Suggestions and or help building a 17Kwh battery pack.

I am about to embark on building a 16.8Kwh battery pack out of the following:

960 3.2V 5.5Ah Tenergy round cells
An SuPower Battery 16S 150A BMS from batterysupports.com. *
10mm 0.15" Pure Nickel strip from batterysupports.com. *

(* Assuming the shipment from China ever actually gets here.)

Using a DIY MOT spot welder.Putting them in a 16S60P configuration. I am planning on 2 sets of 3 rows of 20 cells 8 cells deep and then putting them on top of each other in a reverse direction to get to the 16S and so both terminals exit on the same end of the battery pack. I will end up with a pack 24" long X 25" wide X 7.5" tall battery pack. Those dimensions are the internal cell only measurements the foam, HDPE, and steel will add 1.5-2" per dimension for an overall battery pack size. Now I plan to have this in a foam lined HDPE box embedding in a steel box for protection and support. The terminals will exit the pack such that the battery leads the controller plus the controller leads to the motor will be significantly less than 1 meter to minimize voltage drop. All the cells are new and measuring 3.43 or 3.44V. I will be checking them all before assembly, and if I need to running them all through a discharge and charge cycle to test and match them all before assembly.

Should I replace the stips I plan to use? Will I need to replace the connections between cells with copper? How thick if I do? Can anyone offer suggestions for a first time battery pack builder? Your suggestions and help would be greatly appreciated, thank you.

JimFritzMI said:

I am about to embark on building a 16.8Kw battery pack

Click to expand...

KW, or KWh?

It does matter, as 16.8KW capable pack means it can output that much power (so many amps output while sagging to only so many volts). But a 16.8KWh capable pack means it has that much capacity (but the actual output wattage could be very different).

Too many people (and vendors) don't use the terminology correctly, and it leads to misunderstandings and sometimes problems or (costly) incorrect purchases by those that don't know there's a difference.

960 3.2V 5.5Ah Tenergy round cells

Click to expand...

Which cells? (just to be sure they can support the power level you *do* expect from the pack, without sagging too much in voltage)

Can anyone offer suggestions for a first time battery pack builder?

Click to expand...

Not directly, but I would recommend reading up on all the many 18650 (and other) battery build and design threads; it's an awful lot of reading, but you'll get a lot of information about what has worked and what has not for others (in addition to lots of speculation, unfortunately ).

My bad, I've made the correction. LOL It's late... I meant 16.8 Kwh. I don't intend to pull even 150 amps, more than a few seconds. I don't intend to pull, even at the completion of the project more than .35C, slightly above their suggested 0.2C discharge rate which is way below their continuous discharge rate of 2C. I also don't intend to charge them at a rate higher than 0.09C which is way below their suggested charge rate of 0.2C. These cells won't be pushed hard at all. I won't exceed a continuous maximum of 6Kw, except during acceleration and hill climbing.

I can supply a datasheet if you would like. But they are LFR/Tenergy LiFePO4 32650 5500mAh 3.2V Rechargeable Batteries.

I would add, I am considering using 0.0162 sheets of copper instead of nickel strips because of the cost of the 100 meters of nickel strip I would need. Which would give me a cross sectional area of around 261 square millimeters, more than 2.5 times as much as 4/0 gauge wire. Which should give me very little, if it is even measurable, voltage drop in the battery pack, let alone the minor drop with less than a meter of combined wire leads from the pack to the controller and onto the motor. This large cross sectional area of the sheets would allow me to use commonly available 0.032 aluminum sheet. Unless someone believes the aluminum sheet would be too difficult to spot weld. I do intend for the battery pack case to be able to be sealed, such that oxidation won't present and issue.

For those questioning my loads, the vehicle will have a GVWR of 1000 pounds, a Cd of 0.159, a CdA of .151 from the frontal area of .95 square meters, and a rolling resistance of 0.008.

I should also mention, I will be using pure tungsten electrodes on the spot welder if I do have to copper sheets, instead of nickel. Which is where I am leaning as the $300 for copper sheets versus the 100 meters of 0.15X10mm strips would represent a huge savings. Let along the $100 for the aluminum sheet.

I suggest you practice spot welding on all your material options before you committ to a full pack.

