Charger for LFP cell testing

[zamboni]

I think you will find that battery balancing is not as big an issue as it is made out to be.

Inconceivable!

I've come to regard cell balancing as choosing the voltage at which you want to be able to rely on whole-pack voltage (rather than separate cell measurements) to make decisions. Even "top" and "bottom" are misnomers because i expect a well-informed user would dial their voltages away from the very ends of the knees anyway. It's more like "high" and "low" balancing... I'm shooting for "high" for all the reasons already established in other threads: Solar charged, low-C, many days of reserve capacity.

It is easier to start off with a battery that is balanced but if you have some sort of cell monitoring and alarm or even better a BMS that will shut down the charger or loads

...
Originally i thought i could skip the BMS. If i had bought better cells, i probably could (as you imply above), but watching individual cell voltages just with my meter has shown me that my cells are too unruly and i need cell-level monitoring. I caught one string shooting up very quickly once, and that got my attention. We leave the trailer unattended quite a lot when we go out for a day -- i don't want to come home to a plasma ball. So i just bought a cellog. With the whizbang jr. connected to the Classic and "high" balanced cells, i should be able to count on the classic keeping cell voltages under control by measuring pack voltage, and the JLD can throw the overcharge protection solenoid if pack voltage creeps up beyond that. But the cellog can watch cell voltages on the discharge end, and throw the LVD if any cell gets too drained.

I supposed i could use a second cellog to manage the overcharge protection too...then i wouldn't even need the JLD, or a BMS. But i'm trying to avoid throwing money/complexity at this system. The cellog is worthwhile because i otherwise have no realtime monitoring of cell voltages. And it's really inexpensive.

I agree that it is good to know what the capacity of the individual cells is before you start. If I had had the Reaktor before I commissioned my battery I would have done a capacity test. I would have probably done it on only a few cells as it would be allot of work and take some time to do the 32 cells that I have!

20 cells here...not as big a task as you, but not trivial either! I had to drill and tap all 40 terminals in order to build the pack in the first place and that was pretty tedious. I can be patient and deliberate if it suits me : ) I'm ever so grateful for the advent of podcasts; making tedium infinitely more bearable.

I would be very surprised if you find the variation between cells to be more that a couple of %.

I seem to be a suck-magnet most of the time, so i'm planning on bad news. But maybe not!

Here are the charge/discharge curves that I did for one of the cells from my friend's system...

Did you create these from logged data in the Reaktor?
I have the manufacturer's published charge/discharge curve and i'm curious to compare it with my cells' actual performance, if i have time.

I think you will find that the measured SOC using the current flow too/from the battery will be more than a convenience as it can be accurate to within a few % if you adjust the reading to take into account the ~0.5% coulomb/current inefficiency of LFP batteries. It is far harder to work out the SOC from the battery voltage.

I agree that with the correct setup, coulomb counting can be a very useful tool. But people always get bent out of shape on public forums if you give any hint that you think a $15 "fuel gauge" is useful. So i just err on the extreme side of things and pretend that i'm not going to pay much attention to SOC. But in reality, that's the number i look at the most. I only pay attention to other numbers if something seems wrong with SOC. The system's automated controls are built around voltages, but for a user-facing interface SOC% is much more intuitive.

I have set my SOC counter to read 0% when there is 10% of the rated capacity left in the battery and it resets to 100% every time the average cell voltage is 3.45V with a charge current of less than C/20

I plan something similar with my Trimetric, which is what's visible in the living space.

Sorry, it wasn't intended to be an advert.

Didn't take it that way. Grateful for the link! Upon disassembly i may decide i'm unhappy with my non-braided bus bars. My first choice was braided, but we were rushed when we did the initial conversion to LFP so i didn't source any.

[karrak]

I've come to regard cell balancing as choosing the voltage at which you want to be able to rely on whole-pack voltage (rather than separate cell measurements) to make decisions. Even "top" and "bottom" are misnomers because i expect a well-informed user would dial their voltages away from the very ends of the knees anyway. It's more like "high" and "low" balancing... I'm shooting for "high" for all the reasons already established in other threads: Solar charged, low-C, many days of reserve capacity.

I agree with you that you only need to have the battery balanced at the voltages that you will be charging to. Makes it a hell of allot easier than having to do a special balance procedure at voltages that you would not usually use. Only exception to this would be that I think it is worth doing the initial commissioning balance at around 3.6V as this will give an even tighter balance at the voltage that you intend to charge to on a regular basis.

