The secret to compression

stienman

Solar Addict
Joined
Jan 6, 2021
Messages
226
Compression affects lithium based cells. There are a number of mechanisms associated with compression, as each component of the cell is affected and must be taken into account separately and together.

This forum has many, many, many discussions about compression, and this isn't the first argument over whether a rigid method is better than a compliant method.

For those who really want to understand, here's a number of research papers that discuss the mechanisms that cause capacity degradation which can be reduced with mechanical compression: https://scholar.google.com/scholar?...apacity+degradation+lithium+compression&btnG=

If you choose to continue to assert your beliefs as true about why compression works and why your specific method is better/worse than another persons, please provide references, either to manufacturer's data sheets/application notes/etc or to published research.

Otherwise it's just a bunch of opinions based on second or third hand opinions from others, and you should be comfortable stating as such, with "I believe..." or "My anecdotal evidence suggests..." or "My opinion is...".

And for those who care, my opinion is:

I throw my lot in with those favoring compliant methods of compression.

As far as the pouch vs prismatic cell debate - prismatic cells are halfway between cylindrical and pouch cells. Cylindrical cells do expand, but are still under compression due to their mechanical construction. They don't need external compression.

Pouch cells are unconstrained, and may develop gaps between the materials in the cell which cause degradation. While the effect is significant under stress - such as the formation of gas bubbles due to electrolyte evaporation due to heat - this degradation will still happen under normal use over time, just on a smaller scale. There are many other aspects beside gas bubbles, that's just one example. Compression is absolutely necessary for pouch cells, as normal movement during charge/discharge will cause separation between materials in the pouch. Pressure also resists the state change from liquid to gas within the compressed area.

Prismatic cells are like cylindrical cells on the ends, and like pouch cells in the middle. The compression is NECESSARY on the larger faces for the same exact reasons you must compress pouch cells. The center of the cell is the only place where it's useful - the two edges of the cells on the large faces don't matter as much, and the top, sides, and bottom don't matter at all in terms of compression.

The only datasheet from a manufacturer I've seen regarding compression is EVE's 280AH cell, and compression is given in kgf - kilograms force. This is an old, outdated standard unit, but it is directly convertible to pressure given the size of the footprint. It was commonly used in applications involving spring force - such as bows, line tension, and, of course, springs.

After wading through these forums and the available research I've decided that for safety and cycle life, I'll compress to around 12psi at half charged state using either springs (ie, straps with known elasticity) or foam. Since I have to put something between each cell anyway, foam solves two issues at once. Given battery compression specific foams, this becomes trivial as I can dimension everything and not have to calibrate or measure beyond simple length measurements.

I've concluded that compression isn't just a cycle life issue, but a safety issue. Uncompressed, gas bubbles may form, and that may result in hotspots that eventually could contribute to thermal runaway even under normal conditions. Overcompressed, the mechanical damage that occurs may damage the separator in a cell, allowing high current discharge and thermal runaway. As I'd rather suffer the slower thermal runaway of gas bubble hotspots than the fast runaway of anode and cathode contact, I prefer to tend toward undercompression. However, with the appropriate foam or spring mechanism, it's easy enough to maintain the correct pressure throughout the charge/discharge cycle that I believe the maximum safety and cycle life will occur under compliant tension. A significant impact such as a hammer hit is just a localized overcompression - if the separator fails, you get thermal runaway.

All that said, while many think of threaded rods and other constructions as rigid, in the world of 300kgf forces, they are surprisingly springy. I expect that as long as you account for maximum cell expansion and avoid significant overcompression at the extreme end, then you're going to be just fine. The forces are so large, and the expansion so small, that a common threaded rod will expand as well, and the pressure might not become too great.

Now if your design involves six 1" hardened steel rods and 16 batteries in a row, then I'd suggest you're risking a bit too much.

All that said, I suspect we're spending a lot more time, individually, on this topic than it warrants. People who are using "rigid" mechanisms are probably compliant enough for their small group of cells. I haven't yet seen anyone on here show or describe a truly rigid mechanism.

Compress your cells, not too much, and move on.
 

stienman

Solar Addict
Joined
Jan 6, 2021
Messages
226
Particularly when you account for the fact that you should have a separator between each cell anyway. The foam provides:

- Electrical separation
- Thermal separation
- Compression
- Fire suppression (it has fire inhibitors - not enough to put out a fire, but enough to keep it from spreading quickly)
 

Q-Dog

¯\_(ツ)_/¯
Joined
Dec 22, 2020
Messages
640
Compression affects lithium based cells. There are a number of mechanisms associated with compression, as each component of the cell is affected and must be taken into account separately and together.

