DIY Electric Car Forums banner
1 - 20 of 37 Posts

· Registered
Joined
·
41 Posts
Discussion Starter · #1 ·
Hi,

we are building a Porsche 914 conversion, using battery modules from a Porsche Taycan.
We use the Hyper 9 motor 130V version, so our HV is limited to 125V. That is 5 Taycan Modules in series (one module has 25V max).
To get more capacity we want to parallel 3 of these battery packs, which would be 5 Taycan modules in series and 3 in parallel.
One battery pack will be in the front of the car, the other two in the back.
So there is a problem with the BMS. We used the Orion 2 BMS system in our former builds but it's not built for parallel wiring as the manual says. So I would need 3 seperate BMSs - one for every pack. This would be quite expensive (~3,500€)

So I had a look at the Dilithium BMS, which will probably be the better solution. But there is one question regarding the configuration.
I attached 2 configurations that make sense to me. The first one will be the safest I guess, the second one would safe me one BMS unit.

In the first version I use one controller and one satellite unit for every pack.
Rectangle Font Slope Schematic Parallel

In the second version I use one controller and one satellite unit for the front pack, and one controller and one satellite unit for the back pack, which is directly paralleled.
Rectangle Font Parallel Schematic Slope

I don't know if the second version is possible, because there only the modules are paralleled, but not the individual cells in the modules.

What would be the best option to solve this?

Thanks
Alex
 

Attachments

· Registered
Joined
·
346 Posts
30s packs will make things a little tough, since each module for the dilithium system can only support 24 cells. I suspect this means you will need 2 controllers. In your second diagram, you would need 2 more satellites to monitor the the other modules. The satellites save a little money over having another controller, but it is still going to be a pretty expensive BMS. Unless there is a way to physically connect the cells in the rear packs, you will need to monitor them separately. I have a dilithium system that monitors a parallel pack, although mine is 42S and 2P. I use one BMSC and 3 satellites.

Have you considered a higher voltage Hyper 9, and trying to get down to 2 strings? 42s 2p would only use 14 of your modules, and you would have to find room for 2 more up front. I am guessing that the 914 is a bit limited on space?
 

· Registered
Joined
·
41 Posts
Discussion Starter · #3 ·
Actually all parts are here already and designed for the 130V system. So we have to stick with that.
We plan to build another one but with other components and raise the voltage to use CCS charging.

So version 2 is possible beside the fact I would need 2 more satellites? I wasn't sure about paralleling the modules but not the individual cells.

Also would it be possible to use the 2nd satellites LTC2 for the 3rd pack? So there is the BMS controller 2 and BMS satellite 2 for the 2nd pack but on the satellite 2 only LTC1 is used.
So if I could use LTC2 of the satellite for the 3rd pack, I would only need one more satellite for the rest of the cells.

Luckily the modules are quite handy so space is not the biggest problem compared to a 356 :D
 

· Registered
Joined
·
346 Posts
Also would it be possible to use the 2nd satellites LTC2 for the 3rd pack?
Yes, each LTC is completely independent, and you can wire them in any crazy order you like, as I recall (it lets you remap them into strings).

I think there is a good case to be made for using the OEM boards if there is a way to interface with them, but it was not a subject I did much research on.

I am not following your build, but I am not really following the talk about cutting up the modules with a dremel? Are there not cell level taps at the surface of the module?
 

· Registered
Joined
·
41 Posts
Discussion Starter · #11 ·
Yes, each LTC is completely independent, and you can wire them in any crazy order you like, as I recall (it lets you remap them into strings).

I think there is a good case to be made for using the OEM boards if there is a way to interface with them, but it was not a subject I did much research on.

I am not following your build, but I am not really following the talk about cutting up the modules with a dremel? Are there not cell level taps at the surface of the module?
Unlucklily there are no cell taps at the surface, there is only a 4 pin connector for the OEM boards communication.
 

· Registered
Joined
·
8,631 Posts
Is there some reason for using the Taycan modules? As Porsche parts, they seem likely to be the most expensive salvage modules in existence. They are roughly the physical size (and perhaps exactly the same) as modules from various models from multiple manufacturers. As part of a system running roughly double the voltage of most EVs, they are configured in exactly the opposite way to what would be suitable for a low-voltage system. They have 12 LG Chem pouch cells of about 60 Ah capacity like several other available modules (65 Ah in this case), and has them in a 6S2P arrangement (so about 22 V and 126 Ah nominally).

In the first post of the OpenInverter discussion these are described as 'Alu LG Bricks'; these are the VDA 355 format... the only standardized EV battery module format used in production EVs. Modules of this format are produced in 6S2P, 4S3P, 3S4P, by LG Chem and CALB, and sold as OEM components and aftermarket products.

You could avoid all of the parallel connection hassles by using the appropriate number of modules which use the same cells but have more in parallel and so fewer in series, and simply connecting the modules in series. Even two strings instead of three would make everything easier.

