Lwerewolf

'09 gs450h, EU spec, ~280000km odo.

OK, let's make it weird from the start.
-Engine does NOT consume oil, biggest OCI was 10k miles with OEM 0w-20 (when I first got it - the previous owner said the maintenance menu in the car was correct)... decided that 10k miles is extreme for my conditions (Bulgaria - questionable fuel quality, bad air quality, 25km one way trip to work at a relatively constant rural/highway/speedway), now it's 10k km (~6k miles) on amsoil ss 10w-30 (which is about the only thing that I know that I can get locally that's extremely unlikely to be a knockoff, and is really full-synthetic... why 10w30 - it's practically a 5w (look at CCS), so probably zero VI improvers. At any rate, used to be ss 5w30 for two 5k(km) runs for cleanup and a 10k(km) redline 5w30 before it started being... hard to get here. New OEM air filter at every change.
-Engine does NOT consume coolant, the coolant itself is clear (in radiator, in inverter loop & in expansion/reserve tank), no residue, no color change.
-Engine does NOT overheat. The worst I've gotten (OBD temperature readout via hybridassistant) is 97deg. c (203f) after two consecutive very hard accelerations (0 - 140 - 60 - 140+, all km/h), during 30 degrees (86f) weather. Usual highway cruise temps are around 85 - practically thermostat opening temperature, this is at 120-140kph highway speeds in 30deg. c (86f) weather.
-Heating works fine, haven't noticed any funny coolant smells or residue or anything really on the inside. Again, there's no coolant loss in the first place...
-No frothy/milkshake/white residue/etc. on oil, dry exhaust pipes, no white/blue/any smoke other than condensation in winter (i.e. nothing right now, since it's summer), etc.
-No DTCs (other than the p0420/0430 due to midpipe leak, I'm on it).
-No misfires during any kind of load - that includes force-charging the battery when it's low (i.e. leave the car overnight on 2 bars, start it up, it starts force-charging immediately, no misfires).
-Highest recorded misfire count via Techstream was 10 or so per 650 ignition events, random cylinders but bank 1 was getting most of them.
-The misfires occur only in the first few minutes of startup, if the engine has been off for a few hours. The engine might still be "warm" (60 celsius while running for heat during winter), so I'm guessing either something else gets colder at a faster rate (so metal to plastic contact), or it's oil temperature/pressure-related, since that doesn't necessarily coincide with coolant temp... especially during winter.
-Fuel economy is normal.

One further clarification - OEM means Toyota/Lexus here.

So, from the beginning - I was chasing those misfires (basically the occasional light "shake" or disturbance on the steering wheel while "idling" for warmup or heating). First thing I did - new spark plugs (OEM), old ones looked fine and evenly worn... exception was #5, but as I later found out, pretty much all the blowby from the PCV is going to the 5th cylinder, so didn't give it much of a thought. I then changed the PCV... old one looked fine, new didn't change a thing. I then read about my symptoms coinciding with an intake manifold leak close to a cylinder (and the resulting trainwreck when the AFR is measured on the whole bank) - changed just about all intake gaskets (surge tank rubber gasket, plenum gaskets - whatever they're made of, PCV & breather rubber pipes & clamps), didn't help. Did the port injector seals, didn't help... but then I noticed something down the valley (i.e. between the two cylinder banks), and borrowed a cheap borescope from a friend... and saw this:


This is between cylinders 3 and 5, passenger side (LHD car).
If you were to look at the first picture in this article:
http://www.turbofast.com.au/4L_2GR-FSE_Engine.html
...this is right "behind" (from the picture's point of view) the sideways hole on the valey, right next to cylinder #3 (2nd cylinder from front to back on the left). I immediately thought "head gasket leak, then I started thinking...
1) Everything that was mentioned at the start - no signs whatsoever that there might be a problem like this.
2) The "dip" (not a "drip" - doesn't seem to have flown downwards at all) is small, and doesn't go downwards. The valley plate is mostly clear, save for a few even tinier drops that might be due to another reason entirely - look below:
3) I did change the inverter on the car almost a year ago, which has its own separate "pressurized" (not really, the cap is rated at ~0.35 bars) cooling system - own water pump, radiator cap, shares expansion tank with the engine on the 3GS model, separate on 4GS. Bleeding the air out of the system requires running silicone hoses from one end of the engine bay to the other, and I did spill some coolant (drops) back then - pretty sure I was cleaning "coolant drips/runs" over the intake surge tank (the big black plastic thing over the engine - I'd call it "manifold" but this name is used for just about anything from the intake tract for some reason).
4) The "dip" is conveniently located where the head & block meet (i.e. at the head gasket), but it's also in a place that you can look at in a straight line through the intake manifold - i.e. it's possible that I just "spilled" it there.
5) That whistling when the engine is hot - after highway runs, etc.
6) I googled some 2GR head gasket leaks, and... this is the smallest that I could find: http://carspecmn.com/toyota-and-lexu...aking-coolant/ - the one on the valley side, the other looks like it's been cleaned. Key take away points - still way bigger than what I have, considered "minor", way deeper crust formation & all that, hasn't spilled down into the valley.

