Thứ Tư, 23 tháng 11, 2016

Pics/Info: Inside the battery pack part 8

  • Oct 13, 2015
    wk057
    Yeah, the block on my PM access is getting a bit annoying. It'll definitely delay me being able to finish my contest since I wont be able to get the winners' addresses for prizes securely.

    Anyway.

    As of now I don't think I need another battery pack personally. And if the pack is in Norway that probably doesn't do a whole lot of good for me. Shipping would probably be expensive. I'm sure someone in Norway would be interested in a battery system though given how electric rates work out that way.

    Good luck.

    - - - Updated - - -

    The pyro fuse is inside the HVJB isn't it? I can't remember exactly off hand. It does effectively cut the HV systems though, since it cuts the HVIL.
    �
  • Oct 13, 2015
    Spacela
    I thought it was in the vicinity of the 12v battery. Could be wrong.
    �
  • Oct 13, 2015
    stopcrazypp
    You guys might be talking about different pyro fuses. The conventional "pyro fuse" might be there, but I think jeffro01 is talking about the updated fuse for the pack that boosts the current limit to 1500A instead of 1300A.
    �
  • Oct 13, 2015
    Ingineer
    The Pyro fuse is located on the DC primary junction board (large fuses) which is located under the air intake plenum in the frunk. (Under the cabin air filter housing)

    Depending on your configuration the 12v battery might be under this board, or under the frunk bucket.

    You can follow the wire from the frunk emergency responder cut loop to locate it.

    - - - Updated - - -

    That new "electronic" fuse is supposedly located in the pack. I haven't seen one, so it's unknown how that operates. I assume that it's hooked up to the contactors such that if the electronic circuit "trips" it opens the contactors. Though I suppose it could be a pyrotechnic device. I wouldn't think it would be safe to break a HVDC line without a sizeable gap or in the presence of inert gas or hydrogen like some of the contactors use (yes, Hydrogen!)
    �
  • Oct 13, 2015
    stopcrazypp
    Tesla describes the new fuse like this: "It constantly monitors current at the millisecond level and is pyro-actuated to cut power with extreme precision and certainty."
    http://www.teslamotors.com/blog/three-dog-day

    That is why some people call it a "pyro fuse" although that terminology is already in use for a different fuse.
    �
  • Oct 13, 2015
    Ingineer
  • Oct 13, 2015
    offpist


    Anyways.
    I am thinking 15 000 dollar for the battery, then I guess about 1000 dollar for the shipping to New Jersey or other east cost town.
    Shipping big stuff to and from USA with boat is not a problem at all.

    I regularly import stuff from USA, like motorbikes, pallets of solar panels etc.
    �
  • Nov 1, 2015
    Murmele
    Will the voltage measured at every cell, or only on cellpacks? How the balancing is realized?
    �
  • Nov 1, 2015
    apacheguy
    I believe voltages can be measured for each string of 74 cells. Balancing occurs during and after constant voltage charging (>93%).
    �
  • Nov 1, 2015
    kennybobby
    i doubt there is any balancing done during charging since the maximum balancing current is only ~100mA. Based upon the pcb circuits i think balancing could be commanded at any time and at any voltage, as long as there is no charging or discharging loads on the pack.
    �
  • Nov 1, 2015
    wk057
    In testing with a salvage vehicle and pack, the balancing circuits would never enable until the car was at the constant-voltage stage of the charge cycle, around 93+%. Afterwards they would stay on regardless of charge/discharge for some set period of time calculated to bring that particular cell group down to where it needed to be.
    �
  • Nov 3, 2015
    Oba
    Great data point, thanks.
    �
  • Nov 5, 2015
    Quantum`
    Hello wk, I've been following your various battery adventures with gratitude.

    I'm only up to page 16 here so stop me if you already know this, but some observations on your BMS PCB.

    On your front picture there's a quad-pack IC, and the number on it is legible as 76PL536AQ1. TI is one of the manufacturers of these: BQ76PL536A | Battery Monitor, Protection & Authentication Solutions | Battery Management Products | Description & parametrics
    The datasheet is here: http://www.ti.com/lit/gpn/bq76pl536a

    It does perform battery management functions, including balancing. And here's the good news if you might someday consider re-rendering your (expensive) solar install: "The bq76PL536A is intended to be used with a host controller to maximize the functionality of the battery management system. However, the protection functions do not require a host controller."

    The smallest quad-pack on the front is a microcontroller: Mouser Electronics - Microcontroller

    ... and I can't get a number off the small TSOP8 but I'll bet it's an EEPROM for the microcontroller's software, given the PROG pin nearby. No idea what the microcontroller might do though; the big chip seems to do most.

    I notice also on the board the TSOP16 chip has numbers, but I just can't read them from the picture. This is a mystery chip.

