I'm thinking more about putting the bulk of the cells inside the rear passenger compartment (where the rear seat used to go) and enclosing the front engine (motor) compartment to guard against road spray getting to the battery boxes.
A123 20ah cells, testing the connection method
10
Some of you may have noticed I decided to go with A123's in my 944 conversion. I don't need a lot of range, but I am hoping for quite a bit of power so these cells meet both criteria. It's also hard not to borrow some of Crodrivers findings and he likes these cells so why not. Especially since it appears that he's using them in his Concept_One electric supercar.
![]()
![]()
I'm sure one of his battery modules cost as much or more than my entire conversion including the car, but it is an elegant package and will put my foam insulated boxes to shame.
Visit Rimac Automobili for more info on his amazing creation.
These aren't "new with warranty / cost a small fortune cells" I took a gamble on some grey market cells and these appear to be cells that were rejected and destined for recycling. All of the ones I've tested have good IR (1 - 1.5mohm), but only come in at 18 - 18.25Ah. Because of this the cells were castrated... well had their tabs removed.
There is plenty of tab material still covered by the plastic casing, it's just a pain in the ass to uncover and make useful without going to far and puncturing the cell. Below is a cell that is "half" done, the casing material has been removed but the remaining adhesive needs to be scraped off to reveal the aluminium and copper tabs.
![]()
With some of Yabert's help on connection method ideas I settled on a simple sandwich type connection with block - tab - block - repeat. There are 4 aluminium blocks for each end terminal and 7 blocks for all of the main series connections.
![]()
As you can see from Yabert's rendering above that there will be a lot of different blocks to connect these cells. Initially he had a line on a company that would laser cut the green parts for a decent price, but I would still have to make all of the red and blue blocks, then due to limitations on available material thickness to be laser cut it just makes more sense to make ALL of the pieces myself. Below are a few pieces from the first batch of parts used for assembly.
![]()
Each block goes through many steps before it's done, first the material is cut to length, then marker holes are drilled using the CNC router, then the marker holes are fully drilled on the drill press, along with some parts that get a countersink, some threaded etc. Each 16S3P module has 65 of the "green" blocks above, 15 "blue" blocks, 16 "red" blocks and 2 large terminals shown below.
That's a total of 98 blocks per module or 588 blocks for the 6 modules in the final design. The pink foam spacer in the center of the pack is just what I had on hand, I have 0.06" plastic sheet that will be cut to size for the actual pack assembly. During assembly each side of each cell tab is cleaned with a stainless steel wire brush, along with the aluminium blocks and treated with No-Alox, this is another time consuming process but the initial results are worth it. The battery shown below is only 2S3P for testing, I will be assembling them into 16S3P modules for the car.
![]()
Above charging with my dc/dc converter charger, it's simple and reliable but 3 parallel cells are too much for this method without some current limiting. During the charging process the dc/dc converters overheated and shut down, they do turn back on automatically when they cool off though, but I'm sure doing this too often would shorten their life span. The cells charged up perfectly to 3.5v per cell though. (Please ignore the lack of proper terminals, I have terminals for the high current connections, but I neglected to buy small terminals for this purpose)
![]()
The cell tabs first need to be uncovered, this is time consuming surgery and must be done carefully or risk damaging the cell. Then two holes are drilled in each tab to allow the screws to pass through that hold the blocks together. I created a jig on the CNC router out of MDF that allows me to align and clamp each cell and drill the 4 holes with a built in guides.
![]()
![]()
The jig is simple and effective allowing me to quickly and precisely drill the 4 holes I need. If I were to do it again I would probably find/make metal sleeves to line the pilot holes to keep the holes tight with repeated drilling.
