Chasing info code 123 *without* P3000? Or do I need a better scan tool?

Hello, all!

Quick version: My 2001, with 223,000 miles on it, is giving me the !-triangle and !-car lights, but no Check Engine light. Applying the scanners I have available isn't giving me any powertrain codes (Pxxxx), but I did get it to cough up an information code of 123. Per the Toyota service manual, this may be associated with P3000. Reading the service manual and Googling implies that some other code(s) "should" be set if you are getting a P3000-123, and you should chase those code(s) to find out what's actually wrong.

Questions: Has anybody seen this before - information code 123, with no other code set?

Do I need a better scan tool (Techstream?), or to do something different with the one I have?

Long version:
Long post ahead. Sit down and get comfortable.

Long-term history:
Bought the car new in spring 2001 and have owned it since then. It has about 223,000 miles on it now.

The original traction battery had the TSB for the module sealing (black squares under the module studs, goop on the terminals) applied to it by a dealer.

It has never had the TSB for "misfire code after cold soak" applied to it (which, IIRC, is swapping one of the ECUs). I just live with it in the winter.

In summer 2010, at about 122,000 miles, the original traction battery failed, with at least one bad module. I bought a new traction battery from Toyota and swapped it myself. I re-used everything that didn't come with the new battery (the battery ECU, system main relay, service plug, HV cables).

In summer 2011, at about 143,000 miles, the original (!) aux battery failed. I replaced it with the kit from Toyota that involved a physically bigger battery, a new "cradle", and different battery cables.

In spring 2018, at about 220,000 miles, the aux battery failed again. I replaced it with an Odyssey AGM battery; I didn't think the Toyota battery was a good value.

Short-term history:
The car was parked for a few days earlier this week.

On Thursday, I ran it for a few minutes to move it from one side of the driveway to the other, and then again to move it back. No warning lights or other misbehavior either time.

It rained Thursday night and Friday morning, but nothing out of the ordinary for Missouri in August. (The car was parked outside).

It sat until Friday evening, when I started it up.

The problem, and what I tried:
When I started it, I got the !-triangle on the speedometer, and the !-car on the touch screen. I did not get a Check Engine light. I tried cycling the ignition key a few times, but neither indication cleared. I tried going all the way to "Ready" (with Ready light and ICE running), shutting off the key, and repeating a few times, but neither indication cleared.

I used my "old" scan tool (see below), but it did not find any codes current or pending, from either ECU it can see. It said the Check Engine light was off. I told it to reset the codes, and the !-triangle and !-car cleared. I then started the car and let it run while parked, and in less than a minute, the !-triangle and !-car returned.

I got my "new" scan tool (OBDII-Bluetooth dongle + Torque Pro) working. It did not find any codes current or pending. I told it to reset the codes, and the !-triangle and !-car cleared. I then started the car and let it run while parked, and in less than a minute, the !-triangle and !-car returned.

In Torque Pro, I dug into the extra parameters that are available from the PID file posted here. The battery module voltages looked reasonably OK... they were close to each other at key-on-not-Ready, and when I went Ready and the ICE was running, they all went up a little. The state-of-charge was a bit over 50%. The "Number of Battery DTCs" PID reported 18, but I don't know how to find out what the DTCs were.

The main thing that caught my eye was that information code 1 was set to 123. (The other four info codes were all 0.) The service manual only talks about 123 being set in conjunction with P3000. Reading here, as well as the service manual, makes me think that I should have some other code set, which will point to a more specific problem. But I don't have that code... or I can't see it with the equipment I have.

My "old" scanner comes up by default talking to what it calls "ECU #10". It has a "toggle ECU" function, which results in it talking to "ECU #16". #16 reports some of the same basic ICE-related values that #10 does (intake air temperature, coolant temperature, engine RPM, etc), but not everything. If there is a trouble code set, #16 usually reports the same trouble code as what #10 reports.

