Pack / Cell compression Optimized By Using Springs.

[Johncfii]

I appreciate the solution is not perfect but using springs is a far from perfect solution in the first place so as long as I am kind of in the ball park Im happy. The biggest reason I am using springs is so I dont severely over compress.

The pressure is not dependent on state of charge. It is a constant pressure regardless of SOC. It is not a pressure based on SOC.

Using any spring-based solution to maintain pressure, and with springs of any reasonable length, pressure most definitely will change with SOC. Only something like a pressure-compensating hydraulic fixture would allow constant pressure over a range of expansion/contraction.

Of course there is always room for a “good-enough” solution. But assembling each pack with a dedicated compression fixture is such small added cost for the assurance of better performance that .... just my opinion.

 

[fafrd]

https://www.daytonlamina.com/sites/default/files/doc/DS117_DieMax_XL_Springs-web.pdf

Page 11

Thanks. That’s what I was looking for. The total deflection to solid is slighter larger than the 30% ‘maximum’ travel for acceptable lifetime.

With the Lee Springs, you could select a spring with Solid Load below 17psi and then ‘back off’ the appropriate distance (in terms of turns on the bolt/threaded rod) to calibrate 100% SOC load.

With these springs, sounds like it’s not a good idea to compress to solid, so compressing an appropriate distance from free probably makes more sense...

 

[noenegdod]

I would like to keep this type discussion on the other threads .... and keep this one more about the use of springs, not the theory of the pressure specification ..

If you go back and read what I was replying to instead of taking the statement in isolation you would see it is relevant.

but, there were also graphs showing how the cycle life varied across of spectrum of pressure, and there were direct conversations with EVE engineers.

You are correct, and I did read all of them. Those tests would have been based on cycles at a consistent pressure, not on varying the pressure during the cycle. They set a pressure and cycle the cells through a given number of cycles to get a curve. Increase the pressure and cycle through the same number of cycles. Rinse and repeat to get the graph that EVE provided. They did not vary the pressure based on SOC

 

[Bob B]

You are correct, and I did read all of them. Those tests would have been based on cycles at a consistent pressure, not on varying the pressure during the cycle. They set a pressure and cycle the cells through a given number of cycles to get a curve. Increase the pressure and cycle through the same number of cycles. Rinse and repeat to get the graph that EVE provided. They did not vary the pressure based on SOC

Since there is no way we are going to do a DIY constant pressure mechanism ... do you agree that keeping the pack in a range between 9 and 15 PSI is logically going to increase cycle life of the pack?

 

[noenegdod]

Using any spring-based solution to maintain pressure, and with springs of any reasonable length, pressure most definitely will change with SOC. Only something like a pressure-compensating hydraulic fixture would allow constant pressure over a range of expansion/contraction.

Did you not read what I said?! Slow down and actually read. That is essentially exactly what I said in the first place.

Of course there is always room for a “good-enough” solution. But assembling each pack with a dedicated compression fixture is such small added cost for the assurance of better performance that .... just my opinion.

What everyone is doing here is what they deem as "good-enough" . You dont think that my approach is close enough to your ideal, I dont think that selecting springs that loose or gain over 200lbs for every 1/5", falls into anything higher than the "good-enough" pile.

We are all just trying to do the best we are capable of with the resources we have and having someone make suggestions is great but the attempt push others efforts down to "good-enough" and stating their solution is superior when they have no relevant expertise is pretty arrogant.

 

[noenegdod]

Since there is no way we are going to do a DIY constant pressure mechanism ... do you agree that keeping the pack in a range between 9 and 15 PSI is logically going to increase cycle life of the pack?

Bob, I am using springs....

Where did I ever imply, explicitly or implicitly that I did not feel this was a good idea?

 

[Bob B]

Well .... I guess it was when you posted this ....

The pressure is not dependent on state of charge. It is a constant pressure regardless of SOC. It is not a pressure based on SOC. There are a bunch of theories floating around on where the target pressure should be set but they are all speculation and best guesses.

But after reading the rest of that paragraph, I see you are using springs ..... I also agree with the idea of trying to use ones that have less deviation.

