Thanks to an astute poster in this forum from two years ago, I located a cooperative US distributor for the orange 22-pin Yazaki 7382-6100 connector in the battery management computer. This connector corrodes due to gases coming from the battery pack. I'm curious why the corrosion appears worst around pins #19, 20, and 21 of this connector. In the example below these pins are literally missing from the connector: Does anyone have access to schematics for the battery management computer and could say what voltages appear on pins 19-22? I'm betting that pin #22 is the bottom of the battery stack voltage-wise and pins 19-22 are (probably due to a design error at Toyota in the early 2000s) the opposite end of the battery stack with the highest voltages relative to pin #22. But on further examination there's a 400V electrolytic cap tied to pin #22 so it might be the opposite of what I describe: pin #22 is the top of the battery stack and pins 19-21 are at or near the bottom of the stack. My guess is that someone forgot to sequence the connector's pin-out so that adjacent pins never have more than one block voltage (14VDC) between them. The 2.5mm pin-to-pin spacing in this connector could not long tolerate 200VDC between pins. Nor could the PCB underneath the connector, even with conformal coating on it. I turned the magnification up on my microscope and can see that the skinny copper PCB tracks exiting pins #21 and #22 have served as electrical fuses and blown open as part of this failure mode. The battery block voltages from the Yazaki connector are fed to a group of 7 dual-channel Matsushita/Panasonic linear optocouplers.
Thanks, Chap. It looks like I was reinventing the wheel on my Yazaki connector study. Seems that others have pursued the matter to its conclusion. The photos of totally blown-up Yazaki connectors the posting you suggested are concerning. It would be difficult to sustain a pin-to-pin arc with only one block voltage between the pins. So that suggests that T's supplier didn't sequence his connector pins with adjacent battery blocks between adjacent connector pins.
Didn't realize that there'd been an entire book of postings about it some time ago. Everyone who works on Prii should read that series of postings. Sobering... Matt, the Texas BMS rebuilder, in his responses to Chap's comments a while ago, answered my question above about how much voltage appears between pins 21 and 22 of the orange connector. Matt's answer was 30VDC. Given that the voltage is DC, it's sufficient to sustain an arc in a narrow space. My guess is T's BMS vendor contaminated his Yazaki connector with rinsed-off reflow solder flux which has a ton of ionic crap in it. Subsequently, moisture in the air driven by the voltage differential between adjacent pins, will get the corrosion rolling.
So, to avoid it, what? Dunk the whole unit in distilled water, agitate for half an hour, and then dry thoroughly, perhaps in front of a fan? I recall trying to open that case and it had some peculiar sort of Japanese screw like fasteners in it that looked like Phillips but weren't. When none of my screwdrivers fit well and they wouldn't budge with moderate force I gave up. Because it would of course be better to pull the board out and clean it rather than to dunk the whole assembly.
That sounds like it would be a PITA to fix. Is the board still relatively flat there or did the heat pit it?
Pasadena, sorry for the delay in responding. I was out of town. My stereo microscope doesn't have sufficient magnification for me to say definitively that the damaged PCB copper tracks emanating from the orange connector have consequences in inner PCB layers. I think not, but I can test that. It's also likely that the PCB is only 2 layers, not 4. The PCB is not sufficiently complex to require 4 layers. Prior to any BMS internal fire I think these PCBs can be repaired. The orange Yazaki connector is available through Mouser and Digikey, including the matching plug. I also notice that one of the EBay sellers of fresh bus bars also includes in his kit a new orange wire harness with new plug. *** In the intervening week I've given this BMS orange connector failure additional thought. Poster "Matt" a couple of years ago on this corrosion topic remarked that the two most corroded pins in the orange connector are 0VDC (i.e. bottom of battery stack and the top of the second 14V battery block. The voltage differential between the pins is only 30VDC or so. 30VDC from a zero ohm source impedance can certainly sustain an arc but not likely start the arc. So how is the arc being initiated? I think the answer is tied to bus bar corrosion. Dr. Prius app reveals that the most stressful situation for the battery stack is high speed deceleration using regenerative braking. I saw battery currents during regen braking of more than 100 amps. Ohm's Law (I squared R) suggests that even 1 milliohm of resistance at the corroded interface between a cell's terminal bolt and the buss bar screwed to it causes 10 watts of heat dissipation. 25 milliohms of interface resistance (which is less than the internal resistance of my 2008's battery blocks) would cause 250 watts dissipation at that bolt/bus bar interface. I believe what's happening is that a heavily-corroded bus bar enters thermal runaway during heavy regen braking. The over-temp event is short (only as long as braking lasts) but the voltage across the bolt/bus bar interface briefly rises above 100VDC and arcing is initiated between the two corroded pins in the orange BMS connector. Said differently, the voltage seen by the two pins in the BMS connector is the sum of the two block voltages (30V) plus any voltage across the bolt/bus bar interface. If there's any corrosion on the other 3 bus bars associated with blocks 1 and 2, this voltage also adds. After the regen event ends, the 30VDC background differential voltage on the orange connector pins sustains the arc on the PC board. Some BMS modules flame inside as a result. Interior BMS fires are simply the final closing act of the bus bar corrosion drama. If Toyota had placed series resistors in the voltage sampling wires of the orange harness, the fires would not have happened. However, the corrosion leakage current at the two pins in the orange BMS connector would have disrupted the battery sampling and thrown block voltage errors. It's known that the corrosion at the bus bars is caused by electrolyte leaking out of the cells. The corrosion at the orange Yazaki connector is likely coming from wintertime driving episodes that also cause the vehicle's windows to fog up on the inside. The moist air in the passenger compartment blows across the cold BMS. Fine condensation probably occurs. Maybe there's even some electrolyte vapor wafting inside the battery enclosure to add to the list of corrosion accelerators.