I haven't found a similar "toggle ECU" function in Torque Pro. I feel like maybe one of the ECUs in the car has a code for me, but I haven't figured out how to talk to that ECU yet.

Equipment available:
A mid-2000s OBDII-to-RS232 (!) scanner from Harrison R&D. This scanner has successfully read powertrain codes before (P0xxx), but doesn't know how to read the Toyota-specific codes, like the traction battery voltages.

A couple-year-old "generic" OBDII-to-Bluetooth dongle. It has five LEDs; a red one that comes on steady when I plug it in, and green and orange ones that blink when Things Are Happening. This is coupled with an Android phone that has Torque Pro 1.8.202, plus the PID files available from the "Scangauge or Torque App?" thread and from kutasg on Github.

Both volumes of the Toyota factory service manual, plus the wiring diagram and new car manuals, on genuine dead tree slices.

Digital voltmeter and (if needed) analog oscilloscope.

Plan of attack:
This will probably happen early next week (after Labor Day). The easy stuff first. See if I can hear the traction battery blower running. Pull up the trunk lining and see if the 12 volt harness to the battery ECU looks beat up in any way; maybe disconnect the aux battery, unplug the 12 volt harness to the battery ECU, inspect the pins, and replug. Take the service plug out and put it back in. Check the aux battery to make sure it and its cables are secure, and that its voltage is OK.

Thanks!

 

Interesting, thanks! The only times I have had the traction battery cooling ducts apart were when I was swapping the traction battery (you have to), and when I was swapping the aux battery (taking out the duct above the aux battery makes that process a lot easier).

When you cleaned the traction battery fan, was there a lot of dust/dirt/grime on the fan blade? Mine has never been cleaned, at least not by me.

Mine has a plug with a hole in it, where the vent tube from the factory aux battery and the replacement flooded battery went through. I don't remember if the plug is in the "floor" of the aux battery area, or if it goes a little more inboard and then through the floor, but it's down in that area somewhere.

One thing I did notice back when I swapped the traction battery is that the factory traction battery had a solid piece of sheet metal underneath the secret charging connector on the driver's side (or, in other words, outboard from the service plug, around the corner of the battery, and forwards). The replacement battery had about a 1"/25mm diameter hole in the sheet metal in this area. I covered the hole with electrical tape before I installed the battery in the car; I figured it would be a cooling air leak if nothing else. This ends up inside the trunk, so if the trunk seal is working, it *shouldn't* make a difference, but maybe it does. If I can do it with the battery in place, I will try to see if that tape is still there.

I was able to load up Torque Pro with the definitions in the nearby "Scangauge or Torque App?" thread and see a lot of interesting Prius-specific stuff, but not the trouble codes. (Link: Scangauge or Torque App? | Page 7 | PriusChat )

I remember doing this when I swapped the traction battery. I got the regular OBDII connector to cough up some codes with the scan tool I had, but I also got (IIRC) a brake/ABS code when the traction battery failed. I jumped those terminals and watched the light blink; I don't remember the code exactly, but the way I interpreted it was that the brakes were mad because the battery was mad and therefore they couldn't use the battery to regen anymore. When I got the new traction battery in, the brakes got happy on their own.

Thanks! I will hie myself off to Ebay or Amazon and peruse their selection.

 

I got myself a Mini-VCI from Amazon, and used the copy of Techstream 10 linked at ScanGuage AE Pig-tail, code flasher, and TechStream | Page 2 | PriusChat . (The Mini-VCI also came with a copy of Techstream 12, but I haven't tried it yet.)

Survey says... P3000 with info 123, P3009, C1259, and C1202. P3009 is "Leak detected"... the battery ECU thinks that the high voltage is leaking out somewhere it shouldn't be. Which, correlated with the rainy conditions that happened when it set, is interesting. The trouble tree for P3009 is long, but it basically comes down to taking off the "battery carrier catch bracket" (the cover over the DC cables at the traction battery) and the top of the traction battery to see if that clears it, and if not, to start disconnecting and megging (insulation resistance testing) the DC and AC high-voltage cables. C1259 is the "I can't regen" from the ABS computer, which is probably expected if P3009 is set.