I realize the lack of experience spot welding battery cells, is my major hurdle at this point. I do have a lot of experience with spot welding of thin gauge steel sheet from an industrial point. I'm just wondering if there is something I may not be considering about building large battery packs that will make this practically unworkable. I appreciate your comment though, thank you.

One question I would like to ask, if anyone knows, is there a material difference between the casing material of a common 18650 and a LiFePo4 round cell battery? I'd much rather eat up common 18650 cells than my tenergy cells during the testing phase. I understand they are considerably physically smaller cells, but if I can weld the 18650s then I should be good. Physically the material is slightly thicker in the tenergy cells, but I'm wondering about material metallurgical makeup. I really don't want to get a mass spec done on the materials and yet I can't seem to find anything which talks about the alloys used in the making of these cells case.

Sometimes people practice on dead AA primary cells. I think the construction is similar enough to get a good approximation.

Why would you make a pack that large? Grab a pack from a wrecked chevy volt and be done.

If they are the eBay tenergy cells, some of them have sketchy quality. Like tons of 0v dead cells.

I think the 0.15mm nickel strips for the parallel connections will be fine. I don't think it will be a problem spot-welding the nickel onto the cells, and then drawing the current you plan on using.

For occasional 150A bursts, you definitely need copper on the series connections. There is some research and experimentation going on right now about resistance soldering. It looks like a useful method to connect copper series bars to nickel-strips, without any fear that heat would be able to damage one of the cells. There is no conclusive result yet, the work is on-going.

"Resistance Soldering Bus Strips to 18650 Cells"
https://endless-sphere.com/forums/viewtopic.php?f=14&t=88965

"Common pack design mistakes, how to avoid?"
https://endless-sphere.com/forums/viewtopic.php?f=14&t=84412

"Fuse-wire for individual cells, DIY Tesla style"
https://endless-sphere.com/forums/viewtopic.php?f=14&t=88039

"Sticky thread index for battery technology"
https://endless-sphere.com/forums/viewtopic.php?f=14&t=69683

Thank you for the suggestion of dead alkaline batteries. That will work nicely.

I am building the pack that large for a touring autocycle, think large heavy velomobile. The reason I don't buy a battery from a Chevy Volt is money. I have less than $1200 in the 960 cells shipped and a 150A BMS. I have tested all 960 of the cells now and they are all good and within .1 and .2 volts of each other.

Thank you for the links, I will be reading them all tonight. I appreciate the help.

The reason I am trying to use aluminum sheets is again a cost concern, as well as I have several 48"X144" .025" aluminum sheets on hand. 120 meters of nickel strip would cost a fortune, from the sources I have been able to find.

Thank you, spinningmagnets. The bus bar with fuse wire looks to be a brilliant and inexpensive solution. With a slight reconfiguration, a slightly narrower and taller battery pack, that should work great.

JimFritzMI said:

I am building the pack that large for a touring autocycle, think large heavy velomobile. The reason I don't buy a battery from a Chevy Volt is money. I have less than $1200 in the 960 cells shipped and a 150A BMS. I have tested all 960 of the cells now and they are all good and within .1 and .2 volts of each other.

Click to expand...

I understand you are trying to do things cheaply, but there is more to it than just $ per Kwhr. Long term reliability/cycle life is worth considering as well. Chevy volt packs are ~$1500 usually have a similar capacity and a known quality. The tenergy's I have some as well, and actually it was the full rivers so far that have been garbage the tenergy's have been ok, although I've not tested them very much yet.

Are you just testing the Tenergy cells for voltage, or have you capacity checked them yet? What about IR?

Also, regarding what kind of conductors to use for making this pack, you are hardly drawing any current per cell, so even very light gauge nickel will be plenty.

960 cells with 16s, is 16s60p. 150A would mean just 2.5A max current draw per cell. The 10mm 0.15mm pure nickel you mentioned will easily handle 10-15A continuous. 10A or so being a pretty ideal continuous current rate.

If I were you, I'd probably buy some 32650 specific train track type nickel that will do the parallel and series connections at the same time with no need to make a bunch of cuts. I'm not sure if these is available in pure nickel, although I've seen it in nickel plated steel which would be plenty fine for your application as long as you are not exposing the pack to a high humidity environment. It would make your welding job orders of magnitude easier than hand cutting a couple thousand nickel strips.