I wonder how much of the other threads you have waded through! Like trying to wade through molasses.

Originally i thought i could skip the BMS. If i had bought better cells, i probably could (as you imply above), but watching individual cell voltages just with my meter has shown me that my cells are too unruly and i need cell-level monitoring. I caught one string shooting up very quickly once, and that got my attention. We leave the trailer unattended quite a lot when we go out for a day -- i don't want to come home to a plasma ball. So i just bought a cellog. With the whizbang jr. connected to the Classic and "high" balanced cells, i should be able to count on the classic keeping cell voltages under control by measuring pack voltage, and the JLD can throw the overcharge protection solenoid if pack voltage creeps up beyond that. But the cellog can watch cell voltages on the discharge end, and throw the LVD if any cell gets too drained.

I supposed i could use a second cellog to manage the overcharge protection too...then i wouldn't even need the JLD, or a BMS. But i'm trying to avoid throwing money/complexity at this system. The cellog is worthwhile because i otherwise have no realtime monitoring of cell voltages. And it's really inexpensive.

As you say you could use one cellog 8 to generate the LVD. I am fairly sure that there is a remote OFF input on the Magnum that this could be fed into. On the charger side another cellog 8 could be used to generate the HVD. Again I am fairly sure that this can be used to stop the Midnite. One problem with the Cellog 8s is that there is no hysteresis between the alarm being On or Off. In this case the cellog 8s are your BMS. Have you had a look at my BMS?

Did you create these from logged data in the Reaktor?

Yes that is how I generated the graphs. There are a few tricks to processing the data. If you decide to generate your own plots I can give you some tips.

I agree that with the correct setup, coulomb counting can be a very useful tool. But people always get bent out of shape on public forums if you give any hint that you think a $15 "fuel gauge" is useful. So i just err on the extreme side of things and pretend that i'm not going to pay much attention to SOC. But in reality, that's the number i look at the most. I only pay attention to other numbers if something seems wrong with SOC. The system's automated controls are built around voltages, but for a user-facing interface SOC% is much more intuitive.

Sound like a good strategy to me.

Didn't take it that way. Grateful for the link! Upon disassembly i may decide i'm unhappy with my non-braided bus bars. My first choice was braided, but we were rushed when we did the initial conversion to LFP so i didn't source any.

I would have a good look at the interconnects as your Inverter could draw over 300A from the battery. I ran some SPICE simulations on the wiring arrangement I outlined earlier and found that the resistance of the interconnects need to be less than 1/10 the cell resistance to get good current balance. Is that wiring arrangement the same as how you currently have your battery wired up?

Simon

Off grid 24V system, 6x190W Solar Panels, 32x90ah Winston LiFeYPO4 batteries installed April 2013
BMS - Homemade Battery logger github.com/simat/BatteryMonitor/wiki
Latronics 4kW Inverter, homemade MPPT controller

[zamboni]

I wonder how much of the other threads you have waded through! Like trying to wade through molasses.

I tried to find all the existing LFP posts using searches, but i'm sure i missed something. This forum is basically the only place anybody is talking about the "nuts and bolts" of LFP for renewable energy, which makes it a goldmine. But i have to say -- moreso than most other forums -- the content of the threads diverges wildly from the topic of the thread in most cases. It can be an ordeal as a newcomer ; ) It's a treasure hunt that pays off, though.

Again I am fairly sure that this can be used to stop the Midnite. One problem with the Cellog 8s is that there is no hysteresis between the alarm being On or Off. In this case the cellog 8s are your BMS.

You are correct about remotely stopping the Classic, however i will be adding the Whizbang Jr. Module (for accurate voltage measurements at the shunt instead of at the Bat+/- terminals of the charger). And the WBJ inputs to the Classic via Aux2 -- the same port used to drive the Classic to rest/float/charge externally. So i have to choose. The WBJ will make the shunt-measured voltage available to the Classic for all voltage-based decisions, so i definitely want that. Adding the cellog may be a little redundant, but i can parallel its outputs with the JLD to give both pack-level and cell-level HVD protection, and should not really get nuisance trips from the lack of hysteresis, since the HVD operating point will be several tenths of a volt above my charge voltage.

