatched parallel cells will be the cell with the highest capacity, not necessarily the cell with the lowest internal resistance. Granted, at certain times during a discharge, a lower resistance lower capacity cell may run slightly higher current, but on the overall, the highest capacity cell has to discharge at the highest rate…
To help understand this. Imagine placing an IMR18500-1100mAH cell with 60mOhm resistance in parallel with an NCR18650B-3400mAH cell with 110mOhm resistance and discharging the pair.
Think about how that discharge would go, considering that they are effectively “locked” to hold similar states-of-charge through the discharge. Which cell would deliver more current?
After pondering that a bit, it should be pretty obvious that the 3400mAH cell will wind up averaging ~3X higher current than the 1100mAH cell when discharged in parallel. Capacity trumps resistance in this situation, and even when the difference in cells is narrower, the same will remain true.
…any cell rated for 10A continuous could do that on its own, and stay within spec. So even the worst case is okay, adding more cells just increases runtime, and reduces stress on each cell. And even increases efficiency (slightly) because less of the cells’ energy is wasted in internal impedance.
Yes, if the light draws UP TO 9A, and the cells are all designed to run up to 10A, then it should be a non-issue.
I think the best bet would be to go with bms / pcm to monitor / shut off, if any of the parallel packs starts to act weird…
Found this comprehensive resource of different types of bms / pcm’s:
–
Resource for gaining more insight into batteries (unbalanced including):
Will try not overhelm the topic but been reading a lot lately concerning battery making and this is what I found related to mixing batteries.
It partly gives some clues about that resistance thing and suggests using different gauge wires.
He used lead acid / lifepo system, so the requirements for lipo/li ion might be a little bit different.
‘‘Mixing chemistries for most economic performance and range’’Postby Aerowhatt » Sat Mar 07, 2009 4:49 am
Just a note about resistance and paralleling two non matched batteries. If they are both conected to the load then having resistance in between one of them and the load isn’t as bad as it seems. Intuatively adding resistance produces alot of heat and wastes power. This is very true in a single source system. But with dual sources the system just askes more of the second source of power rather than forcing the power through the resistance causing a large loss to heat.
Similar to the difference between adding resistors in series and adding them in parallel. The results are very different! For example on my motorcycle which has some pretty large peak amp demands. A direct paralleling of the LiFePO4 with the SLA’s using the same guage and length wires would yield higher than desired peak loads on the LiFePO4. A simple solution is to use smaller gauge cable for the LiFePO4. So that at the controller (or junction of the two batteries) it’s voltage drop is more than it otherwise would be. Intutition jumps to the conclusion that the wires would heat significantly and waste power. Why, because if you add resistance and have to pull the same amount of power you get a big voltage drop and an increase in amps resulting in hot cables. What intuition forgets is that there is another source of power (similar to parallel connected resistors)! So rather than FORCE the higher resistance line it pulls from the other one. Basically it doesn’t take much additional voltage drop on the LiFePO4 to get the SLA’s to step up a bit more and give a more ideal balance of current delivery during peaks loads. So if you have a dual battery system that is naturally pretty close to each performing where you want them too. Changing the harness resistance of either, or both batteries to fine tune it results in a very low loss.
In the case of the motorcycle going from a 6ga cable to 10ga cable on the LiFPO4 fine tuned the balance of load for each battery with only a 3 degree F increase in the harness for the LiFePO4. Arguably this is less loss than the losses associated in throttling that battery with electronics would be. With the added advantages of no additional cost or complexity.
I just thought of this very idea, how ever, instead of putting the two batteries together in parallel, what if you had the two batteries working together, with the addition logical piece of hardware that dictates when each battery should be used. e.g. When cruising the power source would be li-ions and when you need a burst of power it switches to the lipo??
I think someone has suggested this idea also, the guy who mentioned it called it booster switch or similar.
I think it could be tried with some high amp switch or a smart switch where you hook it up to a receiver or install a foot switch somewhere on your board.
This could perhaps work better than running with mixed lipo and liion all the time.
Although, we are still improving the database so we have not had the chance to massively start inviting people to join and post & share their eboard data there.
We plan to start inviting more people there, once we have finalized the design and some extra features for the page.
LIPO VS LI-ION
I don’t know much about the subject as I am not a chemical engineer and the stuff being explained to me went way over my head when I took a 4 hour crash course from an actual battery chemistry engineer with a PhD and several years of experience but here is the main takeaway.