C1202 is the brake fluid level switch - either low brake fluid or open circuit. I think this is an old code - the brake fluid level has been dropping very gradually over the past few years, I believe due to normal pad and shoe wear. Several months ago, it was still above the "min" mark when the car was level, but the "brake" warning light came on when I drove up a pretty good hill. I bought a can of brake fluid and filled the reservoir back up and haven't had any problems since then, but I think the ABS computer remembers my sins.

After reading all the codes and saving the data, I cleared them. I drove back and forth in the driveway (only a couple of car lengths) and they didn't come back. I then did a short drive up and down the block - probably less than a mile - and they didn't come back. A little later, I drove to dinner and back - about 3.5 miles - and they didn't come back. (It rained earlier today, but it wasn't raining when I was testing.)

My plan is now to do a visual inspection of the DC cables at the traction battery, under the car, and at the inverter, and as much as I can see of the AC cables coming out of the inverter for MG1 and MG2. I will also check the HV battery cooling ducts. If I don't find anything, I will probably take the traction battery cover off to see if there's anything interesting inside there.

I know I'm preaching to choir, but Techstream is pretty good. I spent a little time just playing with it in the car. The only slight burble I found is when I asked for the list of parameters from the steering ECU (EMPS); I got the whole list and then Techstream complained that it lost communication with the car. I unplugged/replugged the Mini-VCI and the same thing happened again. It didn't do that when talking to any of the other ECUs. I don't think this is a big problem - just something I noticed.

Thanks!

 

Quick version: I checked to see when the car was throwing codes, and then got jiggy with a pile of Prius parts and a megger. The megger says all is OK, but in the process, I found water in the car in suspicious places. Troubleshooting is still in progress.

I cleared the codes, then let the car sit for 1 minute, and then 10 minutes, in ON-not-READY and it did not throw a code again. So far, the code only throws when the car is ON-READY. I have seen it throw at ICE idle in Park, and also not throw until I shifted into Reverse or Drive. I got a megger and meggered the DC cables, the MG1 and MG2 windings, and (gulp) the inverter itself, and all of those readings appear to be in spec. There wasn't any visible moisture or crud under the covers for the MG1 and MG2 cables. I took the big inverter cover off and looked in there and didn't see any moisture or crud either.

When I took the rear seat cushions out, I noticed some moisture/condensation under the seat bottom, and on the DC cables at the battery. I didn't run the car to check codes in that condition, though.

Long version: While waiting for the megger to ship, I did a little reading here about testing in ON-not-READY and ON-READY (to use the system main relay to isolate the battery from everything else), so I tried it. As above, I couldn't get it to throw codes in ON-not-READY, but I could in ON-READY. (I'm fairly sure the very first time also happened in ON-READY but I don't remember for sure.) So, to me, that points to something on the inverter/motor side of the system main relay.

Starting at 21:00 Thursday night (6 Sep), all the way to 15:00 Saturday afternoon (8 Sep), the relative humidity outside was between 93% and 100%. It got as low as 78% later on Saturday evening, then went back up to to 100% for a few hours before dawn Sunday, and got down to around 66% by the time I started working on the car on Sunday. For all that time, the car was parked outside with all doors closed and windows rolled up.

Today, the megger came in the mail, so I set out to use it. (It's a VC60B+; mine is "Vici" brand, but it's also sold as Victor, Samyo, Kiamitor, and probably other names too.) When I pulled the bottom seat cushion, there was some moisture/condensation on the body underneath the cushion, towards the rear of the cushion. Not a puddle, but like those fine droplets you get when you take a Coke can out of the fridge and then immediately take it outside in August in Kansas City. At that time, I didn't notice any water on the parts of the battery cables I could see (just inboard of the driver's side rear door).

