DIY Electric Car Forums banner

Hacking the Tesla Model S Front Junction Box

26K views 41 replies 12 participants last post by  Anatolii Efimtchik 
There is current measurement after the fets, see the curled up busbar.

Wait a minute, why would the fets control the HV to the 'grey' HV connector and measure current?

So would this mean that the PTC heater (water???) be a dumb module? Ah 2013 module looking at some info the battery heater just needs a pwm into the junction box.

Plus nice isolation devices (Big SMD packages) just like on the bms slaves.
I have parts from a 2015 70D...

1) the PTC Cabin Heater operates from CAN. Like some of the other Tesla modules, it has a Tesla designed control board that drives a main board. I did see some recent hacking by others on controlling the mainboard IGBTs directly. The heater is fairly large and I'm 50/50 I can get it to fit into my airbox. I'm hoping Damien gets a look at it.

2) the Air Conditioning Compressor was also CAN controlled. I also bought a 2014 Tesla compressor which is reportedly PWM controlled, but I haven't done any experimenting with it yet (not an easy thing to test without already being in a system) - probably something better learned from a working 2014 Model S.

3) The coolant heater (battery heater) looks like the not intelligent device. Previous suggestions were that a PWM signal to the Front HVJB controlled the power to the heater. My battery (A123 from a 2014 Chevy Spark) already has an integrated heater (controlled locally within the BMS) so I may just be using the heater for cabin heat (assuming I can't get the PTC heater jammed in).

4) The DCDC also has a direct CAN connection. Interestingly in my DCDC there is no Tesla branded board. I'm hoping that the device will simply report CAN out rather than need CAN to operate (at least beyond a 'fail safe' mode). It's only the list of things to test next. I'm also hoping Damien will tackle the DCDC.

Jeff
 
What are the dimensions of the box?
(length/width/height) and size of the holes for the incoming power cable...
I've got it up on the bench at the moment.

Front HVJB
L : 9 1/4in case (10 3/4 to allow for LV connector and ground connection)
W: 8 1/4in case (9 1/4 to allow for HV plugs to AC,DCDC,etc. and HV in)
H: 2 3/4in case (3 1/4 if the mounting bracket still fitted)

HV hole size
less than 15mm, hard to get actual without disassembly, probably 13-14mm diameter.

Jeff
 
Front HVJB

inputs that I've tagged (but not verified with operation)

1 - blank
2 - blank
3 - blank
4 - OR 12v (orange)
5 - RD PWM in
6 - BL STAT out (PWM)
7 - blank
8 - blank
9 - blank
10 - BK Gnd
11 - YL-BL HVIL
12 - YL-RD HVIL

Note: also a large ground strap connected to the protrusion/nut just to the left of the low voltage connector

Connector : Molex 33472-1201 (I'm pretty sure)

Jeff
 
And answers do trigger new questions:
- Can we do anything with pin 6 (STAT out (PWM))? Curious what kind of signal it will give.
- What are the characteristics of PWM in (pin 5)? Is that just a matter of assuming it's 12V (should be logic in a car) and try different frequencies?
My plan was to start with a 5v PWM signal, probably starting at 1Hz and working up (faster). Since the PWM is coming from a controller, I suspect they'll use 5v signals rather than a 12v. The Tesla pumps used a 1Hz signal - which caught me by surprise as I was expecting something in the 1kh range. I would guess that there is really only a need to send a PWM with a hertz that's twice as fast as the device can respond anyways. There's got to be some design balance between what is easy for the controller to parse (how easy it is to read the PWM signal) and how often it needs to see changes in the PWM.

I'm guessing the STAT will be something like 0-5v as a signal to the controller (outside the Front HVJB) that the heater is actually responding. Since the heater response is likely to significantly lag the controller instructions.

The control scheme would be something like 1) update the PWM, 2) wait for the STAT to get into the expected range. I feel some interesting tests coming on... pumping some coolant through the heater, sending PWM instructions, and then watching STAT and coolant temps to understand the response.

Jeff
 
Top