The car on my driveway as arrived
Here is close up
photo of one of the claws open. The battery can be unsnapped in about 15-20
seconds with this ingenious mechanism. All I had to do is to crank the actuator
(worm drive) head, seen on the photo above this one.
Guiding
alignment pins welded to the the machined aluminum battery tray. Entire tray has
grooves and battery modules attachment features CNC milled in it. Quite a work
of art. Massive steel claw on top is also visible.
Close up of the
battery tray. The alignment guide and the bracket with retaining pin.
The battery is
shaped after the standard cavity in Nissan's underbody.
The battery consists
of 8 modules manufactured by A123 Systems. There are 3 module sizes used in this
vehicle, but its clever packaging allows to stack any number of cells using
standard heat sink plates, plastic covers and the BMS components. The long
modules have 3P16S arrangement, the medium ones are 3P11S and the short ones are
3P6S.
The cell type A123
uses here is LiIon AMP20M1HD-A pouch cells produced in South Korea. Nom Voltage is 3.3V, nom. capacity
19.5Ah. A I mentioned, there are three sizes of battery modules used in Qashqai
- 3P16S (48 cells, 3 modules), 3P11S (33 cells, 2 modules) and 3P6S (18 cells 3 modules).
Each single cell is 19.5Ah capacity. Quick calculation on the back of a napkin reveals that
this entire battery then has got total of 3*3*16+2*3*11+3*3*6=264 cells. From
now on, unless specifically mentioned, a "cell" will refer to electrically
single 3.3V (which is 3 paralleled physical cells).With 3P
arrangement in all modules this means 264/3=88 cells in series. At 3.3V nominal
voltage that's 290.4V pack and at 19.5*3=58.5 Ah nominal capacity it's got 290.4
V * 58.5 Ah = 16,988.4 Wh
energy contents. That's about 71% of Nissan Leaf's battery. Assuming same
240Wh/mile consumption as the Leaf claims (at 35mph speeds), that translates to
70.8 miles range for Qashqai, or ~56 miles at freeway speeds. Not spectacular, but keep in
mind that the battery is designed to be swappable and perhaps it was calculated that customer
can always get to the nearest swapping station with this kind of range. Anyway,
this was interesting to figure out, but it was not the point of disassembly...
The power cables snake around the tray and connect the modules in series. The pack has emergency disconnect in about electrical middle, so the cabling it appears more complicated than plain series connection of all the modules. The BMS network (black looms) is routed separately from power cables to reduce interference. Naturally, all the comms are implemented via CAN.
After all the modules were taken off the tray, each
was perked up with external power supply to about 60% SOC for long storage. It
is unlikely I can get hold of more of these modules, so it is not enough for my
conversion progress in progress, and I'm not sure I'd se this type of battery
anyway but these packs may be useful for the BMS tests and as something as a
range extender trailer. The BMS PCBs would be replaced with my own, but it is
hardly worth the effort to design them from scratch to fit in place of stock
PCBs for just one off battery system. For now will wait and see what kind of use
these modules can be put in.
e
The main BMS controller (more on this - below).
The slave PCB of the BMS installed on the side of a
short battery module. Each slave PCB serves up to 16 cells. If it is installed
on the module with 8 or fewer cells in series, only half of the PCB is populated
with components as you cans see on this photo.
Battery module slave BMS board. As expected is conformal coated
with silicone moisture resistant compound which remains flexible. It also has
good heat conducting properties. It may look like epoxy on the photo, but it is
easily peeled and can be scarped off of components and PCB. Note that clear compound covers the circuit but leaves
uncovered circle around mounting screw. They way it is deposited is using a CNC
micro-spray station which is being part of SMT assembly.
Almost all the ICs were identified except one obscure part number - Google search did not return anything close to making sense. However, based on installed parts, processor, memory, opto couplers, CAN transceivers and discrete circuitry interfacing to the cells, the architecture of the BMS became apparent. I was especially delighted to find out that the processor IC (dsPIC30F5011) on the PCB A123 engineers chose was exact part number I have used in one of my BMS designs for a customer at about the same time (circa 2009). This just validated my sense of judgment regarding balance of optimal power and simplicity.
Overall appearance of the main BMS module
The main PCB open. As with slave modules, the type of
semiconductors tells me how this BMS is designed. Also, tracing few interesting
components revealed a few tricks engineers used in their circuit. Granted,
critical components, connectors etc. were identified. You can see engineering
changes (blue wiring) around the chip at the bottom right corner of the PCB.
Zoomed in
photo of the main CPU and ICSP header. Of course the code is protected and not
readable, but even if it was available, there would be little practical use for
it for me - the software is very specific to this battery type and other vehicle components. I found more valuable studying the hardware design
and layout, identifying
components and seeing how interfaces, I/O protection, power supply, isolation and
other seemingly non-essential circuits are implemented. These circuits are critical
for robust and reliable BMS operation.
Next links point to individual pages describing actual reverse engineering work. Warning - this information, terminology and professional lingo are very specific and if hardware electronics and CAD tools are outside of your comfort zone, you will quickly get bored or get lost.