Also, it will let me keep an eye on cell voltages without having to remove the toilet and open the panel in the floor...

Have you had a look at my BMS?

Only briefly. DIY electronics are definitely on my project-fun list...but pretty close to the bottom. I'm a mechanical engineer with enough familiarity with electronics to wire a house or (apparently) set up a small, simple PV system...but a long way from being an electronics hobbyist. Also, i have been spending far too much time with projects and not enough time outside, which is the whole reason i moved into a damn RV in the first place ; ) So i'm trying to limit my "project fun-time" to just the stuff that keeps the rig rolling (am having some truck issues too, at the moment).

The new charger and cellog will give me plenty to play with during rainy days, anyway. If i can't figure out how to plot the logged Reaktor data myself, i'll hit you up for help.

I would have a good look at the interconnects as your Inverter could draw over 300A from the battery. I ran some SPICE simulations on the wiring arrangement I outlined earlier and found that the resistance of the interconnects need to be less than 1/10 the cell resistance to get good current balance. Is that wiring arrangement the same as how you currently have your battery wired up?

My cell interconnects follow the same regime as your diagram...except 5p4s. The connections are just 2-hole bars, from one cell to the next, and overlapping at each terminal. It's not elegant. I don't have the equipment to accurately measure the resistance of a single interconnect, but would not be surprised if they were a problem. A crude deltaV / deltaI calculation from the Magnum control panel (turned off everything in the trailer then turned on a 1500W inverter load), and came up with Ri ~ 4mOhm. This doesn't seem awful to me, but i'm still looking at nearly a full volt of sag when i fire up the water heater or electric teakettle, so i'm interested in improving it. The fundamental problem is that the whole system really ought to be 24V...but i won't be changing that anytime soon. Better interconnects may help. I would love to get multi-hole braided bus bars custom made for my pack, so i don't have overlapping bus bar at each terminal, but i'm not sure if now is the time. I'll look into the link you sent and see what they can make at what price.

[Mike90250]

just remember, when someone is vocal, verbose, and saying stuff thats easy to hear, does not make that person correct.

[zamboni]

And when someone is abrasive, verbose, and prone to 'personal attacks', doesn't mean they should be ignored.

Everything here is "something i read on the internet" and i treat it as such. My LFP bank is my experiment and my responsibility. Since there is no true authority on the subject to defer to, this is more like politics than engineering. Who to trust? Make small changes, measure the results, and make data-driven decisions.

If you this forum has a lot of misinformation mixed in...take a look at diesel truck forums! OMG.

[karrak]

My cell interconnects follow the same regime as your diagram...except 5p4s. The connections are just 2-hole bars, from one cell to the next, and overlapping at each terminal. It's not elegant. I don't have the equipment to accurately measure the resistance of a single interconnect, but would not be surprised if they were a problem. A crude deltaV / deltaI calculation from the Magnum control panel (turned off everything in the trailer then turned on a 1500W inverter load), and came up with Ri ~ 4mOhm. This doesn't seem awful to me, but i'm still looking at nearly a full volt of sag when i fire up the water heater or electric teakettle, so i'm interested in improving it. The fundamental problem is that the whole system really ought to be 24V...but i won't be changing that anytime soon. Better interconnects may help. I would love to get multi-hole braided bus bars custom made for my pack, so i don't have overlapping bus bar at each terminal, but i'm not sure if now is the time. I'll look into the link you sent and see what they can make at what price.

I shouldn't rely on my memory but I remember calculating the battery and interconnect resistance of my battery at around 6mOhms. Going from charging at around 10A to discharging at around 80A drops the overall battery voltage by about a volt.

It might be worth going round the system with a multimeter and measuring the voltage drops across the interconnects, lengths of cable, fuses and circuit breakers individually when there is a high discharge current to get an idea where the losses are occurring.

Simon

Off grid 24V system, 6x190W Solar Panels, 32x90ah Winston LiFeYPO4 batteries installed April 2013
BMS - Homemade Battery logger github.com/simat/BatteryMonitor/wiki
Latronics 4kW Inverter, homemade MPPT controller

[Sunking]

Another reason i was planning to do this cell-by-cell; i want to measure each cell's capacity starting from my operating "full" voltage (that its, voltage i will balance them all to, eg. 3.5vpc). I think i can only do that by individually discharging each cell.