I then pulled the rear seat cushion, the two metal bars, the plastic "wall" behind the seat cushion, the service plug, and the "battery carrier catch bracket" (the cover over the DC terminals). When I pulled that, I noticed a little bit of condensation on the insulation of the DC cables themselves, outboard of those metal rings that contact the ground plate. I also noted some small rusty spots on the top leading edge of the battery cover - again, not like the whole thing had been underwater, but a few spots around 0.25-0.4" (6-10 mm) or so diameter. I went ahead and disconnected the DC cables from the battery.

I went up front and pulled the wipers, the plastic cowl cover, the wiper motor, and that metal cowl tray. I then pulled the DC cables off of the inverter. I went back to the trunk and tested between each individual cable terminal and body ground, per the manual. At 500 volts, at which my megger goes up to 200 megohms, the megger always showed over-range - more than 200 megohms. The spec is 10 megohms, so that test passed.

Up front, I pulled the MG1 and MG2 connectors, and meggered between each one of them (motor side) and body ground. For these, the megger started at a lower reading (around 20 megohms), and then rose over the course of a few seconds to over-range. Again, the spec is 10 megohms, so I believe those passed.

Finally, I meggered between all eight of the MG1, MG2, and battery terminals at the inverter, and body ground. All eight read 1.1 megohms on my meter. However, the meter beeper ran continuously, rather than intermittently, which I think (based on experimentation with fixed resistors on the bench) means the meter was current limiting... if it had the ability to deliver more current, the resistance reading might have been lower. Spec for this is 0.9 megohms, so I think it passed. The meter current limit might be fooling me into a pass, but the fact that all eight were the same lends credence to the pass, to me.

The areas under the MG1 and MG2 terminal covers looked clean and dry. I took the big square inverter cover off and looked in there and that all looked clean and dry. I also took off the plate that the top circuit board in the inverter sits on and lifted up and looked under there; again, looked OK. I didn't try to dig any further into the big busbars or the "wet" half of the inverter.

I had to go do something else, so I replaced the inverter cover, cables, and terminal covers. I also laid the seat cushions back where they normally go - I didn't bolt/clip them back in, but I kind of want to see if they are wet underneath when I pull them up again. I will probably work on it more on Monday.

Plan of action: Reading the service manual just now, I notice I missed a trick. I was supposed to megger from the MG1, MG2, and DC terminals to the inverter body (which I think I did - the inverter body is grounded to the car body), but I was *also* supposed to megger from those eight terminals to the transaxle case. I would think the transaxle would be grounded to the car body (it's bolted to the engine, which I know is grounded to the body), but maybe I should try it their way.

I may try megging the inverter at 250 V instead of 500 V, or possibly at 500 V with different batteries in the megger just to see what happens. As above, I think my 500 V readings were current limited.

I will take the battery cover off and inspect in there. Since I haven't yet got a code with in ON-not-READY, I'm not sure if this is my problem, but I am running out of places to look.

I will also look around at the body sealing. I know there are some beat-up places on the seal on the body around the driver's door, but I didn't think they were enough to let water in. There is also a tear in the leading edge of the "scraper" for the driver's side window (just aft of the outside mirror). The moisture under the rear seat makes me suspect the rear window seal a little bit as well.

 

Quick version: I meggered some more and the inverter, MG1, and MG2 were still in spec. I found a couple of minor body seal issues but I don't think they were the main problem. I popped the case on the HV battery and found some crusty terminals. 76 nuts off, 36 busbars off, soak busbars and nuts in vinegar, rinse in water, wire brush, dry, clean miscellaneous green stuff out of pack, 36 busbars on, 76 nuts on, lid back on, inverter reassembled. Turn car on, get code for "service plug not there, dummy". Put service plug back in, try again. No P3009 in ON-not-READY. No P3009 after READY and ICE idle. Try D and R (foot on brake, car not moving) - OK. Try D and R back and forth 10 feet in the driveway - OK. Drove 3 miles to the quick mart in R, D, and B - OK. Drove 16 miles back home, R, D, and B, speeds up to 50 mph - OK, no codes.