JimFritzMI said:

. I have tested all 960 of the cells now and they are all good and within .1 and .2 volts of each other.
.......The reason I am trying to use aluminum sheets is again a cost concern, ....d.

Click to expand...

Is that voltage range after charging/balancing ?... Because 0.2 v is a lot of difference between cells in a pack .
Have a go at spot welding that Al to a steel battery case....let us know how that works out .

That voltage range is out of the box. I have already test all for voltage, and I am 25% of the way through testing capacity, and so far I haven't found a single bad cell.

The cheapest price I've gotten for nickel strips for the pack is $650, and that would only handle the parallel connections. Is everyone really spending 1/3 of their pack expenses on interconnections?

With a voltage of only 16s 150A isn't going to push all the weight of a 16kwh battery very well at all, so the terrain needs to be flat.

+/- .2V is horrible. I send packs with that kind of balance to the scrap battery section of my battery stash. You really need to test the cells first. Do they have tabs already? If not, I'd strongly suggest welding them in 10 strings of 16s6p, so you can test them in smaller batches. You need to test them after welding, and if you killed a cell in the process, you don't want it dragging a full 60 cells down with it in the process.

The large voltage difference tells me you're dealing with cells of different ages, and therefore capacities. Individual cell testing is warranted, even if it's something as crude as paralleling all and charging them to full voltage (leaving them in parallel for a few days to ensure equal voltage). Then put them all in series and doing a discharge test of say a bit more than half the rated capacity. Afterward measure and mark the voltage of each. Sort them by that voltage and put like cells in 16s strings. I've done that kind of testing and used a series of incandescent lightbulbs as my discharge resistor at nearly 1000V. You'd probably want to break yours into 6 batches for testing as the voltage gets high and dangerous. The problem with batching is getting an equal discharge test for all. Assuming you can't do real individual testing, the best thing would be to sort them first based on voltage, and put like voltage cells together in strings for the relative capacity discharge test above.

I hate to be the bearer of bad news, but at 7 cents per wh retail, you're pretty much assured you have old stock cells that have nowhere near their rated capacity. I've put many man-weeks in work making packs with used cells with good results. Some packs I built in 2008 and 2009 are still in service and that's with cells I got used that were rated for only 500 cycles new. You couldn't pay me to make a pack with 7 cents/wh Tenergy cells. My best advice would be to sell the lot to someone else at a loss and get some better batteries.

Why do you think you need such a big pack anyway with such an efficient to move vehicle anyway? If it's for cross country travel, you'll end up spending the bulk of you time charging anyway, and opportunity charging 18kwh will have you overstaying your welcome. If you're thrifty and patient, great automotive grade batteries can be had for well under 20 cents/kwh, and your vehicle would be a lot happier and safer with a more appropriate pack size of 4-6kwh.

Of course, like most, you'll ignore the good advice and invest countless hours making your pack with what you have. When you finish the pack and it proves to have less than half the rated capacity, at least be man enough to report back so others can learn. For your sake, hopefully the only issue is with diminished capacity, because I'd consider those cells to be dangerous and wouldn't want to be in an enclosed vehicle with them. Be sure to install temperature sensors liberally throughout the pack that are tied to an alarm, and include live individual cell voltage monitoring (not just the BMS). Lastly, be sure you have a plan and practice it for an emergency stop and vehicle escape, because a pack of that size can turn into a nightmare in seconds.

Sorry for the negativity, but I felt a strong dose of reality was warranted.

JimFritzMI said:

That voltage range is out of the box. I have already test all for voltage, and I am 25% of the way through testing capacity, and so far I haven't found a single bad cell.

The cheapest price I've gotten for nickel strips for the pack is $650, and that would only handle the parallel connections. Is everyone really spending 1/3 of their pack expenses on interconnections?

Click to expand...

1KG of nickel is more than enough for your application... $650 for what? What are they quoting you $650 for?

99% all had the exact same voltage, 43 of the 960 had a voltage of .01V higher and 11 were .01V lower. I misspoke when I said .1 and .2. I meant to say .01 and .02. 0.02V being the difference between the highest and lowest. No they have no tabs they are bare cells. I am in the process of building 5 racks which can balance charge 6 cells each at a time and I just got 5 genuine SkyRC iMax B6 balance chargers so I can capacity test them all at least 3 times before proceeding. I got that pricing by buying 1000 of them at a time and cheap, slow shipping. They were all 3.31V to 3.33V, with the vast majority at 3.32V.