This forum has many, many, many discussions about compression, and this isn't the first argument over whether a rigid method is better than a compliant method.

For those who really want to understand, here's a number of research papers that discuss the mechanisms that cause capacity degradation which can be reduced with mechanical compression: https://scholar.google.com/scholar?hl=en&as_sdt=0,23&q=capacity+degradation+lithium+compression&btnG=

If you choose to continue to assert your beliefs as true about why compression works and why your specific method is better/worse than another persons, please provide references, either to manufacturer's data sheets/application notes/etc or to published research.

Otherwise it's just a bunch of opinions based on second or third hand opinions from others, and you should be comfortable stating as such, with "I believe..." or "My anecdotal evidence suggests..." or "My opinion is...".

And for those who care, my opinion is:

I throw my lot in with those favoring compliant methods of compression.

As far as the pouch vs prismatic cell debate - prismatic cells are halfway between cylindrical and pouch cells. Cylindrical cells do expand, but are still under compression due to their mechanical construction. They don't need external compression.

Pouch cells are unconstrained, and may develop gaps between the materials in the cell which cause degradation. While the effect is significant under stress - such as the formation of gas bubbles due to electrolyte evaporation due to heat - this degradation will still happen under normal use over time, just on a smaller scale. There are many other aspects beside gas bubbles, that's just one example. Compression is absolutely necessary for pouch cells, as normal movement during charge/discharge will cause separation between materials in the pouch. Pressure also resists the state change from liquid to gas within the compressed area.

Prismatic cells are like cylindrical cells on the ends, and like pouch cells in the middle. The compression is NECESSARY on the larger faces for the same exact reasons you must compress pouch cells. The center of the cell is the only place where it's useful - the two edges of the cells on the large faces don't matter as much, and the top, sides, and bottom don't matter at all in terms of compression.

The only datasheet from a manufacturer I've seen regarding compression is EVE's 280AH cell, and compression is given in kgf - kilograms force. This is an old, outdated standard unit, but it is directly convertible to pressure given the size of the footprint. It was commonly used in applications involving spring force - such as bows, line tension, and, of course, springs.

After wading through these forums and the available research I've decided that for safety and cycle life, I'll compress to around 12psi at half charged state using either springs (ie, straps with known elasticity) or foam. Since I have to put something between each cell anyway, foam solves two issues at once. Given battery compression specific foams, this becomes trivial as I can dimension everything and not have to calibrate or measure beyond simple length measurements.

I've concluded that compression isn't just a cycle life issue, but a safety issue. Uncompressed, gas bubbles may form, and that may result in hotspots that eventually could contribute to thermal runaway even under normal conditions. Overcompressed, the mechanical damage that occurs may damage the separator in a cell, allowing high current discharge and thermal runaway. As I'd rather suffer the slower thermal runaway of gas bubble hotspots than the fast runaway of anode and cathode contact, I prefer to tend toward undercompression. However, with the appropriate foam or spring mechanism, it's easy enough to maintain the correct pressure throughout the charge/discharge cycle that I believe the maximum safety and cycle life will occur under compliant tension. A significant impact such as a hammer hit is just a localized overcompression - if the separator fails, you get thermal runaway.

All that said, while many think of threaded rods and other constructions as rigid, in the world of 300kgf forces, they are surprisingly springy. I expect that as long as you account for maximum cell expansion and avoid significant overcompression at the extreme end, then you're going to be just fine. The forces are so large, and the expansion so small, that a common threaded rod will expand as well, and the pressure might not become too great.

Now if your design involves six 1" hardened steel rods and 16 batteries in a row, then I'd suggest you're risking a bit too much.

All that said, I suspect we're spending a lot more time, individually, on this topic than it warrants. People who are using "rigid" mechanisms are probably compliant enough for their small group of cells. I haven't yet seen anyone on here show or describe a truly rigid mechanism.

Compress your cells, not too much, and move on.
I randomly clicked on one link in that list of articles you probably haven't read and found this in the abstract of one paper: This thesis experimentally investigates the effects of compression on pouch cells. Cells made from stacked electrode layers, were compressed between 0 MPa – 0.43 MPa (@ 40°C,50% SOC) and cycled 2000 times. It was found that pressure had no effect on the rate of degradation. Compression improved the volumetric energy density by 10% at 0.43 MPa down to a 5% improvement at 0.02 MPa.
Only time will tell.
 