I will test more after I finish reassembling the car. The cowl pan and wipers are out of it right now, so it's not ready for general-purpose driving or driving where the 5-0 will see it. >_>. It's been dry outside for the past 2 days; I also want to test it after being parked outside in the rain again.

Long version: You're in luck, I'm not going to write one of my usual novels right now.

Pictures: Of how it looked when I opened it up, and what the busbars and nuts looked like after I cleaned them, available at 2001 Toyota Prius NHW11 battery terminal cleaning .

Credits: So far, I think @sandy11246 and @Dxta were on the right track, with their advice about cleaning the HV battery. Also @ChapmanF for helping set me upon the Way of Techstream Righteousness.

 

The Black and Teal Book of Truth*, pages DI-333, 336, 337, 338, gives the spec for the wires as "10 MΩ or more". The spec for the inverter terminals is "0.9 MΩ" with no qualifier (page DI-334).

A megger doesn't give you answers quite like a regular ohmmeter; you usually get a minimum value as a spec, so the reading on the megger has to be at least that much - but more is usually OK.

*otherwise known as the factory service manual

 

Quick version: So far, P3009 has stayed gone, but now the service plug microswitch is being dumb. I think I may have caused that problem when reinstalling the trunk trim. After patching that problem, and another 6 miles of driving, P3009 has not returned.

Update: I ran the car again in the driveway today before putting it back together. I let it idle until ICE shutdown, forced ICE with the A/C button, and drove it back and forth in the driveway - no codes. So I proceeded with putting the cowl pan, wipers, trunk lining, back seat, etc, back in the car After reinstalling the wipers, I ran the engine again to bring the 12 V up so I could be sure the wiper arms wouldn't hit things at the ends of their travel - no codes. I got it all back together and stuck the key in again... and got a !-triangle and !-car as soon as I went ON. I tried to go READY and it wouldn't let me. Arrrrgh!

Techstream says: problem with service plug. I went to the back and noticed that the bail or handle wasn't quite vertical. When I put back the trim that covers the top and back side of the HV battery, I was doing it from the back seat side. I remember the hard plastic "tongue" that is attached to that trim and covers the service plug area hanging up while trying to install it, and I reached back from the back seat and moved it so it would sit flat against the back of the battery. I figured I had snagged the bail, so I cycled it down and back up, went up front, cleared the code... and it came right back.

After a little more fooling around, with the service plug microswitch unplugged from the harness, I don't have continuity through the switch in either position of the service plug handle. The tab or ear on the plug that pushes the microswitch lever down is working, and I can see the lever go down and hear it click, but no joy on the ohmmeter. The wires aren't obviously busted where I can see them, but I haven't yet tested each one individually.

I ended up doing the same thing that everybody does when an interlock switch is annoying them >_> and that satisfied the computer. Yes, I am aware that doing what I did will curve my spine, warp my mind, and make puppies cry. I do plan to chase it some more, later this week, but a couple of other parts of my life need to happen first.

The car did OK on a short drive to the grocery store, the Scottish food emporium, and back home - a little over 6 miles total - no codes.

 

I am old enough to have poked an analog VOM across a Big-Arse™ capacitator, on the bench, and watched the same thing happen.

My megger also has a couple of "standard" DMM ranges - 1000 V DC, 750 V AC, 2000 ohms. The first couple of times I megged something, I immediately flipped over to the 1000 V DC range, thinking I would watch the voltage decay. The reading did tend to start at a high voltage and then go down to 0 V, but after about the second time, I realized that there are two different sets of jacks on my meter for "megger" or "regular DMM", so I wasn't really watching what I thought I was watching. I did make it a point to wait a few seconds for things to bleed down before grabbing the meter leads again.