That will work if you have no life and no one cares about you. I certainly don't care. You discharge than all in series until the first battery is drained. Note AH. Now discharge the last cells and count AH. Should not take more than 2 hours of your time and not days.

[Mod note - deleted spam]

You are right, the weakest cell determines the pack capacity. So how do you stop or correct that problem, Just one little mistake or a Vampire/Montitor board fails and you will destroy a cel. BMS systems are the number 1 cause of LFP failures. Us folks in the EV field learned this hard lesson long ago. We ripped out the BMS and sold them. When you Bottom Balance, you eliminate any chance of over discharge. All cells have the same capacity and thus all arrive at 2.5 volts at the same time making it impossible to drive a cell into reverse polarity.

Where did we learn about Bottom Balance? NASA, Boeing, Hughes Aircraft, Lockheed Martin, and Saft Battery Manufacture. Those companies use LFP and lithium batteries in mission critical systems, mainly satellites that orbit the earth. They do not Top Balance or even use a BMS in the sense you are thinking. They like all EV manufactures Middle Balance. Bottom Balance mimics Middle Balance where the cells are never allowed to be fully charged or discharged, they stay in a Partial State of Charge until the satellite falls out of orbit and burns up on reentry.

[Mod note - deleted spam]

Moderator Note - Forum spam consists of posts on Internet forums that contains related or unrelated advertisements, links to malicious websites, trolling and abusive or otherwise unwanted information.

[zamboni]

We ripped out the BMS and sold them

You don't have to convince me; i was about to pull the trigger on an Orion BMS the day i found this forum, but posts by you and others convinced me the Orion would have been money badly spent.

That will work if you have no life and no one cares about you.

Frankly, at the moment i don't...but that doesn't mean i enjoy wasting time. Thanks for the better suggestion.


Oh, and since this thread is supposed to be about cell chargers:
I scanned in the Turnigy Reaktor 300W manual, in case someone else is looking for it. The quality is low for filesize, so the text is pretty fuzzy. Sorry about that.

[Sunking]

Oh, and since this thread is supposed to be about cell chargers:
I scanned in the Turnigy Reaktor 300W manual, in case someone else is looking for it. The quality is low for filesize, so the text is pretty fuzzy. Sorry about that.

That is what you get for listening to Karrak. What you bought is Counterfeit Chi-Com Junk. It is stolen property from New Zealand.Exactly what Karrak sells and promotes here. He rips off his own country's companies. It is a Counterfeit Copy of an iCharger 1010B. Did they even give you the Balance Board shown with the iCharger in the pic below? Now if you want a good copy of the manual, try the real thing, not a fake.

When it comes down to it, the iCharger 1010B is a scaled down version of the PL8. the difference is the iCharger can charge @ 300 watts max, while the PL8 can charge @ 1330 watts or 4.4 times faster. The charger is great for smaller batteries used in RC planes, Quads, and Trucks. (does toys come to mind).For 12 volt batteries you are limited to about 200 AH and 100 AH @ 24 volts. PL8 up to 1000 AH at 12 volts and 500 AH @ 24 volts.

[sensij]

That is what you get for listening to Karrak. What you bought is Counterfeit Chi-Com Junk. It is stolen property from New Zealand.Exactly what Karrak sells and promotes here. He rips off his own country's companies. It is a Counterfeit Copy of an iCharger 1010B. Did they even give you the Balance Board shown with the iCharger in the pic below?

From the site you linked:

Please Note: The iCharger 1010+ standard package does not include a balance board. Compatible JST-XH, Polyquest, Hyperion or Thunderpower balance boards are available separately.

icharger.co.nz distributes products from www.jun-si.com. Sure, that sounds like an authentic New Zealand company to me, too. The iCharger and Reaktor are probably produced in the same factory, or at least were at one time, just like the equivalent models from Thunder, Voltz, NeuEnergy, etc.

Edit again: Here is a 2000+ post active thread in which the Reaktor is thoroughly dissected, compared to other chargers, etc.

[zamboni]

Page 11 of the manual (page 13 of the pdf i uploaded) gives the pinout for connecting LFP cells for balance charging directly to the charger (via its balance port). There is no mention of needing a balance board. Am i being ignorant here? Sounds like i just need to make the correct connections. I believe i read a customer review which mentioned not needing balance boards, too.

