diy solar

diy solar

Pack / Cell compression Optimized By Using Springs.

My first post on the forum, so firstly a big thanks to everyone for the advice and knowledge shared. I have learned a lot, I think – what follows may prove/disprove that!

My build is at feasibility stage, but compression is an important consideration for me as space constraints require my 1P16S battery to be configured 1 x 16 cells in a single row. From the feedback on this forum, I therefore need to manage 8mm expansion across my battery pack through the charging cycle (0.5mm per cell) and maintain specified compression.

I am planning 4 bolts for compression, with springs at either end (so 8 in total), allowing each spring to manage 4mm of expansion. I would also plan to apply 300kgf at 30-40% SOC specified in the latest EVE 280K specification (attached - for which thanks, Amy Wan).

With c 150mm space each end of the pack to limit free length of each spring and prioritising as low change in load/mm as possible, 2 springs off the Lee Spring website really fit my requirements (LHL 1250A 10 & LHL 1000AB 12). Both springs have change in load of c 1.6kg/mm, so that a 4mm change per spring would increase individual spring load by c 6.5kg or 26kg across the pack. Compared to the 300kgf, this is a <10% change in compression force. This was confirmed in conversation with a technician from Lee Spring.

In psi terms, this equates to an increase from 12 to 13 psi if cells expand the full 8mm across the pack. Well below the critical level of 17 psi referenced elsewhere in this thread. They would really have to blow to get close to 17…

One other limiting factor highlighted by their technician: if the springs are deflected 80% or more from their resting state, this compromises their future performance. The 2 springs selected above under my assumptions only reach 70% deflection when 8mm cell expansion occurs, so some headroom there, too.

This spring configuration fits my purpose and space constraints, and appears to keep the cells very close to 12 psi on average - if my calcs are correct! This may be of use to other forum members.
 

Attachments

  • EVE LF280K (3 2V 280Ah) Product Specification( Version B ).pdf
    542.4 KB · Views: 49
I would also plan to apply 300kgf at 30-40% SOC specified in the latest EVE 280K specification (attached - for which thanks, Amy Wan).

I don't think the data sheet mentions compression, so relying on other sources.

The data sheet describes welding of terminals. Do you know what you'll be getting, and how you will make connections?
With 0.5mm per cell of motion, you'll want a busbar which can flex that much.
 
Compression per the datasheet attached above:
- "under 300kgf +/- 20kg" specified when applied to the thickness dimension of the cell at 30-40% SOC (section 3.13 on p2)
- "under the 300kgf clamp" specified for life cycles (sections 5.5/5.6 on pp3-4)

I'll be asking for welded terminals, and will most likely make crimped wire connections to eliminate the risk you highlights. Happy for any suggestions, though. I've seen some beautiful copper braided connections on a build by another forum member, but that might be above my pay grade...
 
Ah, under "cycle life". Document wasn't searchable.

Are the welds good these days? Some people have twisted welded studs off. We've see that for round terminals, intended assembly is a busbar with round hole to fit over the terminal. Both tapped holes and welded studs were after-market hacks. There were some pictures of properly done ones.

One guy built his own braid terminals with copper pipe. I wonder about maintaining crimp force in that case.
Ring terminals can be crimped on braid. Someone did that with a braid ground strap at work.
One member measured better voltage drop with solid busbars. It is possible to have a busbar bent in a shape so it is compliant for space between studs (but maybe not misalignment.)

My system is 8x 6V AGM batteries, not lithium cells. I bought 10x 4/0 cables with crimped ring terminals for $80.

Whatever kind you use, take care of native oxide on aluminum.
 
Ah, under "cycle life". Document wasn't searchable.

Are the welds good these days? Some people have twisted welded studs off. We've see that for round terminals, intended assembly is a busbar with round hole to fit over the terminal. Both tapped holes and welded studs were after-market hacks. There were some pictures of properly done ones.

