[WIP] L322 RR Vogue - The beached whale

[nkiernan]

If you have the time it is highly recommended to be able to close this topic and have definitive results.

Yes, I want to be sure there is nothing that will cause harm. Looks straightforward, but is there any issue running this type of tester on modules and full pack while the BMS modules are connected (not going to harm anything delicate there?) or the isolation monitor in SBox is connected (distorted results)?

[tom91]

Yes, I want to be sure there is nothing that will cause harm. Looks straightforward, but is there any issue running this type of tester on modules and full pack while the BMS modules are connected (not going to harm anything delicate there?) or the isolation monitor in SBox is connected (distorted results)?

Should be fine as the designs should be 2.5kV

The test should be carried out with the system powered down. Best to do component tests then assembly tests to make sure nothing changed.

[nkiernan]

Got hold of an insulation tester to carry out more investigation on the battery pack. Removed all the modules and visual inspection of each module and the pack case looked good. There is an aluminium pad on the underside of each module for heat transfer, so recorded a number of checks on each module in the table below. Modules all seem good

Continued with a range of tests on various elements of the system, finishing up with the full system reconnected. I'm interpreting all these as expected results with no insulation concerns! The SBox HV- tests show lowest results and possibly that's to do with its internal insulation test components?

With everything back connected, I powered up the system seeing 332.7 VDC on the busbars in the HVJB as expected. With the Fluke 115 meter, seeing 4 VDC between HV- busbar and HVJB case and this continues dropping (see video link below). HV+ busbar to HVJB case shows approx. 170 VDC when first connecting the meter and this drops to about 70 VDC (see video link below), but slowly continues to drop to around 63 VDC. It looks to keep dropping but very slowly! Resistance check from each busbar to HVJB case is OL.

This still looks a big concern!?

Short videos showing voltage drop with HV- to HVJB case:
https://youtu.be/3vEcLrav8vk?si=bpw3jZmAtoDWjPpG

Short videos showing voltage drop with HV+ to HVJB case:
https://youtu.be/lSIfJ7toWF0?si=Qxx_X8fEIP58W86Q

[nkiernan]

What am I dealing with!

Appreciate any extra eyes on this issue to confirm if its a real problem or can be explained To me, there should be zero voltage measured in the few tests below.

Well I'm back looking at this X5 battery pack and the concern there's some sort of HV insulation issue. In the previous post, I worked through a range of Megohm tests on the individual modules and variuous components. All seem's ok in those tests.

With the modules back in the original battery case, I have the 12 modules connected for full voltage, but disconnected from the SBox, so the modules and HV should be isolated from the battery case (no internal insulation test routes etc). I read open loop between the HV- and HV+ battery terminals and the battery case as expected, but I still see a DC voltage between each terminal and the case as can be seen in the video below (found a second meter to use on these tests and it seems to show a lower set of readings, original meter issues maybe):



Connecting the full system and activating full HV to the HVJB, I am still reading a DC voltage between the HV+ busbar and the HVJB enclosure:



Some tests run with SBox, BMS loom, and case ground connection disconnected to ensure isolation between battery module voltage and battery case

I split the pack into 4 groups of three modules and checked each group HV- and HV+ terminals to battery case (all disconnected from SBox), and each group can show up to 2.9V DC that drops over a 5 seconds or so to 0.2ish VDC. So they all seem to behave the same, not flagging any one module as a problem (surely should be zero, or is this expected with meters allowing internal test path for voltage!?)

So I'm questioning if there really is a problem, if its the checks between HV- and HV+ to the case that charges the case between checks and throws the readings off, if its the nuances of the meters and internals, but I expect zero voltage in these tests!

[m.art.y]

I have seen that same issue on various battery packs. There is a video on youtube where Arlin Samsone sees the same thing. I don't know how to explain this but this might be some sort of ghost voltage that multimeter picks up. If you insulate the modules from the casing with additional rubber pads do you still see those voltages?