You can produce a high power/low energy cell (VTC6) or low power/high energy cell (Sanyo GA) out of the exact same chemistry by varying the ratios of the chemicals inside the cells ever so slightly (we’re talking 1-2% changes).
Now, these are kind of sheets coated in chemicals and stuff. You can either roll them up as tight as possible and stick them in a tube, 18650 style, the advantage of this is that you get more contact surface between sheets and therefore you can have a more energy dense cell or you can lay them flat in layers in order to have slim packs that can be shaped in a thousand ways.
The bad thing about rolling the sheets into a tube is that you do not get as much heat transfer to the outside so you’re limited as to how much power you can run through them before the chemicals start degrading due to heat and reacting too fast.
The sheets on the other hand have the same reaction occur over a larger volume so the heat is dissipated better and the chemicals remain happier which is why they can produce more power.
That is abouf as much as I remember without going into excruciating details.
All of this of course has an effect on the voltage operating range, internal resistance and therefore discharge curves. So based not only on the chemistry but also packaging, your internal resistance can vary throughout the discharge cycle. The internal resistance is a direct product of the chemical reaction.
Internal resistance is NOT a constant. Not within a single discharge, nor at the same point of the discharge on different cycles. It is always rising as the chemicals brake down.
The problem with mixing Lipos and Liions in parallel then becomes who is providing more power, why, and is it exceeding the safety limits at ALL TIMES.
The internal resistance of LiPos is lower than Liions which is good, it means that they’ll carry the larger current for the same voltage drop. The problem arises when LiPos reach ~3.4V and their internal resistance is almost the same if not higher than that of a liion at the same voltage which means the power distribution is equal or slightly higher on the liions. Are you harming the liions at this point?
What about in 10 cycles, do you know what the internal resistance of each cell is then? How about in 50 cycles half way through the discharge curve?
If your LiPos deteriorate faster than your Liions then they would have a higher internal resistance and you could be harming your Liions.
It’s a very slippery slope and unless you know exactly what you’re doing, I’d recommend against it.
I do however think that you could run on LiPos during high power conditions and then switch over to Liiions on cruise conditions. My company is working towards this in the future with the exception that we would use a fuel cell in combination with a gas engine instead of liions and lipos. But the end goal is similar.
Your controller would be connected to both power sources and based on the power draw, you could choose power source A or B.
I could write more on the subject but have a flight test scheduled in a few hours so I’m calling it a night.
Thanks for posting this! I think we could finally make a ‘‘summary’’ with all the info you have already told on this forum about batteries!
Had one question though, you mentioned this:
So do you really agree here that lipo would detoriate faster than li-ions? I understand that in this ‘‘application’’, where both of them might be working at the same time, the lipos would be the ones which would take the most ‘‘load’’, so it would be logical to assume they detoriate faster than the li-ions only because of the size of the load…
But anyways - I still have not found a source for a good lipo vs li-ion comparison as im not sure how both of these ‘‘chemistry compositions’’ work in the long run, if they are discharged / charged in the same way…
You told it very good about the ‘‘layout of layers’’ in a battery - for lipo vs li-ion. The physical structure of the battery itself gives it its attributes when it is operating… that is why we could sort of say that lipos ‘‘dont heat as much as li-ions’’ or that their heat dispersation is just way better because of the somewhat bigger volume.
Anyways… some good stuff you got there… I wished more ppl would get the chance to speak with battery chemical engineers like you did
–
I think you should agree that there are ppl runing mixed chemistry setups… they just monitor their cells better and for newbie who does not know the difference between 3.4v and 4.2v, of course, you wont let him run both of them in parallel.
I get that it might be damaging the cells when the load exceeds the limits the batteries are rated for - but that is true no matter if they are mixed or not. Shouldn’t a parallel setup like this always be better than the single packs?
I mean when you have a load of say 100A and you only use the Liion pack it will not be healthy for it.
The same is true for the Lipo. Combining them in parallel however should make it at least healthier for them to supply the 100A (the load obviously doesn’t double just because you use two packs in parallel).
Isn’t it also like having a charger with you the entire time? The Lipo will supply the high currents and in times of little load the Liion will actually “charge” the Lipo back to an even voltage across the packs.
Could it maybe even make sense to hook the single liions to the balance wire of the Lipo? Say having multiple 1S4P Liion packs hooked up to the balance leads charging the Lipo while riding and supporting the cells in case of voltage sag.
I also agree that the voltage cut off needs to fit to the Lipo - so you lose some capacity in the liions but that should still be less than what you gain by adding the Lipos to the system, no?
