Project Planning

Wolf

Well-known member
Joined
Sep 25, 2018
Messages
1,461
I'm still having a hard time wrapping my head around.
Ah the wrap your head around the whole idea thing.
Yea initially it is somewhat confusing.
Think of it this way, if you are building a 14s9p Lets make it 14s10p easier to draw out. That are 14 series packs with 10 parallel cells in each pack making a 48V nominal Battery. Each 10p pack is a battery by itself but to achieve your nominal voltage of 48v you need 14 of these packs (batteries) in series. Each "pack will have 1 longmon to measure the pack voltage of 3.7 nominal.
1622733003220.png
If you build a single 14s10p battery in one chunk you will have a single battery with 140 cells that looks like this
1622734313922.png
In either case you will need a longmon on each individual 3.7v parallel section.
The former model is easier to service as you can have a single 3.7v 10p pack ready to change out if you are experiencing an issue with a pack.
The latter you need to take the whole pack off line to service 1 parallel section.
Also if you follow good building guidelines as in IR checking and making sure every cell is up to the strictest standards you should not experience any issues whether you go with design 1 or 2.
Hope this helps some of the head wrapping.

Wolf
 

OffGridInTheCity

Well-known member
Joined
Dec 15, 2018
Messages
1,591
Maybe it would help to define some terms.

Let's define a 'pack' as a group of cells arranged physically as a 'unit' and wired in parallel + and - connections coming out. A 9p pack would have 9 cells. In the picture above, you see 89 cells wired to bus-bars that with bolt-on lugs coming off the left end.

A pack will act as if its 1 'big' cell. https://batteryguy.com/kb/knowledge-base/connecting-batteries-in-parallel/
If an individual 18650 cell (in the pack) is weaker power will flow from the other cells (along the buss-bar) to maintain it's voltage to be the same as all the other cells (in the pack) and vice versa the other way. If you add cells, the voltage remains the same but you increase the pack's power potential.
**Its best that all packs in a DIY powerwall have (pretty close) to the same power potential - e.g. ah rating.
** Each pack will need a longmon. Longmons can handle up to several hundred cells in a single pack. Its not wrong but it's a waste of $ to put a longmon on a 9cell pack - the longmon was designed for 100p - 500p type packs.


Once you have a bunch of packs....

If you want to build a 48v battery then you wire 14 packs in series. This is because the 18650 lithium-ion voltage range is such that 14 in series is a perfect match for 48v equipment.
1622735220127.png


If you want to increase the battery you need to add more packs... in groups of 14. There are 2 methods.

The most common way is to build a 2nd battery of 14 packs and then wire that 2nd battery in parallel to the 1st battery. A cost to this is that you need 14 more longmons - for a total of 28. The upside is you don't have to disturb battery #1.
1622735392744.png


You 'could' double the packs and just have 1 48v battery. This would require only 14 longmons - as each pack is still a 'pack', just double in size. A downside to this approach is you'll likely need to dissemble battery #1 to make the physical changes but you don't need to spend $ on more longmons.
1622736052792.png
 

Attachments

  • 1622736037023.png
    1622736037023.png
    25 KB · Views: 13
Last edited:

Doc3G

Member
Joined
Apr 18, 2021
Messages
56
Ah the wrap your head around the whole idea thing.
Yea initially it is somewhat confusing.
Think of it this way, if you are building a 14s9p Lets make it 14s10p easier to draw out. That are 14 series packs with 10 parallel cells in each pack making a 48V nominal Battery. Each 10p pack is a battery by itself but to achieve your nominal voltage of 48v you need 14 of these packs (batteries) in series. Each "pack will have 1 longmon to measure the pack voltage of 3.7 nominal.
View attachment 25223
If you build a single 14s10p battery in one chunk you will have a single battery with 140 cells that looks like this
View attachment 25225
In either case you will need a longmon on each individual 3.7v parallel section.
The former model is easier to service as you can have a single 3.7v 10p pack ready to change out if you are experiencing an issue with a pack.
The latter you need to take the whole pack off line to service 1 parallel section.
Also if you follow good building guidelines as in IR checking and making sure every cell is up to the strictest standards you should not experience any issues whether you go with design 1 or 2.
Hope this helps some of the head wrapping.