Of course, the manual is written in broken english (not the worst ever, but not the best)...so i'm not exactly going to take that one page as gospel. It would be nice if, for example, they gave some indication what type of connector i need, to build my balance-charging pigtail. I guess i'll just browse through a helicopter forum until i figure it out.

The iCharger was an attractive option but, like every other charger i saw except the Reaktor, it only operates on ~12VDC input. The built-in AC-->DC power supply of the Reaktor is important because space matters more to me than probably a lot of other folks. A separate power supply is yet another box to find a place for.

With electronics especially, i'm skeptical of all products - both imported and domestic. But automatically dismissing all Chinese imports as junk is kinda behind the times. As sensij mentions, i am aware that a lot of products roll off the exact same lines and just get boxed or branded differently. It's hard to have confidence in the country-of-origin of just about anything anymore. So i'm fine with imported goods but i do check reviews first (and do silly things like ask around on forums).

[karrak]

Page 11 of the manual (page 13 of the pdf i uploaded) gives the pinout for connecting LFP cells for balance charging directly to the charger (via its balance port). There is no mention of needing a balance board. Am i being ignorant here? Sounds like i just need to make the correct connections. I believe i read a customer review which mentioned not needing balance boards, too.

Of course, the manual is written in broken english (not the worst ever, but not the best)...so i'm not exactly going to take that one page as gospel. It would be nice if, for example, they gave some indication what type of connector i need, to build my balance-charging pigtail. I guess i'll just browse through a helicopter forum until i figure it out.

The "balance board" is not needed as it is built in to the Reaktor. The balance board just splits one balance port connector up to several different size connectors so you can plug prebuilt batteries that have a different size balance connector dependant on how many cells there are in the battery.

I think your best bet is to get a premade cable assembly something like this, cut off the male connector and wire the individual wires to the cells.

Simon

Off grid 24V system, 6x190W Solar Panels, 32x90ah Winston LiFeYPO4 batteries installed April 2013
BMS - Homemade Battery logger github.com/simat/BatteryMonitor/wiki
Latronics 4kW Inverter, homemade MPPT controller

[Sunking]

Page 11 of the manual (page 13 of the pdf i uploaded) gives the pinout for connecting LFP cells for balance charging directly to the charger (via its balance port). There is no mention of needing a balance board. Am i being ignorant here? Sounds like i just need to make the correct connections. I believe i read a customer review which mentioned not needing balance boards, too.

No it is not needed, just very handy to have. If you make a JST Balance Plug and cut-off one of the keys, you can make it work. Just one wrong, connection or mistake, and you let the SMOKE out of the charger.

[PNjunction]

zamboni - if you want to do the single-cell 3.2v LFP charger thing, see the category for that here. Looks like you'd want the 20 or 40A unit:

http://www.batteryspace.com/smartchargerforlifepo4batterypacks.aspx

3.65v is not a problem for individual cells. The usual 3.6v spec is because manufacturers know that you'll be using the cells as batteries and not as single units. If you dig the docs back to China, you'll commonly see 3.65v as being ok for single-cell charging.

Still, that's a lot of maintenance for so many of your cells, but at least you know that there are no adjustments to mistake. If you walk away, the only thing that will be done to that cell is fully charge it and stop.

I gotta hand it to you - thats a LOT of cells to maintain, but as long as you are safe and sane about it...

[zamboni]

Thanks for the link. Those chargers are a lot more capable than the little Reaktor i have now, but they seem to be pretty simple as far as additional features. I know it wil take a while but i am very interested in measuring at least some cell's capacities. Perhaps not all of them...just the weakest few...but i think i need the charger to have that feature built-in. The Reaktor manual says that it supports an external resistor for faster discharging (since it can only dissipate 20W internally). The wording is very wonky but it sounds like it supports up to 300W discharge rate in that config. I could pick up a couple power resistors and set that up, to reduce how long my capacity measurements take?

I remember looking at that site "AA Portable Power" when shopping for cells but wasn't too impressed with them when i called up to discuss the order. Do you have an opinion of them (for future reference)?

Yeah, it's a lot of cells. If this weren't an RV, i would have stuck with FLA...But it wasn't doable with the space/weight capacity we have. Still, it grants us a LOT of flexibility and that's worth spending serious time baby-sitting the cells in order to measure and re-balance the pack before re-commissioning the rig.