One guy built his own braid terminals with copper pipe. I wonder about maintaining crimp force in that case.
Ring terminals can be crimped on braid. Someone did that with a braid ground strap at work.
One member measured better voltage drop with solid busbars. It is possible to have a busbar bent in a shape so it is compliant for space between studs (but maybe not misalignment.)

My system is 8x 6V AGM batteries, not lithium cells. I bought 10x 4/0 cables with crimped ring terminals for $80.

Whatever kind you use, take care of native oxide on aluminum.

I either save things as .pdf .... or print to .pdf .. One of my favorite things about the PDF docs it the search function.
 
Here’s a picture of my fixture. I used the same springs as 100 Proof (post 272).

3/4” plywood on ends (had some spare maple plywood)
metal on both sides of springs is 3/16” x 1 1/2” steel (local metal shop)
1/4” all thread (10 ft piece from Home Depot)
1/4” id plastic tubing (Home Depot)

As 100 Proof, I used a 162 lb. weight to compress the spring and measure it’s length.5DF67CC8-DAA8-469B-9256-3F582918A7F3.jpeg
Springs we’re compressed to 1.660”
 
Here’s a picture of my fixture. I used the same springs as 100 Proof (post 272).

3/4” plywood on ends (had some spare maple plywood)
metal on both sides of springs is 3/16” x 1 1/2” steel (local metal shop)
1/4” all thread (10 ft piece from Home Depot)
1/4” id plastic tubing (Home Depot)

As 100 Proof, I used a 162 lb. weight to compress the spring and measure it’s length.View attachment 76012
Springs we’re compressed to 1.660”
Looks great. Where are you mounting the bms?
 
1640651582405.png
I have read through this thread (and a few others on the "clamping" issue) and this way of distributing the load on the cell faces has me the most puzzled. It was mentioned by noenegdod but there doesn't seem to have been any discussion about it.
My understanding is that if you have one EVE LF280K cell with a surface area to be compressed at 12psi that would be :
173.7mm X 204.6mm = 35,539.02 or 55 square inches......X 12psi = 660 pounds of force on that surface.
If you were to use 4 bolts for the clamping then you would divide 660 pounds by 4 and apply 165 pounds at each bolt, therefore applying a total of 660 pounds on that cell face. I think everyone agrees on this.
If you were to use 2 bolts for the clamping then you would divide 660 pounds by 2 and apply 330 pounds at each bolt, therefore applying a total of 660 pounds on that cell face.
If you were to use 6 bolts for the clamping then you would divide 660 pounds by 6 and apply 110 pounds at each bolt, therefore applying a total of 660 pounds on that cell face.

If you had two cells side by side and used one plate to cover both cell walls and 4 bolts to do the clamping you would need (660 X 2) or 1320 pounds of clamping force or; 1320 divided by 4 equals 330 pounds at each bolt. This is assuming there is no deflection in the plate.
If you assume there might be deflection in the plate you might consider putting two more bolts at the midway point to counteract that deflection bringing your total to 6 bolts. That would be 1320 pounds total divided by 6 equals 220 pounds at each bolt.
I don't understand how varying the pressure at different places on the clamp will create an even pressure on the cell walls.
I would be very happy if someone could explain this to me. Thank you.
By the way, I love this forum, how everyone works to be helpful and considerate and all the good information that is passed around. I certainly wouldn't be where I am without it.
 
If you assume there might be deflection in the plate you might consider putting two more bolts at the midway point to counteract that deflection bringing your total to 6 bolts. That would be 1320 pounds total divided by 6 equals 220 pounds at each bolt.
I don't understand how varying the pressure at different places on the clamp will create an even pressure on the cell walls.
I would be very happy if someone could explain this to me. Thank you.
Your first assumption is that you want to use a plate and bolt configuration that does not allow enough deflection to worry about. A gross example of too much deflection would be trying to make an end plate out of aluminum foil. Since you want a calibrated total deflection, the springs should do most of the flexing while the plate stays flat.