    On the back side of the PCB you'll notice that the heavy-duty IC has actually had its numbers sanded off. This is probably a just set of power MOSFETS to shut off power to one set of cells or t'other, but whatever it is, it's controlled by the mystery TSOP16. Looks like it's switching heavy currents in planes internal to the PCB, probably charge current.

    Also on the back there is a connector for 'cells 0-6'. This is going to go to 7 stacks of multiple cells, I'm betting 4 pins per stack, for monitoring and balancing.

    Then there's the smaller connector for 'TS1-2', + and -. Referring to the main chip's datasheet, these are 2 differential temperature sensor inputs. (probably just thermistors toward one end and the other of the module)

    So these BMS boards are daisy-chained together and are eventually connected to the main controller, which I think you'll found as the odd one. The thing that's puzzling though is how they communicate, unless it's somehow connecting to J2. But according to the datasheet there need to be 13 comm lines (SCLK, SDO, SDI, CS_N, CS_S, DRDY_N, DRDY_S, FAULT_N, FAULT_S, ALERT_N, ALERT_S, GND, V), and you need both north and south connected if you're going to daisy-chain. We're 3 pins short. I gotta infer that the 12V system of the car is all that's needed to power the BMS', or else how would they work when the pack is flat?

    Hang on a sec... I notice that the microcontroller's P0.2 and P0.3 pins (general I/O pins) go to the big chip's CS_H and SDI_H. So the microcontroller is acting as the big chip's host. I'll bet it connects to the other vital functions too. Probably each BMS PCB has its own address in the EEPROM.

    Tracing each PCB land may answer more questions. Maybe you're already way beyond me already, but I'll catch up this weekend.
    �
  • Nov 5, 2015
    Ingineer
    Great analysis, Quantum!

    The microcontroller is probably used as a comms gateway.

    Does anyone have a picture of the main BMS board?
    �
  • Nov 6, 2015
    scaesare
    If you go find wk057's "custom BMS" thread, I think he covers a bunch of this and has some board pics.
    �
  • Nov 6, 2015
    tom66
    What big IC on the bottom? If you mean J1, that's a connector... Q1 thru Q6 are the sink FETs sinking thru the 158 ohm resistors, or 39.5 ohm in parallel, or roughly 90mA.

    The BMS boards are chained together. U4 isolates them, it's a Silabs digital isolator, IIRC

    The BMS ICs (bq76PL536A) are capable of being chained to each other via a current sink communication protocol, however that would present a risk as if a module were to experience a short, the IC would be overstressed and could fail. This might cause further damage to the pack electronics or prevent other cells from being balanced properly which would be bad.

    I wonder what interface the controller will use to communicate. I reckon it would be possible to reuse these module controllers but you would have to observe the pattern of data from a working pack to understand how to send request messages, presumably it supports things like forcing early cell balancing. I wouldn't be surprised if the main controller handles all the balancing and the module controllers only balance if there is a serious over-voltage problem or some other issue.
    �
  • Nov 12, 2015
    Quantum`
    So it does seem to be labelled J1 and we have to assume it is a connector. But for what? Maybe for incoming current for the module as all are heavy pins, but what specifically, and is so much current shunted out those tiny pins on J3 to the cell stacks? How can that be? The BMS IC is not doing comms with other boards, the microcontroller serves as the Host for the BMS IC and so is its master.

    As the microcontroller is the comms gateway, it has two possibilities, SPI and a UART. My money's on them using SPI, which is differential and so is almost immune to noise, plus its a bus architecture, with 4 lines. An EEPROM isn't going to have alot of storage to hold functions, so I doubt comms are encrypted.

    Just don't have enough info to get further.
    �
  • Nov 12, 2015
    wk057
    The BMS boards do not do any charge shuffling. They do top balancing only by bleeding off excess power in cell groups that are highest to allow other cell groups to charge more.

    Why the connector for the communication chain cable is so huge I'm not really certain. Probably just ease of install. There is no real power flowing through those pins at any given time, just enough to power the comm side of the isolation IC on 16 boards.

    The microcontroller on the board is definitely an interface chip between the isolated communication bus and the TMS BMS chip, which makes it all the more difficult to reverse engineer since the firmware of this chip is completely unknown.
    �
  • Nov 12, 2015
    Quantum`
    There is a good reason why the pins on J1 are so large. Tne cell stacks have to get charge from somewhere, and I'll bet that is it, although how it gets current to the cell stacks is a mystery.

    I wish I had the numbers from the TSOP16 so I could get an idea exactly what it does. And I wish I had an idea of how the BMS boards daisy-chained together. And I wish we had some pics of the one odd BMS board, to try and get an idea of its role.