Wolf

Thanks for the drawing, it helps a bunch. So, for the first drawing, each group of 10 has one longmon, that much I get. For the second, each group of 10 also had one longmon?

Thus, if I were to scale everything up, I could build 14 100P packs and still only need 14 longmons (assuming I follow strict cell selection). However, if I wanted to build the second design, each 14s10P pack would require 14 longmons, correct? Or, if I built the 10 of the 14s10P packs, would I be able to tie all ten of the first group of cells in each pack together with one longmon?
 

Attachments

  • Screenshot_528.png
    Screenshot_528.png
    328.3 KB · Views: 13

OffGridInTheCity

Well-known member
Joined
Dec 15, 2018
Messages
1,591
Thanks for the drawing, it helps a bunch. So, for the first drawing, each group of 10 has one longmon, that much I get. For the second, each group of 10 also had one longmon?
Yes. In the second, its still 14 packs (e.g. groups of 10 cells in parallel). The physical arrangement of cells is typically done for ease of hooking up the wiring/buss - but it's important to remember that physical is not what determines the electrical hookup. You can do many different physical arrangements and still wire it up as 14s10p :)

Thus, if I were to scale everything up, I could build 14 100P packs and still only need 14 longmons (assuming I follow strict cell selection).
Yes.

However, if I wanted to build the second design, each 14s10P pack would require 14 longmons, correct?
Yes

Or, if I built the 10 of the 14s10P packs, would I be able to tie all ten of the first group of cells in each pack together with one longmon?
Unsure what you mean.
 

Doc3G

Member
Joined
Apr 18, 2021
Messages
56
Maybe it would help to define some terms.

Let's define a 'pack' as a group of cells arranged physically as a 'unit' and wired in parallel + and - connections coming out. A 9p pack would have 9 cells. In the picture above, you see 89 cells wired to bus-bars that with bolt-on lugs coming off the left end.

A pack will act as if its 1 'big' cell. https://batteryguy.com/kb/knowledge-base/connecting-batteries-in-parallel/
If an individual 18650 cell (in the pack) is weaker power will flow from the other cells (along the buss-bar) to maintain it's voltage to be the same as all the other cells (in the pack) and vice versa the other way. If you add cells, the voltage remains the same but you increase the pack's power potential.
**Its best that all packs in a DIY powerwall have (pretty close) to the same power potential - e.g. ah rating.
** Each pack will need a longmon. Longmons can handle up to several hundred cells in a single pack. Its not wrong but it's a waste of $ to put a longmon on a 9cell pack - the longmon was designed for 100p - 500p type packs.


Once you have a bunch of packs....

If you want to build a 48v battery then you wire 14 packs in series. This is because the 18650 lithium-ion voltage range is such that 14 in series is a perfect match for 48v equipment.
View attachment 25226

If you want to increase the battery you need to add more packs... in groups of 14. There are 2 methods.

The most common way is to build a 2nd battery of 14 packs and then wire that 2nd battery in parallel to the 1st battery. A cost to this is that you need 14 more longmons - for a total of 28. The upside is you don't have to disturb battery #1.
View attachment 25227

You 'could' double the packs and just have 1 48v battery. This would require only 14 longmons - as each pack is still a 'pack', just double in size. A downside to this approach is you'll likely need to dissemble battery #1 to make the physical changes but you don't need to spend $ on more longmons.
View attachment 25229

OK, I think I've got it now. So the gist is that, at least if using the batrium system, building 48-volt packs and stringing them together is not really a viable option.

My hope was to, for example,

1. build a couple of 14s9p packs and string them together and makes sure everything worked
2. Finish building 14s9P packs out of my remaining ring batteries (I think I have ~ 1400), and string the rest of those packs in parallel for V1 of the powerwall.
3. At some point in the future if I have time and find a good deal on other 1860 batteries, to build additional 48 volt packs in 14sXp configuration to achieve packs with roughly the same ah rating as those in V1 of the powerwall.