![]()
I wanted to do some high current discharge testing to make sure that the connections are sound and don't heat up at high currents. I was only able to test this pack to about 700A but all of the terminals and the cells themselves were cold and the pack only sagged to about 3.1xV per cell (measured ~6.25V under load after 5 or 6 seconds when stuff started to melt and smoke). I can't say the same for the load, I wasn't exactly setup for 700A discharges and melted the insulation off some 4 awg wire and made my resistor bank hotter than I ever have before. OOOPS! The cells perform exceptionally well though in all tests I've done to date.
My final comment is it's a good thing these cells perform otherwise all of this work to prepare and connect them just wouldn't be worth it. I can't wait for the day that a large prismatic has the same or better performance to these A123 pouches but already has 2 big screw terminals at the top.
Okay, ONE more comment, I'm sure you are wondering "what did these cells cost!?!?!?! Around $1.25 per Ah assuming 18Ah per cell.
I'm sure one of his battery modules cost as much or more than my entire conversion including the car, but it is an elegant package and will put my foam insulated boxes to shame.
Visit Rimac Automobili for more info on his amazing creation.
These aren't "new with warranty / cost a small fortune cells" I took a gamble on some grey market cells and these appear to be cells that were rejected and destined for recycling. All of the ones I've tested have good IR (1 - 1.5mohm), but only come in at 18 - 18.25Ah. Because of this the cells were castrated... well had their tabs removed.
There is plenty of tab material still covered by the plastic casing, it's just a pain in the ass to uncover and make useful without going to far and puncturing the cell. Below is a cell that is "half" done, the casing material has been removed but the remaining adhesive needs to be scraped off to reveal the aluminium and copper tabs.
With some of Yabert's help on connection method ideas I settled on a simple sandwich type connection with block - tab - block - repeat. There are 4 aluminium blocks for each end terminal and 7 blocks for all of the main series connections.
As you can see from Yabert's rendering above that there will be a lot of different blocks to connect these cells. Initially he had a line on a company that would laser cut the green parts for a decent price, but I would still have to make all of the red and blue blocks, then due to limitations on available material thickness to be laser cut it just makes more sense to make ALL of the pieces myself. Below are a few pieces from the first batch of parts used for assembly.
Each block goes through many steps before it's done, first the material is cut to length, then marker holes are drilled using the CNC router, then the marker holes are fully drilled on the drill press, along with some parts that get a countersink, some threaded etc. Each 16S3P module has 65 of the "green" blocks above, 15 "blue" blocks, 16 "red" blocks and 2 large terminals shown below.
That's a total of 98 blocks per module or 588 blocks for the 6 modules in the final design. The pink foam spacer in the center of the pack is just what I had on hand, I have 0.06" plastic sheet that will be cut to size for the actual pack assembly. During assembly each side of each cell tab is cleaned with a stainless steel wire brush, along with the aluminium blocks and treated with No-Alox, this is another time consuming process but the initial results are worth it. The battery shown below is only 2S3P for testing, I will be assembling them into 16S3P modules for the car.
Above charging with my dc/dc converter charger, it's simple and reliable but 3 parallel cells are too much for this method without some current limiting. During the charging process the dc/dc converters overheated and shut down, they do turn back on automatically when they cool off though, but I'm sure doing this too often would shorten their life span. The cells charged up perfectly to 3.5v per cell though. (Please ignore the lack of proper terminals, I have terminals for the high current connections, but I neglected to buy small terminals for this purpose)
The cell tabs first need to be uncovered, this is time consuming surgery and must be done carefully or risk damaging the cell. Then two holes are drilled in each tab to allow the screws to pass through that hold the blocks together. I created a jig on the CNC router out of MDF that allows me to align and clamp each cell and drill the 4 holes with a built in guides.
The jig is simple and effective allowing me to quickly and precisely drill the 4 holes I need. If I were to do it again I would probably find/make metal sleeves to line the pilot holes to keep the holes tight with repeated drilling.
