rapidly serviceable battery settup

Joined
Oct 10, 2018
Messages
79
Hi guys,

looong time no post. I have been lurking and designing the chassis for my build and it has taken a long time to land for me as I want to be able to quickly service my modules (P120) and not have to solder or cut anything to do so. props to another forum member AJW22 here for their STL files that i have taken and modified to suit my purposes. Thanks mate!


image_ppconp.jpg


I have enclosed one side of my battery chassis (Negative will be in black, in the future thepositive side will be red) to create a top and bottom opening that will act as a duct for the heat that rises through it. the long mons can sit in the narrowed top opening to measure any difference in temperature as the air flows over them.

image_vwzrks.jpg


The plan is to haveconnections viasimple sprung battery terminals that arespot welded to 8mm nickel stripsrunning the length of the chassis and with criss cross connections. The nickel and sprung terminals will be attached via a compression plate on each side to match the profile of the chassis and seal off the ends of the batteries. one side will have glass fuses between sprung connection and nickel strip.

image_wfvabk.jpg


The whole idea is rapid deconstruction when cells need testing/replacing. as I am hoping for 6 banks of 14 modules serviceability becomes the priority.

image_vjiixt.jpg



8mm nickel strip is cheap.
the sprung battery terminals are cheap.
the glass fuses are cheap.

I am still in the prototype stages but can anyone bring some helpful critique to this design to improve it from here? open to experienced voices.
 
hm... the design looks oddly familiar ;-)

Looking forward to seeing how you plan to cram the nickel strip, sprung terminal, and fuse into/onto that.
Remember that nickel strips are very thin with limited current carrying capacity. So you may have to add a thick copper bus, or at least several attachment points between each pack.
 
ajw22 said:
hm... the design looks oddly familiar ;-)

Looking forward to seeing how you plan to cram the nickel strip, sprung terminal, and fuse into/onto that.
Remember that nickel strips are very thin with limited current carrying capacity. So you may have to add a thick copper bus, or at least several attachment points between each pack.

Hey mate,

Thanks again for the design share. I added the stop bars in each cell pocket to give me a consistent distance for spring connection.

Re the packaging the fuses are the real design challenge but will be incorporated into one side of the compression plate that sandwiches the whole thing together.

Re current yes. I will need to calculate my material thickness.
 
The 3D printed design looks nice .... but
I don't understand why so many powerwalls are designed with the cells so close together so that ventilation isn't good.

Can someone explain this?
 
Bubba said:
The 3D printed design looks nice .... but
I don't understand why so many powerwalls are designed with the cells so close together so that ventilation isn't good.

Can someone explain this?

Because heat is a non issue with these units as current draw is so low.
 
Bubba said:
The 3D printed design looks nice .... but
I don't understand why so many powerwalls are designed with the cells so close together so that ventilation isn't good.

Can someone explain this?


As Scottietheyoung says, cooling is not really required for a well dimensionedpowerwall. My typical max charge/discharge rateis justabout0.2A per cell, the average being much lower. It'stoo low to create any noticeable heat.

Nonetheless,I've given cooling a lot of thought when designing my packs. The main motivation being to keep any defective cell from overheating and causing more problems.
Firstly, there is plenty ofspace under the pack to draw in fresh cool air - very very important.
With the ubiquitous square cells layouts, the airflow would cool just the sides of the cells. The cell surfaces facing up/downreceive hardly any airflow.
With the honeycomb layout,the air has to navigate over all the cell surfaces. Moreover,the air gap between the cellsis actually a bit wider than the typical square designs, which I suppose may just offset the longer path the air has to take.
 
ajw22 said:
Bubba said:
The 3D printed design looks nice .... but
I don't understand why so many powerwalls are designed with the cells so close together so that ventilation isn't good.

Can someone explain this?

Also why I like this pack layout thanks AJW. Airflow is enough to provide better than passive cooling if in the unlikely scenario its needed...
 
Imagine the space extra needed for My powerwall that have 29000 cells. I wouldnt add something not needed during normal load :)
 
daromer said:
Imagine the space extra needed for My powerwall that have 29000 cells. I wouldnt add something not needed during normal load :)

you have a 29000 cell powerwall???????? not sure what you are saying above sorry?