Maybe I am sourcing the nickel strips from the wrong source. I calculated a little over 24 feet of nickel for both sides of each 60 cell parallel group with 16 groups. Around 120 meters of nickel strips. Is my pricing way off? I've only found solid nickel strips at around $5/meter. Do you have suggestions for a more inexpensive source of nickel strips? I would love to hear about if you do, otherwise my present course of action is copper or aluminum bus bars with fuse wires to the individual cells. I.E. power wall setup as suggested by spinningmagnets.

The vehicle is for long distance travel, across the United States and back at least once and then Michigan down to Costa Rica and back more than a couple of times. My primary concern is finding charging sources throughout Central America on unsupported trips, which is why I felt it would be wise to allow myself at least 2 days travel between charging opportunities. I should have a pack range at 35MPH of around 500 miles, although the vehicle is capable of 70MPH which would give me a range of only 200 miles. Though travelling at the higher speeds will depend upon my ability to register and license the vehicle as a motorcycle, should I decide to do so. The top speed differences would require a gearing change to maintain efficiencies in the drivetrain, however the plan is not to do such unless an off road opportunity comes along. The vehicle has a GVWR of 1000, fully loaded vehicle including passengers, with 7KW on tap that should result in good acceleration and maintaining a speed of 35MPH only requires 1KW average with losses.

I have an Elcon TCCH-48-35 to charge the completed pack in the vehicle which will charge depending on the voltage available to me at 29-35 amps. You don't believe I will be able to charge the pack overnight? Eventually the plan is to install a Level 2 charging port. This plan depends on me being able to successfully find a source of the port hardware. Pull in, plug in, sleep, unplug, pull out.

I never ignore experienced advice, and always appreciate it, your concerns have registered and I will be doing a lot of testing before, during and after I build the pack. Although as a back up I do realize that I may end up buying a couple thousand 18650s to replace the pack, my original plan before these cells were offered to me for testing. However, I don't have much to lose other than time with these current cells, and at the very least a lot of experience to gain.

If you make it to Costa Rica, look me up. I can fix you up with some proper batteries.

You're going to need some shade for that kind of riding, so I'd suggest some solar power. Just don't go for those supposed flexible panels made with regular cells, as repeated flexing will turn them to dust. That doesn't mean you have to go for expensive low efficiency thin film stuff, nor does it mean putting heavy glass panels meant for rooftops. One of my early motor customers traveled the entire perimeter of Australia on solar power with about a 1kwh battery if I remember correctly, so to me it's crazy to carry so much battery on a low power vehicle.

BTW, doing 35 with a rig geared for 70 is a bad choice, especially when you get in the mountains, because it makes more heat in the motor and controller. Our systems are happiest and have greater drive efficiency at a properly geared WOT.

John, I will take that offer seriously. As for the gearing, I have a human drive of 8-208 gear inches. and I will be carrying 2 sets of gears for the final drive. One for a top speed of 35 and one for 70, if I find race tracks where I can use them. A visit would be great though. I will be moving there permanently in about a year and would love to make friends with a like minded people.

Cheers,
Jim

JimFritzMI said:

Maybe I am sourcing the nickel strips from the wrong source. I calculated a little over 24 feet of nickel for both sides of each 60 cell parallel group with 16 groups. Around 120 meters of nickel strips. Is my pricing way off? I've only found solid nickel strips at around $5/meter. Do you have suggestions for a more inexpensive source of nickel strips? I would love to hear about if you do, otherwise my present course of action is copper or aluminum bus bars with fuse wires to the individual cells. I.E. power wall setup as suggested by spinningmagnets.

Click to expand...

Well, your calcs are out by a factor of 2... You shouldnt need more than 65m even doing it your way,..but you do not need that strip across all cells, ..the fuse wire + bus bar (could still be strip) is a better plan for low amp draw.

.

The vehicle has a GVWR of 1000, fully loaded vehicle including passengers, with 7KW on tap that should result in good acceleration and maintaining a speed of 35MPH only requires 1KW average with losses.
.