    The microcontroller's software is in the EEPROM. These can be unsoldered and clipped to read from a fairly simple USB reader. Of course it is machine-code, but there are disassemblers.
    �
  • Nov 13, 2015
    scaesare
    Unless I'm misunderstanding what you mean, the cells (96 series connected groups of 74 parallel connected cells), are charged via the main power harness they are wired to for also delivering power. Think about it: at a supercharger, there's 300+ amps being delivered to those cell groups... that board connector would evaporate.

    No charging of cells takes place by any of the BMS system as far as I know.
    �
  • Dec 1, 2015
    jat
    Have you charged a whole module using your PowerLab 8? It seems like either you could hook up the cell voltages from the connector to the BMS board to the balance wires to the PL8, or you could rely on the BMS board doing the balancing itself.
    �
  • Jan 29, 2016
    Jessing
    Dear All,

    Does anyone here have some information of the valves in the bottom of the battery casing?

    There are 6 of them shown on wk057' picture 2014-09-09 2021.21.31.jpg. wk057 call them "some kind of safety relief vales".

    I believe they are meand to drain the battery casing from condensate, and at the same time to block any inflow of water from the inside, fx if the car is parked in 6-7 inches of water, after heavy rain. A one-way-valve in other words.

    I need similar valves for a project, and I have searched everywhere I can imagin, but without luck.

    I also asked wk057 if he had some of them in excess, but unfortunately he throw them out as they was not useful for his project. :-/

    Any information, such as a brand, vendor, data, more pictures, or a link to an alternative valve with similar funktion would be helpful ... and ofcourse; if anyone has one or more samples, I would be very interested to buy.

    Thanks :smile:
    �
  • Feb 3, 2016
    Ingineer
    They are fire vents. They are designed to let excess pressure vent safely underneath the car while not allowing air (or water) into the pack in normal circumstances. There are 6 in each of the 15 "brick" compartments which are isolated from each other. If there is ever enough moisture in the pack to need draining, your pack is already done. Other than the 1-way vents, the pack is pretty well sealed.
    �
  • Feb 4, 2016
    Otmar
  • Apr 8, 2016
    zZZzzzSnooze
    They can have a P200DL & 200D Model S Today!! that battery alone would cost around $30K US.

    Just looking at the pictures of the open 85kW battery pack, They can bump up the capacity to 200kW, boosting range from 435km on the P90DL to 970km & from 460km on the 90D to 1,030km... The P200DL & 200D can be made now... Charge times to 80% on the P200DL & the 200D would take something like 1H:29M to 80% & fully charged would take 2H:47M on a Supercharger. At home on a 240V outlet would take 26+ hours. 110 volt? lol 6 days...

    I was just looking at the pictures, I could be wrong...! There is 7,040 batteries in a 85kW Model S battery with a fully charged voltage of 336DCV. 16x 21 volt fully charged banks with 440x 1850 batteries, and that 21V bank is divided into 5x 4.2 volt 1,000mAh capacity 1850 battery connected in parallel. Samsung has a 2.5Ah 1850 battery that can be discharged at 25A. Panasonic has even a larger capacity 3.4Ah but a discharge rate of 3A... but you have 88 of them connected in parallel, The Panasonic should still be enough to power the Model S to its full power. Wow 272kW Model S battery, and them Panasonic's are at least twice the cost of the Samsung!!

    With Panasonic batteries charge times to 80% on the P272DL & the 272D would take something like 2H to 80% & fully charged would take 3H:47M on a Supercharger. At home on a 240V outlet would take 36+ hours. 110 volt? 9 days...

    [?IMG]

    Opened 85kW Model S battery
    [?IMG]

    Voltage on a Bank
    [?IMG]

    Voltage on a 85kW Model S battery?
    [?IMG]

    Front dual layer banks
    [?IMG]

    Thanks for the Pics/Info wk057!
    �
  • Apr 8, 2016
    Ingineer
    You are a bit off: In the 85/90 pack there are 16 modules in there each containing 444 cells for a total of 7,104. In each module there are 6 series strings of 74 cells in parallel, so that's essentially 96 large cells in series for the total pack.

    Fully charged the pack is at around 400 volts, so that makes about 4.17 volts per cell.
    �
  • Apr 8, 2016
    Andyw2100
    Welcome to the forums!

    Just a bit of advice: if you're going to use someone else's photos in the way you just did, you may want to note that the photos are someone else's BEFORE you start including them and describing them. As I was reading your post I was thinking to myself, "Is this guy really taking credit for Jason's pictures," until I got to the very end, where you did include the note above. A simple note along the lines of, "the following pictures were taken by wk057" in the future would avoid any such confusion.
    �
    • Người đăng: vào lúc

    Không có nhận xét nào:

    Đăng nhận xét

    Đăng ký: Đăng Nhận xét (Atom)