Instrument Cluster Rebooting / Components That Fail in the Cold

[trogdor1138]

I have an '08 Lexus LS600hL with an instrument cluster that reboots when cold (<45F or so); here's a video of the problem in action: Video

As the car warms up after starting, the cluster remains on for longer and longer until it's fine. It will continue to operate okay after stopping and starting, assuming it hasn't been allowed to cool down to ambient temperatures.

I know that failing capacitors can be temperature-sensitive like this, and the previous generation of the LS specifically had problems with capacitors in the instrument cluster, so I replaced all of the capacitors on the boards with new parts. The originals were quality caps from Panasonic and Chemi-Con, and I used polymer replacements from Panasonic, Chemi-Con, and Nichicon.

However, the issue is not resolved. I went out to the car this morning (about 32F) and the problem persists. I realize that cold solder joints are a thing and this could just be thermal expansion at play, but are there any other components that behave in a similar manner? I'm asking here because the service procedures only replace the entire cluster as a unit, but I'm sure it's something that can be repaired. This was also a relatively low-production model, so parts specific to it are hard to find or no longer available.

I've attached images of the boards in question for review.

[p.larner]

can you not plug it back in and try some freezer spray on it?,to at least localize the problem area.

[themadhippy]

 a can of freezer spray will help find any thermal problems

[trogdor1138]

Yeah, I know that's the common method to pinpoint issues like this. The only problem is that I'm not sure I can bench test this, and there's not much room in the car to work. I'll have to give this a think to see if I can come up with something

[Gyro]

a can of freezer spray will help find any thermal problems

Not to mention condensation related ones

Actually, condensation is one possibility. It would be worth looking at high impedance areas, particularly around the crystals.

[dmulligan]

Try hooking it up to a power supply when it's that cold.  What if the unit is not receiving enough power rather than a failure of the unit itself.

[trogdor1138]

OK, I managed to rig up a small harness for testing. I discovered the connectors are manufactured by TE Connectivity and was able to locate the correct part directly from them, bypassing the dealer. A little freeze spray and +13.8V later and I've isolated the problem component: Video identifying the problem

I'm unable to identify the QFP component in question though, so if anyone has insight that would be helpful. I'm attaching a higher-resolution image of the board, but it has these markings:

D (company logo?)
457789-0723
F016A9007
0721 Z02

I don't have much experience in this area, but I haven't come up with anything while searching. There are a few Chinese suppliers that list "457789" or "D457789" chips and their use by Toyota, but without any further details or datasheets. There have been a few references to NEC, which would make sense as a Japanese manufacturer, but nothing concrete. If it is something bespoke just for Toyota, is it likely that the chip is failing internally, or is it more likely to be bad solder joints? Is it possible that getting this component re-worked would be enough?

I'm also getting some coil whine (audible in the video) from a large transformer on the power supply board that I'm also unable to identify. It's marked:

P13
K7327

Obviously schematics would be the way to go, but given that these are boards internal to Toyota and were only replaced as part of the larger assembly, that's out of the question.

[lugaw]

Wire up a ceramic resistor and stick it on top of the chip, that would keep it warm.

[J-R]

Did you check each pin on the chip to make sure there are no bad solder joints?

[fmashockie]

That D457789 IC is probably some kind of FPGA or CPLD.  So even if you could replace it, you'd still have to reprogram the new IC.  That would prove rather difficult considering a datasheet is not available. 

It sucks when these kinds of issues occur.  I've had PCBs at my work that control robotics go bad - everything on the board is fine except the CPLD or FPGA goes bad.  In cases like these, sometimes the most economical option is to replace the board.

But still give other options a try and continue to isolate the issue to just this IC first, before looking at replacing it.  Good luck!

[Haenk]

Did you check each pin on the chip to make sure there are no bad solder joints?

I'd say: Just try to resolder all pins, maybe that's enough to fix it.

[tunk]

Could you try to push it, or maybe flex the board a bit - if it's
a bad solder joint then it may be enough to make contact.

Edit: Or if you push or flex it at room temperature, then you
may break a bad solder joint leading to a reboot.

[J-R]

I suppose you could resolder all the pins, but that could be a lot of work depending on what equipment you have.  Solder bridges, etc. could make things annoying.  So I vote for checking first by gently wiggling each pin with something sharp and pointy (with the power off!!).

[edpalmer42]

That chip appears to be covered with a conformal coating.  You can see the edge of the coating where it reflects the light and appears as a white line running below the chip near C52 and C53.  That will complicate any repairs because you have to remove the coating to check or repair any solder problems.

When you tested the chip, how careful were you with the freeze spray?  To the left of the chip is the crystal and C61 and C69.  The capacitors are both enclosed by an outline.  Crystals are typically flanked by capacitors to ground.  If one of them is cracked and your freeze spray splashed onto it, that might make you think the chip is bad.

Ed

[trogdor1138]

Really appreciate all of the responses and discussion.

I agree that it's most likely a programmable component; my thought was to get a dump from it if someone recognized it and knew how. It may very well be something completely bespoke for Toyota though, as no one I've contacted so far has known what it is.

I also had the thought of having someone more experienced than me rework the chip, but it does indeed have a coating over the pins. Does that generally mean it's not going to be removable without damage?

I may be too hung up on identifying and replacing this one chip though; I think I may buy a used cluster and harvest this board from it. I'm pretty sure the 8-pin EEPROM above and to the right stores the mileage for the odometer, and I should be able to swap that onto another board without too much difficulty.

[Haenk]

Try to cover parts and only leave small sectors to cool them down, one by one. So you could narrow down the failing area a lot.
I don't think the chip is damaged, very likely only a dangling leg. If you can source a cheapish replacement (that's probably not going to happen on a 600h), and you were able to find the area, I'd just fire up the soldering iron an heat up the legs, while pressing them down. Ignore the coating.

[darkspr1te]

I saw this and remembered a mate of mine worked for Toyota SA and had mentioned the cpu's once so I called him and he said they are a NEC part and the common name is HC4C05 or later models are HC5C05 ,
Later models had h/w canbus bootloader but earlier models did not, but chip is the same for pin and code wise, it's some common motorola clone part, the HC4. He also mentioned that the reason canbus boot/programming was required now is a blank car via a new ecu would setup all the devices with the right firmware and VIN numbers , a one time deal when the ECU booted and only when connected to a master PC with software from japan that could not be touched (it had barcode scanner only as input, if a mouse was connected it crashed and patched the system to not function until japan fixed it via remote) , bar code scanner was PS2.


Also mentioned was bootloader/reflash pin code was a per VIN based value and bootfiles/replacement code can be found on the russian ECU forums for key programming and the like as well as vin decoder and you may need a win95/win7 virtual machine.


He is not sure how they are these days as he does not work anymore (toyota/ford etc all closed any local electronical support just before covid)
I checked the two sites he gave and non exist anymore and not on wayback either.


He did mention there are some low level EEPROM ECU programmers out there for high end vehicles that can reprogram this, It's due to them having all the dumps from the CPU's for key programming and ecu resets and even VIN changes.
but these devices are big bucks h/w wise and subscription for the files too so it would be cheaper for a scrap vehicle unit.




One thing to try is when it's cold use the eraser end of a pencil on the legs and see if one is cold soldered.




darkspr1te