15 of the 33 modules from a 93 kWh Taycan pack or 15 or the 2 from a 79 kWh Taycan pack (the two battery sizes just use different numbers of the same modules) would provide about 42 kWh (nominal, less than that usable). There must be easier ways to build this capacity at the desired voltage, and it would be even easier with the higher-voltage version of the HyPer 9.

I realize that people are often working with something which is available to them, and restricted to that hardware, but often people choose components and twist the design to fit those components even if they don't really make sense. Even when the parts are already in hand, they can be sold and replaced by more suitable parts.
 

· Registered
Joined
·
157 Posts
This is why I keep looking at my battery options and coming back to the iPace/eTron modules. They're the LG VDA355s in 3s4p for 11v nominal. 14 in series just fits for the Hyper9 HV for 36.4kwh. I'm glad I didn't buy batteries first in this conversion, it would have locked me into all kinds of decisions.
 

· Registered
Joined
·
41 Posts
Discussion Starter · #14 ·
This is why I keep looking at my battery options and coming back to the iPace/eTron modules. They're the LG VDA355s in 3s4p for 11v nominal. 14 in series just fits for the Hyper9 HV for 36.4kwh. I'm glad I didn't buy batteries first in this conversion, it would have locked me into all kinds of decisions.
Yeah, would have been better of course. On the other hand we got these batteries reeeally cheap.
 

· Registered
Joined
·
5 Posts
Agreed, series is best if possible! I have installed and serviced a number of BMS brands including Dilithium and Orion. A module is a string, and there are tricky pitfalls with managing strings of any size in parallel. The method of paralleling strings together to increase current seems sensible at first. As does maintaining balance by paralleling cells or bricks. But paralleling the cells or bricks of separate strings is only acceptable if the parallel connections can handle full pack current. Heres why.
The notion of using small wires to parallel bricks of two strings for balancing fails when the bricks begin to behave unequally. And they all do sooner or later. What will your BMS do when this unequal load causes a few too many balancing amps and one or more small parallel wires fail? Most likely this will happen when brick potentials vary the most, such as finish charging, or hard accelleration at low SOC. Assuming it happens, having BMS cell taps on all individual bricks is smart. Yes it will detect an out of spec condition, and yes it should stop a charger, regen, or discharge. But it won't stop a higher potential string from continuing to overcharge an unintended good brick or two in a lower potential string with a higher resistance brick. Do you smell something burning?
The OrionBMS2 does have a solution that can individually authorize up to 8 separate strings onto a DC BUS. I'm building my second vehicle using this feature of OrionBMS to manage two separate strings of 5 TESLA S modules in series. That's two strings of 30S. The BMS individually parallels each string to a common DC BUS if it is safe to do so. It requires no paralleled bricks, and allows bricks to age differently.
If you must parallel strings, there is more to learn about what Orion can do, but this document is a good place to start, even if building with another BMS.
Good luck with your build!
 

· Registered
Joined
·
3 Posts
Hi,

we are building a Porsche 914 conversion, using battery modules from a Porsche Taycan.
We use the Hyper 9 motor 130V version, so our HV is limited to 125V. That is 5 Taycan Modules in series (one module has 25V max).
To get more capacity we want to parallel 3 of these battery packs, which would be 5 Taycan modules in series and 3 in parallel.
One battery pack will be in the front of the car, the other two in the back.
So there is a problem with the BMS. We used the Orion 2 BMS system in our former builds but it's not built for parallel wiring as the manual says. So I would need 3 seperate BMSs - one for every pack. This would be quite expensive (~3,500€)

So I had a look at the Dilithium BMS, which will probably be the better solution. But there is one question regarding the configuration.
I attached 2 configurations that make sense to me. The first one will be the safest I guess, the second one would safe me one BMS unit.

In the first version I use one controller and one satellite unit for every pack.
View attachment 127442
In the second version I use one controller and one satellite unit for the front pack, and one controller and one satellite unit for the back pack, which is directly paralleled.
View attachment 127443
I don't know if the second version is possible, because there only the modules are paralleled, but not the individual cells in the modules.

What would be the best option to solve this?

Thanks
Alex
Curious, where did you source the batteries?
 

· Registered
Joined
·
62 Posts
I'm also very interested in where these Taycan modules were sourced as they sound like excellent candidates for high power (tesla motor) applications. Most OEM EVs now are designed around a 300-400v setup and wire the full pack in series, so in order to get that voltage you need a full pack (tough to fit in a lot of conversions) . But since the Taycan runs at 800V that opens the possibility of getting to 400v with only half an OEM pack. At 21.9v nominal and 129Ah per module, 16 in series gets you to 350v and 45kWh (with a 5C rate of 645A) which is sounds ideal for a small tesla drive unit.

Also interested in module weight and dimensions.
 
1 - 20 of 37 Posts
Top