Before I meggered the car, I tried my new meter on the bench. I meggered my regular DMM and the megger said 10 megohms (about right) and the DMM read the voltage. I was also able to watch the voltage decay on the DMM. I also had a couple of high-voltage resistors up to a few dozen megohms laying around (as you do), and I meggered those as well... the readings were at least plausible.

The only time I got bit in this whole adventure was when I went to disconnect the cable on the front of the far passenger battery module. It was a tiny bite, but real, when I touched the cable end. The fat cable should have been open circuit at the system main relay, but that also has a sense wire to the battery ECU. I suspect there's a small capacitor (a few uF max) across the most positive and most negative sense leads inside the ECU, so I got to relieve it of about 274 V. (Back when I swapped the traction battery, I also remember seeing a small spark when I hooked up that cable on the bench - I think that was the capacitor in the ECU charging up when I connected the wire.)

Semi-related: when I bought my house, the inspector I hired brought his IR camera and looked at the walls and roof for water leaks, poor insulation, etc. He mentioned that when he's not doing house inspections, he goes to factories and warehouses and points it at big breaker panels, bus bars, transformers, etc, looking for hot spots. I think that was partly driven by downtime concerns, and partly by safety / insurance company concerns.

I found a link to an earlier version of this pamphlet, I think on a review of the meter I got. The manual that came with the meter is reasonably close to OK, but it was clearly translated out of something else and into English, and not reviewed by a native speaker. The Biddle book was (mostly) easier to understand.

 

Things I would look for:

Make sure there isn't a bunch of leaves and crud on the outside edge of the trunk seal on the body. On mine, stuff likes to accumulate in there.

The water might not be coming in from the bottom, but it's easy to check. Take the spare tire out and make sure the two rubber plugs, about 1.5" diameter each, are still there in the floor of the spare tire well.

Maybe recruit a kid or grandkid to get in the trunk with a flashlight and look for leaks, while you spray a garden hose at the trunk lid from outside. They have the escape lever on the inside trunk latch, so they don't have to worry about getting stuck in there

If you take the bottom rear seat cushion out, you will see that the HV battery cables run from the battery and under a piece of black sheet metal, just inboard of the rear door. That piece of metal comes out (one plastic peg for the floor carpet, and four 10 mm bolts - if it's never been out before, the rear inboard one is probably hiding under some grey foam tape) and you can inspect a little more of the battery cables. Once you have that metal piece out, you'll also see the positive aux battery cable coming in to join the HV battery cables. All three of them run together for 8 or 10 inches, and then go through a big rubber grommet to the outside of the bottom of the car.

Once the cables get on the bottom of the car, they run in a plastic channel, just inboard of the pinch weld (that metal lip that the factory jack has a notch to fit over). You can kind of see the channel if you lay down on the ground next to the car; a small mirror might help you inspect it better. You can at least run your hand over the plastic and see if you can feel any beat-up or broken places, like from running over a piece of truck tire or something. When the channel gets to just forward of the driver's door, it runs up along the bottom of the floorboard and ends on the firewall / back side of the engine compartment. The two HV battery cables go up to the inverter, and the aux battery cable goes to the fuse box on the driver's fender well.

On mine, I also tried looking around the back window for leaks there. It's hard to see the seal at the bottom of the window, but once you have the back seat cushion out, it's not too hard to take off the plastic that covers the C-pillars - grab it up by the ceiling and pull it inboard, away from the body, to disengage all the plastic pegs, and then pull it forward and up to disengage the "tail" that plugs into a slot further back on the C-pillar. This lets you inspect the side of the back glass. (To put that plastic back, start the "tail" back into the hole it goes in, slide the whole piece back and down, and then push it outboard to get the pegs back in their holes.)