It is a little odd that there seems to be no protection on those sense wires. In another similar application in the same car the sense wire from the 12V battery to the inverter has a fuse right next to the positive post. I guess Toyota engineers figured that since the module sense wires are inside a metal box they were unlikely to short to the wrong bit of metal and cause a problem. Still, if they had put even 1K resistors near the bus bar it would have greatly reduced the magnitude of possible currents down the wire, and if something bad happened surely better to replace the wire than the computer. Super easy to diagnose a burned out resistor in a long wire. Voltmeter inputs are typically very high impedance, so 10K would likely have worked too, and that would have limited the current in the absolute worst case to roughly 200V/10000, or 20 mA. Besides protecting the electronics, it would have been a safety measure for anybody working on the live battery, pushing the maximum current below the 100mA - 200mA range which is most dangerous for potentials near 200 VDC. Does it though? The BMS is inside a mostly sealed metal chamber which is walled off from the cooling air from the cabin which flows through the HV battery. Under a carpet (sort of). Under a floor. I wouldn't expect the air to exchange very quickly between the vicinity of the BMS and the cabin, although it probably does so slowly. Another possibility, at least in some instances, is that the Prius has a leak, most often from one of the seams just above the hatch hinges. That could put a macroscopic amount of water close to the HV battery, and that would surely increase the humidity in the region, if not even wick up into it.
You could be right about the nature of source of the source of the corrosion on the Yazaki connector. Condensed electrolyte vapor is a stretch explanation. Another thing occurs to me: I remember reading that T's design for the HV inverter includes a boost inverter circuit so that they're able to drive the electric traction motor at voltages substantially greater than the traction battery's voltage. The boost inverter architecture has a large inductor at its input. One side of the inductor is anchored to the + pin of the traction battery. The other end goes to an N-channel switching device like an IGBT. T might be using the traction battery's low intrinsic resistance as the primary means for stabilizing the voltage on the input (unswitched) side of the boost inductor. (A prudent man would save his electrolytic capacitors for the output [HV] side of the free-wheeling diode.) If the effective source impedance of the traction battery and its corroded bus bars grows dramatically during regen braking, you could briefly see some of the boost inductor's back EMF appearing across the bolt/bus bar interface. If true, the voltage seen by the Yazaki block sampling wire harness could briefly be hundreds of volts, not just the 230V stack voltage of the battery. And is there any thermal cut-out switch in T's NiMH cells? If so, does this switch ever open up and briefly allow the a cell to become an open circuit? That, too, would create high differential voltages on the Yazaki block sampling wire harness.
No. There are three temperature sensors for the whole pack, and if it gets too hot they can cause the car to throw codes and effectively turn off the HV pack (and so the car). I don't know if over temperature is enough to cause the car to open the HV relays in the pack and isolate the modules from the rest of the car.
As far as I'm aware, over or under temperature triggers merely change the charge and discharge limits to significantly lower values.
Toyota's got electrolytic capacitors on both the battery side and the boosted side of the boost converter. As for temperature, I think their usual response to extreme temps is just to dial back the limits on maximum current to demand from or deliver to the battery. Actually lighting the warning triangle and opening the main relays would be reserved for more serious faults. (If I remember right from the fail-safe chart, there may be conditions where the main relays won't be opened if you're already driving, but the condition becomes a hard no-READY once you've stopped and turned the car off.)