Printing off the first of many of the positive side of the chassis. I ordered rolls and rolls of"watermelon red" PLA for the positive side of the chassis...turns out its pink :D :exclamation: :huh: :D

image_aedgnc.jpg



looks like I am stuck with Black for negative and PINK for positive lol.

anyway here is the basic concept of the sprung connectors with mechanical attachment to nickel strip.

image_uhvrrr.jpg



from here a sandwich pressure plate on each side screws on to give consistent tension across the surface.

image_zcpyxt.jpg



In this configuration my idea wasthere would be 9 nickel strips terminating (spot welded)onto a wide nickel plate for connection between 120P modules. I wasthinking about using magnets to connect between packs (either that or wooden pegs lol)

My question for those experienced in current and draw is what thickness nickel strip would be needed to carry the expected load? lines of 12-13 cells (will connect them across also)
 
Im saying that compacting the packs together is key for us doing larger installs :)
 
Scottietheyoung said:
anyway here is the basic concept of the sprung connectors with mechanical attachment to nickel strip.

image_uhvrrr.jpg

Have you done a resistance test on those springs? They look like the standard steel springs with nickel plating. You may find you loose a lot of power to heat while under medium to full load of the packs.
 
Korishan said:
Scottietheyoung said:
anyway here is the basic concept of the sprung connectors with mechanical attachment to nickel strip.

image_uhvrrr.jpg

Have you done a resistance test on those springs? They look like the standard steel springs with nickel plating. You may find you loose a lot of power to heat while under medium to full load of the packs.

Yes nickel plated steel springs.I am hoping that with 6 banks of 14 modules load will be extremely minimal for each cell/pack. though to beginwith I intend running two banks of 14.That should limit draw to less than .02 of an amp for each cell under full house hold load. Do you see this being an issue?
 
A problem with "steel" springs not only the higher resistance, but they become brittle over time from being compressed.

Let's say under 200mA load the steel spring wastes 10mA. Over an hour of that load is 100mAh, 10 hours is 1Ah. Now multiply that by how many cells you have in parallel. Even though the waste is small, it still can add up over time. Even if the waste is 5mA, it still adds up.
That's why I asked if you measured the resistance because it will tell you how much will be wasted.

Obviously, the more cells in parallel, the lower the waste as each cell will be less current flowing through. So, there is a kind of wasted energy curve getting closer to 0mA as the parallel count gets larger and the current per cell gets lower.
 
Korishan said:
A problem with "steel" springs not only the higher resistance, but they become brittle over time from being compressed.

Let's say under 200mA load the steel spring wastes 10mA. Over an hour of that load is 100mAh, 10 hours is 1Ah. Now multiply that by how many cells you have in parallel. Even though the waste is small, it still can add up over time. Even if the waste is 5mA, it still adds up.
That's why I asked if you measured the resistance because it will tell you how much will be wasted.

Obviously, the more cells in parallel, the lower the waste as each cell will be less current flowing through. So, there is a kind of wasted energy curve getting closer to 0mA as the parallel count gets larger and the current per cell gets lower.

I understand the resistance side of things but the only reason they become brittle is through heat cycling. So if current is low there is no reason to believe it will get hot and become brittle. I worked with all kinds of metals and spring steels as a fabricator for 10 years.


Scottietheyoung said:
Korishan said:
Let's say under 200mA load the steel spring wastes 10mA. Over an hour of that load is 100mAh, 10 hours is 1Ah. Now multiply that by how many cells you have in parallel. Even though the waste is small, it still can add up over time. Even if the waste is 5mA, it still adds up.
That's why I asked if you measured the resistance because it will tell you how much will be wasted.

Thanks for this thought angle, my thinking was that even with just two banks running of 120P 14S the maximum load a cell would see is .018 of an amp. Do you think the losses would be worth worrying about at that kind of draw? I had assumed the current being so low would negate the need for genuine concern here?
 
Scottietheyoung said:
I understand the resistance side of things but the only reason they become brittle is through heat cycling. So if current is low there is no reason to believe it will get hot and become brittle. I worked with all kinds of metals and spring steels as a fabricator for 10 years.

Kewls. You're way ahead of the game, then :)

Scottietheyoung said:
Thanks for this thought angle, my thinking was that even with just two banks running of 120P 14S the maximum load a cell would see is .018 of an amp. Do you think the losses would be worth worrying about at that kind of draw? I had assumed the current being so low would negate the need for genuine concern here?

That's where knowing the resistance of the springs/connectors will help out. If the resistance is really low, then it's probably no big deal. Also, measuring the resistance (especially if you do everyone one), you could find which ones might need a little extra solder if it's higher than the others. Altho, from the strip you showed, I think you got that part licked pretty good, too ;)
 
Back
Top