If you then put your hand between the headliner and the roof, you can pull down on the headliner and disengage the plastic peg halfway between the C-pillar and the third brake light. With that, you can inspect about half of the top seal of the back glass. You can do the same thing on the other side to see the other half; I think you have to take the brake light down to look under it.

I think the connectors at the inverter where the HV battery cables connect are sealed pretty well, but if any dirt or crud got into them, you might get the P3009 code. They don't have any bolts - you just push the latch on top and slide the connector back. It's a little tough to do this with that pan under the wiper motor in place, but it might be possible. If nothing else, maybe take a rag and clean around the outside of those connectors, without disconnecting them.

Maybe this will help...

Imagine a plain old table lamp in your house. The insulation on its cord is rated for 300 volts, so it works just fine on 120 volts. Now, say there is a little nick in the insulation somewhere. If you unplug the lamp from the wall, take out the bulb, and measure the resistance (ohms) across the prongs on the plug with a regular digital ohm meter, the meter is going to tell you infinite ohms, or open circuit. That meter can only ever apply 9 volts to the cord, though, because its power is a 9 volt battery, and the meter doesn't do any tricks to boost that voltage up. According to that meter, there's nothing wrong with the lamp cord.

If you had an ohm meter that had a 900 volt battery in it, and you measured the resistance (ohms) across the prongs on the plug with it, the 900 volts is enough for a little current to leak out of that nick in the insulation. The meter will give you a reading of something less than infinite ohms - it uses enough voltage that it can tell that there's something wrong (the nicked insulation).

That's basically what the megger does - it is an ohm meter that applies a much higher voltage than normal to the circuit it is testing. Older meggers had a hand-crank generator on them to make the high voltage; newer ones step up the high voltage electronically from a few regular AA cells.

 

Quick version: My service plug microswitch wasn't reliably closing its contacts - the wires to the switch were OK. After operating the switch off the car with a small load several times, applying biosourced compressed air to the switch housing, and gently bending up the lever a little, it seemed to be more reliable. I reinstalled it on the car and drove it for a few miles and it seems to be OK.

Long version: I got some time on Friday evening to poke around the service plug microswitch. Initial testing on the car was the same as the other night - no continuity across the two connector pins, in either position of the switch, whether I pushed the lever down with the service plug, or manually.

I thought I would have to take the battery apart again to get at it, but I figured out that if you remove the service plug, and then the fuse cover, you can remove the microswitch from outside the battery. I used a pair of needle-nose pliers to release the peg that holds the connector to the service plug assembly, and a 1/4" hex Phillips screwdriver bit, by itself, as a tiny stubby screwdriver to remove the single screw and get the switch out of the service plug assembly.

On the bench, the wire from each connector pin to the switch pins checked out OK; I could flex and jiggle the wires and the circuit didn't open. I operated the switch by hand a few times, and measuring resistance at the switch pins showed open circuit with the lever free, but a number wandering around between 20 and 40 ohms with the lever pushed down.

Since the switch is only rated 0.1 A, has probably only been operated maybe a dozen times in its life, and probably only switches a few milliamps, I thought maybe all that "logic level" switching had caused the contacts to get crudded up. I rigged up a little circuit with an LED, so the switch would have to switch about 20 mA at 3 V DC, and operated the switch by hand several times. The LED lit up with the switch closed, but sometimes it flickered just a little. I also observed that if I was careful, I could push the lever far enough that the plunger would go down and the switch would click, but it wouldn't close the circuit. I had to keep pushing the lever just a little (overtravel) for the circuit to close.

Toyota doesn't give a spec in ohms for this switch; it just says "continuity" or "no continuity". The switch is a Panasonic AV3 series, so I got the brain wave to download their spec sheet on it: https://www.panasonic-electric-works.com/pew/eu/downloads/ds_av3_en.pdf (brief) or Download | Automation Controls | Industrial Devices | Panasonic (more information).