Short_Shot

Photon Sorcerer
Joined
Jul 13, 2021
Messages
1,820
I randomly clicked on one link in that list of articles you probably haven't read and found this in the abstract of one paper: This thesis experimentally investigates the effects of compression on pouch cells. Cells made from stacked electrode layers, were compressed between 0 MPa – 0.43 MPa (@ 40°C,50% SOC) and cycled 2000 times. It was found that pressure had no effect on the rate of degradation. Compression improved the volumetric energy density by 10% at 0.43 MPa down to a 5% improvement at 0.02 MPa.
Only time will tell.
However, only cycling 2000 times may not be sufficient to start showing evidence of degradation in that "test".
 

RCinFLA

Photon Sorcerer
Joined
Jun 21, 2020
Messages
1,147
The EVA spec has created a false impression on compression.

The chart is for a 1C charge and discharge rates where cell internal self heating is a factor.

It takes over two hours to do one test cycle at 1C charge and discharge. The chart peaks at 20,000 cycles. At 2 hours per cycle that would take 40,000 hours or 4.56 years to complete just that one data point on the chart. Do you really think that was actually done or just a theoretical estimate projection?

Below 0.5C charge and discharge rates there is little cell internal self heating and compression benefit is questionable.

Hard fixed, flat surface compression can create extreme pressure points which can cause a separator punch through cell short.
I would not be surprised if this was fhorst meltdown issue.
 
Last edited:

stienman

Solar Addict
Joined
Jan 6, 2021
Messages
226
Only time will tell.

Only careful reading will tell. In this case they used a rigid fixture, and measured the pressure. In most cases they observed 0 compression when the cells were empty, thus the cells were not compressed correctly. They performed relatively few cycles - 200 at their maximum - less than 10% of a typical lithium cell's minimum rated cycles. Their pressure sensors, they admitted, were not the correct range for this project and the measurements were inaccurate within the pressure levels these cells are supposed to be under.

There's a lot more to this paper, but if you really want to use it as an example against compression, please understand the test methodology before taking the abstract and applying it to situations it doesn't apply to.

Full paper access here: https://spiral.imperial.ac.uk/handle/10044/1/68739

The one thing I'd pay attention to in this paper is its excellent discussion and literature review that precedes the test results, and the excellent bibliography.

Further, a good discussion is had in section 5.1 which shows how the jelly roll configuration (which we are using, and they were not testing) differs from the pouch configuration (which is all they tested).

So while I had, up until now, assumed they were similar enough to be considered similarly in terms of compression, this paper makes several good points as to why the configuration does matter, and why their work doesn't necessarily apply to jelly roll configurations.
 

Turponieminen

New Member
Joined
Sep 17, 2020
Messages
21
I am also thinking why cylindrical cells are not rated 6000+ cycles since they don't need compression?
 

cinergi

1.21 Jigawatts
Joined
Aug 9, 2020
Messages
1,166
BTW with respect to threaded rods providing enough (or too much) spring-like force ... I found that you cannot rely on them for this. I've used anywhere from 1/4" to 3/8" medium and high strength and in all conditions, you're talking about a minimum of 2mm of movement (4 EVE 280's) which is far more than those rods can properly handle (tight at 100% SoC, loose at 0% SoC). Yes, they flex - but not in a useful way for this application.
 

Q-Dog

¯\_(ツ)_/¯
Joined
Dec 22, 2020
Messages
640
Only careful reading will tell. In this case they used a rigid fixture, and measured the pressure. In most cases they observed 0 compression when the cells were empty, thus the cells were not compressed correctly. They performed relatively few cycles - 200 at their maximum - less than 10% of a typical lithium cell's minimum rated cycles. Their pressure sensors, they admitted, were not the correct range for this project and the measurements were inaccurate within the pressure levels these cells are supposed to be under.

There's a lot more to this paper, but if you really want to use it as an example against compression, please understand the test methodology before taking the abstract and applying it to situations it doesn't apply to.

Full paper access here: https://spiral.imperial.ac.uk/handle/10044/1/68739

The one thing I'd pay attention to in this paper is its excellent discussion and literature review that precedes the test results, and the excellent bibliography.

Further, a good discussion is had in section 5.1 which shows how the jelly roll configuration (which we are using, and they were not testing) differs from the pouch configuration (which is all they tested).