Sorry if I have helped side track the discussion.

 

[noenegdod]

Well .... I guess it was when you posted this ....

The pressure is not dependent on state of charge. It is a constant pressure regardless of SOC. It is not a pressure based on SOC. There are a bunch of theories floating around on where the target pressure should be set but they are all speculation and best guesses.

But after reading the rest of that paragraph, I see you are using springs ..... I also agree with the idea of trying to use ones that have less deviation.

Sorry if I have helped side track the discussion.

I could have been clearer. I should have said:

The tests we are using to validate our attempts to compress cells were all done at a constant pressure regardless of SOC. The pressure the cells were exposed to did not vary based on SOC through the life of the test.

When using springs, where individuals are placing the ideal 12psi based on SOC is speculation, theory and best guesses.

I will try harder.

 

[Johncfii]

When using springs, where individuals are placing the ideal 12psi based on SOC is speculation, theory and best guesses.

That statement is way too dismissive of the skill and intelligence of many dedicated membets of this forum. Good empirical work has been done that makes it much more than guess and speculation. We might not have access to expensive strain gauges and pressure sensors, but this is a community that includes many skilled builders who make good use of shared data and information.

 

[cinergi]

Zen moment here ....

You guys aren't understanding @noenegdod . The intent is not to dismiss the value of spring-based compression or any of the testing and research we've done. @noenegdod is saying these two factually-correct things:

1) The published increased lifecycle is based on a constant 300kgf fixture. Our spring-based fixtures do not provide constant 300kgf.
2) Choosing what SoC at which to set our spring-based fixtures is only based on "centering" the minimum and maximum load -- it is not based on any data that indicates at which SoC our spring-based fixtures should be calibrated.

That said, @noenegdod is a proponent of the spring-based fixtures and acknowledges that EVE (and others) have stated that a "good PSI range" is 6-17 (or thereabouts, I don't exactly remember).

 

[Bob B]

Thanks @cinergi ..... I'm sure you are correct. I just don't like these kind of distraction ..... that seem to have no real value.

 

[noenegdod]

Zen moment here ....

You guys aren't understanding @noenegdod . The intent is not to dismiss the value of spring-based compression or any of the testing and research we've done. @noenegdod is saying these two factually-correct things:

1) The published increased lifecycle is based on a constant 300kgf fixture. Our spring-based fixtures do not provide constant 300kgf.
2) Choosing what SoC at which to set our spring-based fixtures is only based on "centering" the minimum and maximum load -- it is not based on any data that indicates at which SoC our spring-based fixtures should be calibrated.

That said, @noenegdod is a proponent of the spring-based fixtures and acknowledges that EVE (and others) have stated that a "good PSI range" is 6-17 (or thereabouts, I don't exactly remember).

Thank you Sir. Either I need to do a better job of expressing myself or the issue is on the other end but obviously there was a break down in communication.

 

[Bob B]

CAN WE PLEASE GET BACK TO THE SUBJECT OF THIS THREAD.

 

[noenegdod]

CAN WE PLEASE GET BACK TO THE SUBJECT OF THIS THREAD.

Lets do it!

Space is an issue for a lot of us, if space was not a consideration, how long do you think you would go with the spring.

This example: https://www.leespring.com/compression-springs-hefty?search=LHL750B15

is an 8" spring with a 60lbs/ft rate. You would compress it about 2.5" to get the 162 lbs for a 4 rod clamp and have a pretty stable pressure from 0-100%.

Bob B, do you think that is worth the effort or is that too much for not very much return?

 

[Bob B]

That would kinda be the "Cadillac" of a spring installation, but it would require at least 9 and probably 10" of threaded rod outside the compression plate ..... I'd probably go with more of a "Chevy" installation using a shorter spring.

 

[noenegdod]

I was thinking you might be able to get away with 8 or even 7" of rod outside the case. The spring could be flush with the end of the rod and get the first quarter inch of compression by hand or if you were really strong, the first inch. but if space was not an issue leaving 10 outside the box wouldnt be a problem.