Thus, I was hoping to be able to scale it up a little at a time as I had free time, but it seems like, at least if I'm using the batrium bms, it's something where you need to commit to a big chunk of power at once in order to make it cost effective.

It seems like what I would need for the scalable version would be a quality 14s BMS for each 14s9p pack in the ~$50 range, which doesn't seem to exist.

So then, the next big thing that I want to avoid is all the soldering of individual fuses. I really want to use the cell fusing strips that battery hookup has, but I don't quite understand how the design plays out in planning my pack sizes.

They state on their website This 5p nickel fuse is perfect for the 4 cell wide parallel packs. You can connect the 5th row to a wire or busbar that runs up the length of the pack.

They also offer it in 6p wide, and I saw in a video by HBpowerwall where he uses the traditional copper wire soldered to the BH nickel strips.


I'm just unsure if there is a way to make the packs any wider than 6p using this strip while maintaining the protection of cell level fusing.
 

Wolf

Well-known member
Joined
Sep 25, 2018
Messages
1,461
Thus, if I were to scale everything up, I could build 14 100P packs and still only need 14 longmons (assuming I follow strict cell selection).
For a 14s100p i.e 100p packs X 14 in series yes 14 longmons.
However, if I wanted to build the second design, each 14s10P pack would require 14 longmons, correct?
Correct each parallel section would require a longmon.
Or, if I built the 10 of the 14s10P packs, would I be able to tie all ten of the first group of cells in each pack together with one longmon?
No that cannot be done as each of the 14s10p belong to its own individual battery.
Here is a drawing of how to work an expansion where only 14 longmons are required.
1622738005042.png
For the other option 2 you will require 14 longmons for each battery.
1622738305110.png
Wolf
 

Doc3G

Member
Joined
Apr 18, 2021
Messages
56
For a 14s100p i.e 100p packs X 14 in series yes 14 longmons.

Correct each parallel section would require a longmon.

No that cannot be done as each of the 14s10p belong to its own individual battery.
Here is a drawing of how to work an expansion where only 14 longmons are required.
View attachment 25231
For the other option 2 you will require 14 longmons for each battery.
View attachment 25232
Wolf

OK, that clears that up, thanks.
 

Wolf

Well-known member
Joined
Sep 25, 2018
Messages
1,461
They state on their website This 5p nickel fuse is perfect for the 4 cell wide parallel packs.
True but I would look at a bus bar system like mine, not that it is the answer for everyone but it works really good. A bit labor intensive but it spreads the miniscule resistance of the nickel fuse sheet around very well. You can see my build here https://secondlifestorage.com/index.php?threads/wolfs-powirwall.7804/post-61191
I went with 5 cells wide as it worked really good for my bus bar design. Good support for the pos and neg connections in my case. Wth 4 wide you can get away with 2 bus runs and just use a terminal connector to join the loop.
Example
Wolf
1622740608170.png
 

Doc3G

Member
Joined
Apr 18, 2021
Messages
56
True but I would look at a bus bar system like mine, not that it is the answer for everyone but it works really good. A bit labor intensive but it spreads the miniscule resistance of the nickel fuse sheet around very well. You can see my build here https://secondlifestorage.com/index.php?threads/wolfs-powirwall.7804/post-61191
I went with 5 cells wide as it worked really good for my bus bar design. Good support for the pos and neg connections in my case. Wth 4 wide you can get away with 2 bus runs and just use a terminal connector to join the loop.
Example
Wolf
View attachment 25233

I like your design but it might be a little labor-intensive for the short amount of vacation I have left. Is 6 wide viable with two bus runs (spaced evenly) as in the picture above?