It says (page 6) that the initial resistance is supposed to be 100 milliohms max, and you are supposed to measure that by putting 0.1 A at 6 to 8 VDC through it. So I rigged up some more LEDs and a fixed resistor as a load, until I had about 0.1 A at 7 V DC. Operating the switch with this current running through it gave me voltage drops of around 40 to 100 mV when the switch was closed, which corresponds to about 0.4 to 1 ohm (or about 400 to 1000 milliohm) contact resistance. This seemed high.

It looked like there was an opening to the interior of the switch near the lever pivot point. I may also have been able to open up the switch body - there are a couple of tang-and-loop-looking things on each side, between the mounting holes - but I wasn't brave enough to try. I really wanted to hit the switch contacts with a can of air, or nitrogen, or Deoxit, or Cramolin, but since I didn't have any of those things, I applied a custom, biosourced mixture of oxygen, nitrogen, carbon dioxide, water vapor, and trace gases to the switch body at the lever pivot.

In other words, I blew in it.

Doing that reduced the voltage drop through the switch at 0.1 amp to around 2 to 4 mV, or around 20 to 40 millohms. That seemed to be a much more reasonable number. I tried the switch several times and this number seemed to be consistent.

I also (very gently) held the lever near the plunger with one hand, and bent up the "simulated roller" end of the lever just a little bit (maybe a millimeter or two), such that the ear or tang on the edge of the service plug bail would push down the switch just a little harder.

I put it all back in the car, cycled the ignition key several times, and no complaints. I drove back and forth in the driveway - OK. I then drove a few miles to buy a cheeseburger from Warren Buffett, parked and ate it, and drove back - OK.

So... microswitch problem solved, I think. I'm still waiting for a rainy day (which isn't likely until the end of next week) to see if the P3009 comes back.

Microswitch data: Mine has "Matsushita Made in Japan" molded into the body, underneath the lever. On the side, it says "AV39310 0.1A250V AC 00329T". The body is black plastic, the plunger is red, and the lever has a "simulated roller tip" - a semi-circle loop.

Breaking down the part number according to the data sheet, I think I have:

AV3 - AV3 series
9 - This should say what terminals it has, but the enum in the data sheet only lists 2 and 4-8. I suspect this means "special" - maybe Panasonic supplied it with the wiring harness and connector already on it?
3 - This currently means "long hinge lever", but that doesn't match the lever on the switch. The lever actually on the switch is pretty close to 4, "simulated roller lever".
1 - Should be operating force, but the current enum is only 0, 2, 4, and 5. 0 is 0.25 N and 2 is 0.49 N, so maybe this is somewhere in between?
0 - This is either the contact material or the agency approval, but again, 0 doesn't match anything in the data sheet.
Red plunger - CuNi alloy + AgNi alloy + Au-clad triple layer contact type

The "0.1A250V AC" is the voltage and current rating, and I'm pretty sure the other number is a date code. According to one Panasonic date code format, it might be 0-03-29, or March 29, 2000. The build date on the whole car is June, 2000, so this seems at least plausible.

It's possible that only the "AV3" matters, and the rest of it is a part number that denotes a special build. Panasonic built the batteries for these cars, so there was probably a big enough production run to justify a custom switch if they thought they needed it.

If I was going to try to order a "stock" switch to replace this one, I'd probably look for an AV324x61 , where x is 0 or 2. None of the usual suspects stock exactly this switch. Mouser stocks an AV3244613 (higher contact force) or an AVT3244613 (higher contact force, and the terminals are at right angles to the AV3 series) for about $3. Digi-Key stocks an AV3432619-A, $5 each, but with that one, you'd have to cut off some of the lever before installing it. RS Australia stocks an AV3232613 (cut off the lever), or an AV3832613 (cut off the lever, mod the terminals), or an AV343461 (cut off the lever) for AU$3-$5 each, and RS UK seems to have a similar range for about £2.50 to
£4 each. Allied Electronics, which is RS USA, doesn't seem to stock any Panasonic AV3 at all.