 

[fafrd]

I had considered that as a concern but decided it was not a significant issue because:

• They are supposed to be under the load so assessing their physical dimensions not under load is not something that can really be evaluated.
• When stacking cells together, if the areas that have bulged are in different locations (one in the top half and the next one in the bottom half vs two cells where the swelling is both in the middle) the overall dimension is meaningless.

I am adding 3mm drawer liner to the sides of each and every cell to help mitigate. I appreciate the solution is not perfect but using springs is a far from perfect solution in the first place so as long as I am kind of in the ball park Im happy. The biggest reason I am using springs is so I dont severely over compress.

The pressure is not dependent on state of charge. It is a constant pressure regardless of SOC. It is not a pressure based on SOC. There are a bunch of theories floating around on where the target pressure should be set but they are all speculation and best guesses. The springs I chose have a large difference between free and solid dimension and a relatively low spring rate so as the cells expand and contract based on their state of charge there will be relatively little change in force on the cells so Im setting mine at 50% soc. They are 4 inch spring but I have room so I get that luxury.

I really like these springs you’ve found and will probably get my springs from that same supplier.

One think I like about that supplier is that they provide lifetime based on compression % and more or less force you to retain headroom (since they don’t recommend going above a % compression which is below solid).

Those springs allow you to calibrate lbs of force through rotations on a thread (as will pretty much any specified compression spring). The specs may be off and it might be nice to ‘trust but verify’ with a force sensor, but you could say that about almost anything you purchase, so going by specified spring rate is the pragmatic way to pr

I’ve been thinking more about how I will calibrate my fixture once it is ready and here are my thoughts:

I’m targeting 4mm of travel between 0% SOC and 100% SOC for my 1-row 8S battery but that is only after several cycles and initial travel will be as much as 8mm or even possibly as much as 12mm (which is what I get at 100% when not using any fixture).

I think I’ve found a spring solution that will provide ~2psi (110 lbs total) over 4mm and can reach a maximum compression of over 14psi (with headroom), so my goal is to ‘settle’ at ~13.1psi @ 100% SOC at which point 0% should end up close to 11.1psi @ 0% SOC.

And here is my plan to calibrate my fixture to settle at that level:

1/ Attach fixture to cells at unknown SOC (50% +/-40%) and calibrate springs for max target pressure (13.1psi in my case).

2/ Discharge to 0%, estimate psi by loosening to Free, then tighten to 6psi (minimum recommended compression).

3/ Charge to 100%, estimate psi by loosening to Free (should be below 100% max target pressure), then tighten to max target pressure (recording that delta).

4/ Discharge to 0%, estimate psi by measuring spring length, and repeat step 3 until delta approaches 0 (should take 2-3 cycles).

At this point, your max pressure when fully-charged will be where you want it to be (13.1psi in my case) and your min pressure when fully discharged will be determined by the actual compression (anti-expansion) of your pack (11.1psi in my case assuming actual expansion is ~4mm or 0.5mm/cell).

 

[natec]

Its still in the paper phase, but here is my planned design with calcs for the spring selection and compression plate thickness:
github.com/natecostello/van_two_point_oh/blob/master/electrical/battery/Battery_box_design.md

 

[Hedges]

I was thinking you might be able to get away with 8 or even 7" of rod outside the case. The spring could be flush with the end of the rod and get the first quarter inch of compression by hand or if you were really strong, the first inch. but if space was not an issue leaving 10 outside the box wouldnt be a problem.

Compression spring on the end of a rod extending beyond end of battery stack occupies a particular volume. In a few installations, might be able to interleave with other objects (e.g. a vent pipe) so not wasted.

Different configuration of compression or tension spring, could occupy space to the side of battery pack, between end plates.

You could get a very long spring of low compression rate. Start at say 10" long and compress down to say 2" long (it will buckle). It would then vary +/-2mm with battery charge/discharge and have negligible change in force. Cut off excess 8" of threaded rod after compressed.

Alternatively, use 2" tension springs between end plates and stretch to 10"

But given the published curves showing battery cycle life improved varying amounts from 6 psi to 17 psi, the 2" (uncompressed) springs people use are reasonable.

 

[fafrd]

Lets do it!

Space is an issue for a lot of us, if space was not a consideration, how long do you think you would go with the spring.