The second assumption is that it is the 12 PSI that is important, so you multiply that by the surface area that will be under compression. In your example, that is 2 X 55 = 110 square inches. So the force you need is 12 X 2 X 55 = 1320 pounds applied by your non-deflecting plate. Divide that by the number of spring rods you plan to use.

The third assumption is that the geometric center of your fastener pattern lies at the geometric center of the areas under compression. If it isn't, you risk applying the force unevenly. A gross example would be putting eight spring rods at the bottom of the pack and none at the top.

Once you have a non-deflecting plate with the bolt pattern centered at the geometric center of the surface area of the cells, you are good. Spring rods between the two stacks of cells simply allow you control deflection of your plates a little better. In other words, they can be a bit less beefy since the cantilever on them is half as much.
 
Dear all,

I intend to use springs of green colour for compression on 12 mm rods as follows:

2 strings of 8 batteries lf280k
6 springs for both packs (2 one side 2 mid and 2 other side)
end plates oak of 40 mm thickness ( consider no flex)
each spring has the following characteristics
- green springs _ low load
- lenght non compressesd 127 mm
- compression at 25% 31.8 mm/531 N
- nothing in between cells as no clear what material is ok tbd

Tight them at 24 mm compression when battery discharged and i suppose they will go to 32 mm compression when battery charged. ( 8 mm travel)

Is still unclear for me how much the travel is 1 mm or 0.5 mm from empty to full. But this spring seems to cover a rather okish range of force for the application.

Any taughts?

Much appreciated
Flo
 
I don't know if this has been mentioned before, but in case it hasn't,
If building a rig intended to provide manufacturer's specified "clamping" force to a multi-cell pack (without compressible isolators between cells) using a fixed endplate and a spring-retained endplate, both end plates and the cells themselves will have to slide slightly on whatever surface they're resting on as the cells expand, slightly increasing the center-to-center spacing of the cells. Also, the busbars will have to be flexible enough to accommodate the change in center-to-center spacing that will occur twice with each charge-discharge cycle.

The only way to avoid this is to put the "springs" in between adjacent cells, fix the center-center distance at the top & bottom with hard plastic spacers (similar to Fortune prismatic cells), which will accommodate the expected cell expansion while providing the manufacturer's specified "clamping" force on each cell face without changing the center-to-center spacing between cells.

...or just put the cells in what has come to be known as the "snake" configuration, which also ends up putting both positive & negative end terminals on the same narrow end of the pack. Providing the manufacturer's specified "clamping" force is just as easy in this configuration, and more likely to be more consistent, because you have only two cells adjacent on the larger sides, instead of as many as 16.
 
Thanks for pointing this out.
The bus bar issue i will manage it with flexible connections, that is a minor issue
The “distribution of the force to the entire surface” as i will calll the need to manage the higher force created by the belly in the center of the cell i intend to manage it with some sort of compressible material… Not yet clear what i can find in my area.
I think about using 3 mm cork sheets with a not so big density like the one to be used under floors.

One remaining concern will be the lack of air flow using this kind of inter cell spacers.

I will experiment with resistor pressure senzors to see what is happening in different points of the cell and not overthighten the belly vs margins.

All the best
 
Is it possible to calculate how tight the nuts are tightened with a torque wrench (nm) that gets a pressure of 300kgf?

I have 4pcs 174x150mm cells, so the size of the pressure plates is 204x150x15mm (pvc) and four m6 rods.

Does tightening have to be done when the cells are full or empty?

I have read on the subject so much that my brain is in a knot. Can someone count it for me, please.
 
Last edited:
Is it possible to calculate how tight the nuts are tightened with a torque wrench (nm) that gets a pressure of 300kgf?

I have 4pcs 174x150mm cells, so the size of the pressure plates is 204x150x15mm (pvc) and four m6 rods.

Does tightening have to be done when the cells are full or empty?

I have read on the subject so much that my brain is in a knot. Can someone count it for me, please.
Such calculations are possible, and there are several being made earlier in the Thread. I think it's better to measure the spring compression distance, times the spring rating per millimeter of compression, to determine the compressive force being applied by each spring. Tightening should be done on cells which have not become swollen from high SOC - but you don't need to make them "nearly empty". Any charge state in the range of 20-60% SOC should create excellent results.
 