I wanted to do some high current discharge testing to make sure that the connections are sound and don't heat up at high currents. I was only able to test this pack to about 700A but all of the terminals and the cells themselves were cold and the pack only sagged to about 3.1xV per cell (measured ~6.25V under load after 5 or 6 seconds when stuff started to melt and smoke). I can't say the same for the load, I wasn't exactly setup for 700A discharges and melted the insulation off some 4 awg wire and made my resistor bank hotter than I ever have before. OOOPS! The cells perform exceptionally well though in all tests I've done to date.
My final comment is it's a good thing these cells perform otherwise all of this work to prepare and connect them just wouldn't be worth it. I can't wait for the day that a large prismatic has the same or better performance to these A123 pouches but already has 2 big screw terminals at the top.
Okay, ONE more comment, I'm sure you are wondering "what did these cells cost!?!?!?! Around $1.25 per Ah assuming 18Ah per cell.
41 - 49 of 49 Posts
After months of testing I found that my old connection method resulted in higher than expected connection resistance (most likely due to the large number of connection blocks)
So to reduce the number of parts required and try to reduce the connection resistance of the pack, groups of 3 tabs are sandwiched together. For a 3P series connection there are now 3 blocks instead of 7, and the end terminals use two blocks instead of 4. Now there is only a single block between each series connection instead of 5. This is a drastic decrease in the parts I need to make, as well as the complexity of trying to hold and align 6 tabs, 7 bocks and 2 screws. I also moved from M4 to M5 screws and from a flat head (countersunk) to a normal cap screw. I now use a 4mm allen key vs a 2.5mm to tighten the screws which allows much higher fastening torque to be applied. I also incorporated Nord-lock washers into the design.
I still have to revise the end terminal design slightly for proper clearance as well as the extreme end terminals that will connect to the 2/0 ring terminals. When testing with the PL6, the IR of the cells is much closer to the expected value of 3 parallel cells. I can also tighten all of the connections much better and the Nord-locks seem to be doing exactly what they are designed for.
I believe I now have the build it and forget it pack that I was hoping for initially. I will have cell level monitoring using my Cell Log 8 Breakout Modules, but I believe any worry of loose or bad connections over time will be virtually eliminated.
So to reduce the number of parts required and try to reduce the connection resistance of the pack, groups of 3 tabs are sandwiched together. For a 3P series connection there are now 3 blocks instead of 7, and the end terminals use two blocks instead of 4. Now there is only a single block between each series connection instead of 5. This is a drastic decrease in the parts I need to make, as well as the complexity of trying to hold and align 6 tabs, 7 bocks and 2 screws. I also moved from M4 to M5 screws and from a flat head (countersunk) to a normal cap screw. I now use a 4mm allen key vs a 2.5mm to tighten the screws which allows much higher fastening torque to be applied. I also incorporated Nord-lock washers into the design.
I still have to revise the end terminal design slightly for proper clearance as well as the extreme end terminals that will connect to the 2/0 ring terminals. When testing with the PL6, the IR of the cells is much closer to the expected value of 3 parallel cells. I can also tighten all of the connections much better and the Nord-locks seem to be doing exactly what they are designed for.
I believe I now have the build it and forget it pack that I was hoping for initially. I will have cell level monitoring using my Cell Log 8 Breakout Modules, but I believe any worry of loose or bad connections over time will be virtually eliminated.
Attachments
Choice, availability, and personal preference. If I switched to inch sizes it probably would have been easier but I've been able to maintain almost entirely metric fasteners in the car. By switching to this style of fastener it allows me to use the Nord-locks so I don't believe it will be necessary to check torque.
I also like that approach, but the 944 has a rear transmission so my battery box is designed to fit around it. I think a well designed pack using that method has potential though.
Attachments
Any updates on the build????
Related to the connection method no. The new method simply works well, requires a minimum number of parts and is simple to assemble.Any updates on the build????
For the car I put some of the updates in my build thread, but I usually capture everything on my blog, link is in my signature.
41 - 49 of 49 Posts