The next question I have is about spot welding. For the first ~ 100 cells or so I was taking the nickel strips off the batteries with various methods. Then I read a post by one of the mods that noted it was preferable to leave the old nickel on since it was easier to spot weld nickel to nickel and because it allowed you to put less heat into the cell when welding. That seems to make sense so I've been leaving the old strips on since then. However, some of the strips are pretty small and I'm not sure there is enough strip left to spot weld to. Should I remove these undersized strips and weld directly onto these batteries, or should I remove the old, spot weld a new piece of nickel, and then spot weld my fuse level sheet from battery hookup to the new nickel. The latter seems to make sense because I imagine the settings on the kweld will be different for joining directly to the cells vs joining nickel strips to other nickel strips.

Picture for reference:
 

floydR

Well-known member
Joined
Aug 23, 2017
Messages
1,150
There is a piece of nickel there? ;) welding over that should pose no problem just make sure you apply a firm pressure to the electrodes. although I am no welding expert.
later floyd
 

Wolf

Well-known member
Joined
Sep 25, 2018
Messages
1,461
short amount of vacation I have left
Is 6 wide viable with two bus runs (spaced evenly)
Oh a vacation project? Well then we must proceed with all post haste and forget about planning and careful battery construction. I do not mean to be critical here but lithium ion cells of any kind are dangerous to a certain degree. Similar to having a 5 gallon gas can in your shed. With proper care and caution the fuel will never cause an issue but step beyond the bounds of safety, who knows what can happen. Well I do, seen it, not pretty.
So if this is a speed race to get a 14s XXp battery up and running be my guest, but I persuade you to step back and really thing about the design, future expansion possibilities, and above all safety. It took me almost 2 years to build my first battery, from finding cells, testing, and construction to finally putting it into production. OK enough of that.
Since you are using fuse level sheets I suppose you can get away with 2 bus runs as the nickel cover sheet is a reasonably good conductor.
If you are going 6 wide (I assume you haven't ordered the holders yet) then I would run this type of bus. But that's just me. I believe in give electrons the least amount of resistance to cry about and they will treat you right. True story I've seen them cry they emit Pions (π) and cause heat. 🤣 Anyway.
1622806674241.png

Can you get away with this design? I believe so
1622806750602.png
leave the old nickel on since it was easier to spot weld nickel to nickel and because it allowed you to put less heat into the cell when welding
There I have my way of doing it which is remove as much of the strip as possible clean the weld stalagmites or is it stalactites as much as possible. Don't want to make the electrons cry because of resistance. Also if you are soldering sometimes it is easier to solder to the nickel strip rather than straight to the battery. I personally never had a problem soldering directly to the battery. Proper technique, proper solder, proper heat, and no issue. Just takes a bit of practice "and" the right equipment. I personally don't solder as I do use the nickel sheets but if I didn't have them I would not be opposed to it. OK back to technique. Just before spot welding, with the cells in their holders I inspect each pos and neg side with a magnifying glass and make one last verification of the cells physical condition. If any strip residue is visible I try removing it as best as possible with flush cutters. I do not dremel of sand as it may take off the thin coating to protect the terminals from rusting or corroding. The final touch before spot welding, is, I take a heavy rounded steel ( In my case scissor handles (yea I still have really good scissors with steel handles not plastic junk ) ) and run them across the terminals on all the cells as they are in the holders in a medium pressure rubbing motion that rounds off the stalagmites or is it stalactites. It helps a lot with spot welding. I also do 2 welds per connection so I have 4 spot welds on each connection. Again maybe overkill but thats me. Spot Welding is the way to go as it introduces the least amount if any heat into the cell.
Wolf
20210212_200622.jpg151636481_1137445830028625_9029565754818234303_n.jpg
 
Last edited:

Doc3G

Member
Joined
Apr 18, 2021
Messages
56
Oh a vacation project? Well then we must proceed with all post haste and forget about planning and careful battery construction. I do not mean to be critical here but lithium ion cells of any kind are dangerous to a certain degree. Similar to having a 5 gallon gas can in your shed. With proper care and caution the fuel will never cause an issue but step beyond the bounds of safety, who knows what can happen. Well I do, seen it, not pretty.
So if this is a speed race to get a 14s XXp battery up and running be my guest, but I persuade you to step back and really thing about the design, future expansion possibilities, and above all safety. It took me almost 2 years to build my first battery, from finding cells, testing, and construction to finally putting it into production. OK enough of that.
Since you are using fuse level sheets I suppose you can get away with 2 bus runs as the nickel cover sheet is a reasonably good conductor.
If you are going 6 wide (I assume you haven't ordered the holders yet) then I would run this type of bus. But that's just me. I believe in give electrons the least amount of resistance to cry about and they will treat you right. True story I've seen them cry they emit Pions (π) and cause heat. 🤣 Anyway.
View attachment 25239

Can you get away with this design? I believe so
View attachment 25240

There I have my way of doing it which is remove as much of the strip as possible clean the weld stalagmites or is it stalactites as much as possible. Don't want to make the electrons cry because of resistance. Also if you are soldering sometimes it is easier to solder to the nickel strip rather than straight to the battery. I personally never had a problem soldering directly to the battery. Proper technique, proper solder, proper heat, and no issue. Just takes a bit of practice "and" the right equipment. I personally don't solder as I do use the nickel sheets but if I didn't have them I would not be opposed to it. OK back to technique. Just before spot welding, with the cells in their holders I inspect each pos and neg side with a magnifying glass and make one last verification of the cells physical condition. If any strip residue is visible I try removing it as best as possible with flush cutters. I do not dremel of sand as it may take off the thin coating to protect the terminals from rusting or corroding. The final touch before spot welding, is, I take a heavy rounded steel ( In my case scissor handles (yea I still have really good scissors with steel handles not plastic junk ) ) and run them across the terminals on all the cells as they are in the holders in a medium pressure rubbing motion that rounds off the stalagmites or is it stalactites. It helps a lot with spot welding. I also do 2 welds per connection so I have 4 spot welds on each connection. Again maybe overkill but thats me. Spot Welding is the way to go as it introduces the least amount if any heat into the cell.
Wolf
View attachment 25242View attachment 25241

Haha, thanks for the info, I appreciate all of the detail. On the cells where I have had to remove the nickel entirely for one reason or another, I've been using the side of my sidecutters to round off the nubs too.

I'm definitely concerned about safety, I'm not trying to throw caution to the wind, which is why I'm asking questions before hooking everything up :D. I'm just trying to figure out what is practical.
 

Redpacket

Well-known member
Joined
Feb 28, 2018
Messages
1,283
I was watching some videos on installing the longmons yesterday and it looks like it just hooks to the positive and negative of a pack? So, does that mean the individual cells aren't monitored? My understanding is that in a theoretically ideal system the BMS would monitor each individual cell?
A Longmon connects across a "1s" group of cells, eg one of Off grids packs in the pic above. Yes they connect straight to the cell +ve & -ve with comms wires to the next longmon in the loop. in a 14s pack you have 14 Longmons, one for each "1s" group of cells.
 

Doc3G

Member
Joined
Apr 18, 2021
Messages
56
OK, all the cells have been charged. Of course, the most recent ones still need to sit for a week or so but most of them have been sitting long enough that I can begin to sort through them to find any bad cells. At this point, I have about 1400 cells that I dated (the date they were charged) and recorded discharged capacity (all on Liitokala chargers). They are separated into 100 mAh bins.

My MY1035+ finally arrived and I've been spot-checking IR, but I plan to use it to check the voltage and resistance of each cell as I go back through them.

My understanding is that I can either pack via the 100 mAh bin method or use either the excel repacker or the web repacker. If I use the excel repacker I guess it makes sense to build some sort of grid container to number the cells, thus making them easier to locate?

I know wolf had mentioned in the excel repacker thread that he would like to see a controlled experiment using the 100 mAh bin method and I would consider trying it. What would be the conditions required to make a reasonable comparison to packs built with the repacker?