Such calculations are possible, and there are several being made earlier in the Thread. I think it's better to measure the spring compression distance, times the spring rating per millimeter of compression, to determine the compressive force being applied by each spring. Tightening should be done on cells which have not become swollen from high SOC - but you don't need to make them "nearly empty". Any charge state in the range of 20-60% SOC should create excellent results.
If I don't want to use springs, how tight do I tighten them? I install the cells in a small box where there will be thermal insulation, so the springs can be challenging to fit.
 
If I don't want to use springs, how tight do I tighten them? I install the cells in a small box where there will be thermal insulation, so the springs can be challenging to fit.
This Thread is ALL about optimizing with the of springs. Without springs, when the cells expand, your fixed length "compression rods" will not move at all. Too much cell expansion causes very high cell surface pressure. When the cell shrink in width, your fixed length "compression rods" stay wide, and the cells will fall completely loose from the rods.

With springs, shrunken cells have LESS pressure, but they are still being pressed together. As the cells expand, more and more total pressure is exerted by the springs (because that's how compression springs work). But the springs should never be allowed to bottom-out fully compressed, that would end in a situation similar to the one you propose - with far too much "pressure" exerted on the expanded cell packs. Springs optimize the pressure - a bit less when it isn't needed, more and more as the cells become expanded and distorted.

Can the rods and anchors extend outside the body of your box, through holes? In another widely used alternative, the space between the "extra" compression rod length (containing the springs and held with a nut, maybe a washer) is used for mounting the BMS unit against the end-side of the battery pack.
 
This Thread is ALL about optimizing with the of springs.
My mistake, sorry.

If I try to fit the springs, is an 80x20x2.5mm = 5.2kg/cm spring enough?

I calculated it myself and got something like the car's spring class, I don't think it's right ?
 
My mistake, sorry.

If I try to fit the springs, is an 80x20x2.5mm = 5.2kg/cm spring enough?

I calculated it myself and got something like the car's spring class, I don't think it's right ?
With only 5.2 kg/cm, you're nowhere near the compressive strength you will need. Adequate springs cost a lot (upwards of $6 each, and with a $40 minimum order total where I buy mine). If you want to use only 4 at roughly 170 lbs each, my most recent spring will not be adequate - but it would work for a set of 6, compressed to right around 120 lbs each. The free length is only 1-1/4", the maximum load is 168 lbs at 1.013 inches (fully compressed, with 0.237" of length pushed in). When the spring has been compressed by 0.169" (with total length down to 1.081") you will have reached 120 lbs of pressure.

This particular spring is very short, but has a very high rate (814 lbs per inch). In tightening this spring, you will want to use a micrometer to measure your spring length. "eyeballing', to the nearest mm or so, could result in significant errors. https://www.thespringstore.com/pc135-609-7500-mw-1250-cg-n-in.html Note also the relatively large interior diameter, if used with 1/4" diameter threaded rod it would need washers to avoid going off-center around the rod.
 
Last edited:
Is it possible to calculate how tight the nuts are tightened with a torque wrench (nm) that gets a pressure of 300kgf?

I have 4pcs 174x150mm cells, so the size of the pressure plates is 204x150x15mm (pvc) and four m6 rods.

Does tightening have to be done when the cells are full or empty?

I have read on the subject so much that my brain is in a knot. Can someone count it for me, please.

I am using Four 1/4 inch galvanized steel threaded rods on my 8S battery fixtures with the nuts torqued to approximately 5 inch pounds using a torque wrench while my cells were about 70% charged. The wood end plates are 1x10" with thin plastic sheets (gallon ziploc baggies with the top cut off) between each cell. The fixture is more or less just retaining the cells with very little pressure applied. The cells are what I would call snug, but not more.

I am currently not using a spring configuration.
 
Back
Top