Click to expand...

Is that 1000 lb ,..or kg ?
Either way, remember not many roads are flat and downhills go by fast....but uphills are many, and slow, and will need every bit of power you can muster ..even for 1000 lb.

Hillhater, would you define your definition of low amp draw? 50 amps, 100amps?

It is 1000 pounds. As I sit now, that represents 400 pounds for passengers and cargo, 250 pound battery, and a 300 pound autocycle w/o battery.

I realize I am grossly overpowered, to the point even that I am bordering on inefficient. However, I should be able to go between 200 and 500 miles depending on how I want to drive, and it still will only cost me less than $0.02/mile considering electricity costs and maintenance costs including battery replacement. A man can afford to do nothing but travelling at that rate. Maybe a micro tiny house trailer, exact same machine and it's still less than $0.03/mile.

I realize the grade will be my enemy, especially in curvy mountain roads, and I am massively partial to curvy mountain roads. Which is why I have a motor which will generate that power without overheating and holding top speed, and a battery pack which can supply the amps for that climb without passing 0.4C. Which could hardly be called even roughing up the pack. So with a range of 500 miles and a pack worth 2000 cycles. I can spend the next million miles deciding exactly at which angle I want the rear facing camera/mirrors. LOL

I should mention now as the testing is proceeding. I still haven't found a single bad cell. However, the time it is taking me to test these is reminding me that these are cells meant for power storage, not delivering high energy. They want to live at 0.2C all the time, they do tolerate high overcharge, and will handle at most 2.5C discharge. These cells are meant to be in large packs at the power we require from them or put into service in low draw applications. At the pack dimensions and weights regular e-bike users need, these cells would be nearly worthless, you would barely get 100 cycles pushing them the way LiPos can be used.

JimFritzMI said:

Eventually the plan is to install a Level 2 charging port. This plan depends on me being able to successfully find a source of the port hardware. Pull in, plug in, sleep, unplug, pull out.

Click to expand...

You can get one from here:

http://www.tucsonev.com/

You probably want to have both NEMA and J1772 though. You won't always find an Electric Car charging spot. Some people use a harness to have both available at all times. I converted mine into an adaptor box, because I don't really know how to isolate the side which is not used. (E.g if you use the J1772 socket, the NEMA plug becomes live.) I guess in theory, you could make a three way switch, so that only one end is connected to the Elcon at a time... Easier to make an adaptor for me.

As for charging overnight, you'll need more like 10 hours to go from flat. If you need a balance, that will take longer, but you should get most of a full charge.

By the way, Level 2 mode 3 can provide up to 6.6kw of power... If the vehicle is that big anyway, thought of putting 3 Elcons in parallel, and getting 30% of your power in an hour? Will be great for when you can find a electric charging port - buy a coffee, take your time drinking it, and you've got your power for the day. Have lunch, and you've almost got a full charge.

Low amp draw..... For the fuse wire cell connections is below 20 amps.
You suggest a 2.5 C max load , so those individual cell fuse wires wont see more than 13-15amps momentarily..hence you dont need the full strip connect direct to each cell.
you could run a single strip between two rows of cells with fuse wire "spurs" off to the individual cells. An easy way to reduce you strip usage by half again.
Each "group" of 60 parallel cells now only needs 2 mtrs of bus strip. So 16 of those means 32 mtrs of strip needed
..plus a few meters of 20 amp fuse wire !
.....and a lot of patience and work to assemble.! :wink:

Thank you, Sunder. Exactly what I needed, thank you my friend.

Hillhater, that is currently the plan. Copper strip, I'm planning to use what they call architectural grade copper sheet which is widely and inexpensively available. And I finally found a good source, with the math explained, on calculating the fuse wire size. It definitely seems inexpensive, although a lot of work, but time I have. At this point I have recalculated my cell layout and intend to assemble 3 48V packs. They can fit into a bike camper/trailer I am building and allow extremely long distance trips with my current 2 wheeled recumbent work ebike.

I wouldnt spend much time "calculating". fuse wire size.
If you know the max you are ever likely to draw..(15a?) just shoot for something rated to that +50% and go with it.
The fuse is only really needed to deal with serious dead short or internal cell shorting , which will be several times your planned max amps.