So while I had, up until now, assumed they were similar enough to be considered similarly in terms of compression, this paper makes several good points as to why the configuration does matter, and why their work doesn't necessarily apply to jelly roll configurations.
From the paper: "The effects of temperature on the electrochemical performance of the battery are significantly larger than those of any compression."

I don't need to read more. There really is no conclusion. Mostly it says it depends, or the topic needs more study. Fine. I could pick a bunch of stuff out of this paper and make arguements either way all day long. I have better things to do with my life. In a few years we will have real answers or better batteries.

My last set of battery cells were bulging when they came off the hot truck then shrank as they cooled down in my house. And my observations on my constrained and unconstrained packs show clearly that when I charge them at high rate they warm up a few degrees and swell, but when charged at lower rates they don't warm up and don't swell. My conclusion, keep the cells cool and charge at the minimum rate to accomplish your goals. I'll keep mine in a snug fitting space and find something else to worry about. Have fun with your springs and rods and washers if that makes you feel better.

Looking at photos of pack failures when the cells short or get overcharged, they swell up. Is that heat causing the swelling? Or chemical reaction? Both? Probably needs more study? Sure, but not by me.
 

jmole

Solar Enthusiast
Joined
Jan 25, 2021
Messages
19
Looking at photos of pack failures when the cells short or get overcharged, they swell up. Is that heat causing the swelling? Or chemical reaction? Both? Probably needs more study? Sure, but not by me.
Both - typically it’s thermal runaway, which causes, and is exacerbated by swelling. And another good reason to use compression pads between cells, since it will thermally isolate them to some degree. This is a big deal in vehicle applications where you don’t want a single hot cell to destroy your entire bank.
 

Gazoo

Dumb Dumb
Joined
May 12, 2020
Messages
2,810
Location
Texas
Below 0.5C charge and discharge rates there is little cell internal self heating and compression benefit is questionable.
My EVE cells expand quite a bit using very low apx. .06 C charge rates.

Hard fixed, flat surface compression can create extreme pressure points which can cause a separator punch through cell short.
I would not be surprised if this was @ fhorst meltdown issue.
I don't recall at what SOC he compressed his cells but if the cells were compressed at a lower SOC then that makes sense.
 

RCinFLA

Photon Sorcerer
Joined
Jun 21, 2020
Messages
1,147
My EVE cells expand quite a bit using very low apx. .06 C charge rates.
Graphite Anode Li-Ion cells expand when charged. Only matters that it is charged, not the rate it was charged

Compression does not prevent expansion of graphite when charged or electrolyte decomposition bloating if overcharged. Its intended purpose is to reduce delamination.
 

rhino

Solar Addict
Joined
Jun 6, 2020
Messages
522
Location
Minnesota

Also known as Poron:

Foam compresses 25% at 11psi.

If you buy 1/4" foam, then plan for 3/16" between your cells when you design your busbars and your box, and you have 11psi compression with no springs and no hassle.

Worked great for me.

I designed my box to be slightly longer than the uncompressed length, and added a compression plate to push the cells and foam together to achieve the compression I wanted.
Curious if you tried testing what happens to the foam when exposed to open flame preferably a blow torch.
 

Short_Shot

Photon Sorcerer
Joined
Jul 13, 2021
Messages
1,820
Curious if you tried testing what happens to the foam when exposed to open flame preferably a blow torch.
It would be interesting to see, but McMaster lists this stuff as fire retardant so I wouldn't worry too much about needing proof.

The safety data sheet shows virtually zero concern about flammability other than the expected dangers of polymers being burned and giving off smoke. Flame retardant doesn't necessarily mean it 'doesn't burn' so much as it can be self extinguishing.
 

RCinFLA

Photon Sorcerer
Joined
Jun 21, 2020
Messages
1,147
Fire retardant is of secondary importance to me.

Getting the right durometer, thickness of material, and appropriate amount of compression percent on the pad to achieve desired psi is toughest part.

Then have a material that does not take a set and relax over time/ambient time is next. The lower the required percent thickness compression necessary on pad the less problems with it taking a relaxation set over time. No more then 20% thickness compression is a good number.

Next important criteria is some thermal insulation to prevent a cell with thermal runaway from transferring too much heat to next door neighbor cells setting off a full cascade battery array meltdown.

Red silicon rubber pads are low cost and checks many of the requirement boxes, even pretty good fire resistance rating
 
Last edited:

Zwy

Solar Addict
Joined
Jan 3, 2021
Messages
1,050
Top