 
Last edited:

OffGridInTheCity

Well-known member
Joined
Dec 15, 2018
Messages
1,591
If those are RING cells (above) then its probably about 60% blue / 40% green - I'd just evenly distribute by 100mah differences and cells type and I think you'd get very good results.

The OPUS (type) test process is not exact - e.g. you could easily get 50mah differences retesting the same cells every time. The packers are not wrong but look like too much work (to me) for 1,000(s) of cells and I'm not sure how beneficial it is. If you can make the battery with all the same cell types - its even better.

The cargo trailer 14s88p is made with all blue RING cells and 100mah method - e.g. very similar mah and IR. Look at it today - with no balance since last Feb's 20mv sync up - same 20mv difference as then. Of course this pack is not nearly as heavily used as the home powerwall.
1623697512984.png


Compared to my overall 84 packs (100mah method and a few different cell types) which runs 40mv to 70mv max difference.
1623698074979.png
 
Last edited:

Doc3G

Member
Joined
Apr 18, 2021
Messages
56
If those are RING cells (above) then its probably about 60% blue / 40% green - I'd just evenly distribute by 100mah differences and cells type and I think you'd get very good results.

The OPUS (type) test process is not exact - e.g. you could easily get 50mah differences retesting the same cells every time. The packers are not wrong but look like too much work (to me) for 1,000(s) of cells and I'm not sure how beneficial it is. If you can make the battery with all the same cell types - its even better.

The cargo trailer 14s88p is made with all blue RING cells and 100mah method - e.g. very similar mah and IR. Look at it today - with no balance since last Feb's 20mv sync up - same 20mv difference as then. Of course this pack is not nearly as heavily used as the home powerwall.
View attachment 25408

Compared to my overall 84 packs (100mah method and a few different cell types) which runs 40mv to 70mv max difference.
View attachment 25409

When you say "make a battery with all the same cell types" do you mean all panasonics or all LG's? You are correct that all of the cells are from the rings packs and probably pretty close to a 60/40 distribution in favor of the LG's.
 

floydR

Well-known member
Joined
Aug 23, 2017
Messages
1,150
6 blue 4 green ratio in 100mAh groups
looks like you have 3 cell groups 3000-3099, 3100-3199, 3200-3299 this is assuming that the 60/40 ratio is the same across all the groups so 33 cells from each group 18 blue 14 green+1 your choice per 100mAh group and 1 again your choice to make each 100 parallel pack?
Least that is how I understand what OGITC said.

Later floyd
So NOT what OGITC meant/said.
 
Last edited:

OffGridInTheCity

Well-known member
Joined
Dec 15, 2018
Messages
1,591
When you say "make a battery with all the same cell types" do you mean all panasonics or all LG's? You are correct that all of the cells are from the rings packs and probably pretty close to a 60/40 distribution in favor of the LG's.
Yes - all blue / all green. I bought extra RING packs to make this happen... and created an all blue for the trailer and all green NCR18650A(s) to extend the powerwall. Here's the all green 14s88p ready to install...
1623708002831.png


This left me with 350 extra blues. But.... I don't have any proof that it matters. These RING pack blue/green cells seem very evenly matched in capacity/IR/% capacity left - so evenly distributing them should give you very good results.
 

Doc3G

Member
Joined
Apr 18, 2021
Messages
56
OK cool, I get what y'all are saying. I'm not building two independent systems so I'll have to go with an even mix. Once I go through and recheck the voltages and IR's, I'll tally up all the cells I have and post the numbers and my proposed mixing to see if it looks correct. I actually have 4 groups:

1) 3000-3100 (< 10% of the cells in a 60/40 blue/green mix)
2) 3100-3200 (~ 20% of the cells in a 50/50 blue/green mix)
3) 3200-3300 (~ 55% of the cells in a 70/30 blue/green mix)
4) 3300-3400 (~15% of the cells in an 80/20 blue/green mix)

The above are rough estimates, but overall, the blue LG's tended to not only have more cells but also to have higher capacity on avg, though, of the hundred or so that I've checked the LG's also have slightly higher IR.
 
Top