[tom91]

So I'm questioning if there really is a problem

No problem here, the voltage drop to acceptable levels. As mentioned before this is due to having a high charge in one part and then an airgap/insulator and then metal. This sort of mechanical/electrochemical setup will act as a capacitor and store a small charge.

The multimeter has a high resistance on its measurement inputs thus will dissipate this charge when you complete the circuit when measuring. Note how the rate of drop slows with the decrease in voltage, this is Ohms law at work.

[nkiernan]

No problem here

Phewww, this is what I was hoping to hear, thanks for the reply and explanation Tom

[nkiernan]

Charge Port:

Tidying up loose ends, time to finalise a location for the charge port socket. Space was tight behind the fuel filler cap, so mounting behind the front grill leaves it close to the HVJB

[nkiernan]

Charge Port:


Which means the grille needs some attention! This grille shape and pattern don't make it easy, so just started cutting!

Plan is to have a 3D printed hinge detail with 0 and 90 degree stops and accommodate an over-center spring option to hold it open or closed as needed. Some reshaped prototypes later and design works well

Parts to be bonded behind grille

Will tidy some cut edges and possbily repaint grille at some stage but good enough for now

The grille pattern dictates the cut lines, which dictates the hinge orientation and doesn't give the most ideal opening but still enough space

[Marat]

Very good solution, I need to fix it myself.

[nkiernan]

Range Rover firmware module

Have been looking at the vehicle integration again. CAN logs were limited from before starting the conversion as the crankshaft had snapped before I got the car, but got enough to help control the main dash functions. Using one of Damien's older GS450H VCU's as a CAN bridge/MITM, RPM, gear selection and key dash warning lights can be controlled. The bridge was also to fix the clashing X5/SimpBMS and SBox CAN ID's, but I'll be removing the SBox in favour of the ISA shunt and Gigavacs so won't need the bridge.

Then some fun and learning to see if this could be translated to the Zombieverter firmware! Looking through the existing vehicle modules, Damien's BMW E39 module was close and so became the foundation of a JLR_L322 module for this conversion. Kelju's tread here: viewtopic.php?t=5911 was very helpful and now a test branch is compiling and looks to be working well. This car won't be a popular choice for conversions, so don't know if this module will make it into an official ZV release

[marcexec]

How's the project going mate? Dublin hackerspace summer meeting and Dublin Maker are up in August

[nkiernan]

How's the project going mate? Dublin hackerspace summer meeting and Dublin Maker are up in August

I've been off on other projects, but should get the fully welded battery box back this week! And as with many projects, I'd go about these a little different the next time! Unfortunately though, I'm out of the coutry in August so will miss the event

[marcexec]

Neat, I love the braketry (Project Binky reference)!

[nkiernan]

Battery boxin':


Got the aluminium welded battery box back and working on a mounting detail that makes use of the original battery box mounting frame. Keeping the original transfer case and the rear prop over complicate things a little!

Works out well reusing the original X5 pack HV connector and swapping the S-Box for more capable contactors. New low voltage connector located and original BMS master ready for use with SimpBMS

Cradle will be fixed to the original S-Box and BMS mounting plate

Ready for some testing with the latest ZV release

[nkiernan]

I'll be reorginising the order of the modules compared to the original X5 pack layout, which means the original BMS slave connectors don't reach or are in different order.

Anyone know if the BMS loom needs to retain the order of modules or if its ok to connect them to the closest modules regardles of order!?

[m.art.y]

Works out well reusing the original X5 pack HV connector and swapping the S-Box for more capable contactors.

Without S-Box you will be missing the shunt. While you can use simpBMS voltage reading you will still need at least a current sensor. You will need that for fast charging and to enforce any current limits in ZV. In the X5 S-box contactors are somewhat uprated but I was thinking if it could be possible to use the S-Box still but connect other contactors to it if need be.
The battery module order should not matter, as I think the address is assigned automatically by the position within the chain but I'm not 100% sure.

[nkiernan]

You will need that for fast charging and to enforce any current limits in ZV.

Yes, I am using an ISA shunt alright. Its located in the HVJB