This example: https://www.leespring.com/compression-springs-hefty?search=LHL750B15

is an 8" spring with a 60lbs/ft rate. You would compress it about 2.5" to get the 162 lbs for a 4 rod clamp and have a pretty stable pressure from 0-100%.

Bob B, do you think that is worth the effort or is that too much for not very much return?

The new springs (identified in one of these threads) allows you to have +/-1psi with 2 1-1/4” springs per rod, so if you can get to a range of ~11.1-13.1psi with a total of 2-1/2” of spring length, I struggle to see why you’d want to push for something more ideal in terms of constant 12.1psi pressure...

 

[noenegdod]

The new springs (identified in one of these threads) allows you to have +/-1psi with 2 1-1/4” springs per rod, so if you can get to a range of ~11.1-13.1psi with a total of 2-1/2” of spring length, I struggle to see why you’d want to push for something more ideal in terms of constant 12.1psi pressure...

which ones are those?

 

[Bob B]

Its still in the paper phase, but here is my planned design with calcs for the spring selection and compression plate thickness:
github.com/natecostello/van_two_point_oh/blob/master/electrical/battery/Battery_box_design.md

Really good write-up over there .... and like the options with explanations.

 

[fafrd]

I am struggling with 2 fixture design issues I would appreciate anyone’s perspective on (and especially those that actually have a working fixture already):

1/ single-rod vs. dual rod: I was originally thinking I would go with a single-rod design, meaning one pair of threaded rods compressing a horizontal 1-1-4” square tube on either end (crossing the middle of plastic of wooden ‘ends’) but I’m now worried that design may allow the ends to tip/angle so that expansion at the bottom of the cells may exceed expansion at the top of the cells. I’m not sure a bit of ‘accordion’ expansion is a problem (and have the flexible busbars to accommodate it) but the dual-rod design (meaning 2 square tubing braces per end) assures more uniform pressure and the only negative is the additional cost.

2/ single-spring vs. dual-spring: with a spring on only one end, the cell farthest from the spring is only going to move ~1/4mm while the cell closest to the spring will move ~3-1/2mm (for my 8S pack), while with a half-length spring on either-end of each rod, the two central cells will only move ~1/4mm while the outermost cells will only move ~1-3/4mm. Less movement is probably good, more symmetrical movement is probably good, but the main thing making me think I may want to put springs on either end of each tod is that that means the rods themselves don’t move and can be used as the primary support beams for the entire structure with the cells themselves just ‘floating’ in between the two pressurized endcaps. It means twice as many springs but if they cost 1/2 as much, it’s pretty much a no-brainer. If the total investment on springs is double, I’m not sure the minor benefits would be worth it.

 

[fafrd]

which ones are those?

The DaytonaLamina springs found by Johncfii: https://www.daytonlamina.com/sites/default/files/doc/DS117_DieMax_XL_Springs-web.pdf

I’m considering the 3/8” rod 1-1/4” long spring which delivers 181lbs at max compression of 37.5%. In a single-beam (2 rods total) dual-spring (4 springs total) design it can be calibrated to deliver 10.95 to 13.25psi (so a range of 2.3psi from 0% SOC to 100% SOC).

 

[Hedges]

Its still in the paper phase, but here is my planned design with calcs for the spring selection and compression plate thickness:
github.com/natecostello/van_two_point_oh/blob/master/electrical/battery/Battery_box_design.md

"Aluminum sheet will be used for the compression plates."

Any "sheet" will bow badly.
"beam calculators".
"3/16 aluminum ... Deflection = 0.0876 in", "1/2 plywood ... Deflection = 0.0455 in"
vs. "1 mm expansion between 0 and 100% SOC" which equals 0.04 in.

Instead of a "sheet", ought to be a structure such as "I" beam, or "V" shaped bends in the aluminum sheet.
(Consider how rigid corrugated roofing is against bending in one direction, vs. the other.)

If your application is ultralight aircraft then consider honeycomb:

McMaster-Carr

McMaster-Carr is the complete source for your plant with over 595,000 products. 98% of products ordered ship from stock and deliver same or next day.

www.mcmaster.com