Right, so next up I cleaned the area as much as I could without removing the manifolds again, so that I can monitor potential buildup - basically this much:



I looked for other crusty pieces in the valley and managed to snag a photo of this:



...by memory, these were on the other side of the valley, near the front, so I'm guessing something dripped down, formed a crust and couldn't stay attached to whatever it formed on.

I then got a radiator pressure tester and tested both the engine & inverter loops... and my '03 corolla. On the GS, the pressure went from 1.2bar to 1.0bar within 15 minutes (engine loop), the corolla went from 1.0 bar (cap opens at 0.9 so I used lower there) to 0.9bar in 10 minutes, and I couldn't get a good seal on the inverter loop radiator cap for some reason. I did these several times with the coolant temperature of the cars near ambient (the GS was 30c with 19c ambient on the last measurement), and also put the pump in a bowl of water - bubbles were coming out of it (every 20-30sec or so), but given that it should have some air in not-necessarily-airtight parts (the indicator & its plastic shield, the back side of the air piston, etc), I'm not sure if that means much. The almost identical drop rates on the two cars does make me suspect it, though (the corolla has no issues either, and it's way easier to spots leaks on a small 4cylinder in a roomy engine bay).

I also checked the "suspect potential leak" dip after a day of driving again - it was the absolute same thing.

My future plans so far are:
1) Leakdown test of all cylinders, just in case.
2) Testing the pump that I got for its ability to hold air - need to block the silicon hose somehow... a trip to the hardware store should solve it
3) Testing the pressure tester cap for leakage - current idea is to build a "bucket" around the water outlet (where the radiator cap goes) and fill it up with SLLC (or water) once I attach the cap, thereby submerging it... then I'll watch for bubbles.
4) Idle oil pressure test & replacing the oil relief valve, spring and so on - a member on the Bulgarian lexus forums apparently solved a similar misfire case on his rx400h (yeah different engine, still has oil-driven VVT-i though) by doing just that. I've always wanted an oil temp/pressure gauge anyways - that + a coolant pressure sensor would be nice.
5) Finally change the water pump, thermostat & SLLC - I got the parts half a year ago and I've only put in the new radiator cap, since I was reading up on vacuum bleeding the system. I'm probably overthinking this way too much.
6) Possibly get a used oil analysis done on the next change, to check for any trace amount of coolant - the closest supposedly good labs that I know of are in Germany... still got more to read on this.
7) Get a tester cap with a schrader valve fitting - my MTB's air shock pump is known good in terms of holding 250PSI over minutes (vorsprung corset air can - HV negative spring, need a lot of pressure ^_^), so it really shouldn't be leaking at all at normal radiator pressures.
8) Find out where that whistling noise is coming from (happens after long rides) - I'm looking at what tools I should be getting... I'm guessing that a stethoscope won't be enough - some ultrasonic hiss spotter thing or... something.
9) Attempt to assess whether I'm over-the-top OCD about this.

So to sum it up - something that "looks" like a tiny external head gasket leak without any other symptoms whatsoever (as far as I could find anyways), found while chasing occasional poltergeist engine misfires that happen only at the start of a warmup & only during idle (warmup & running for heating) conditions.

So... if anybody has any comments, advice or really anything to say on this, please go ahead. Much appreciated in advance!

Lwerewolf

Daily "update":
-Still looking for proper fittings to test my leak detection hardware.
-The "whistle" (that sounded a bit like a varying coil whine from a CPU's VRMs or something, to be honest) did not appear the whole day. I intentionally drove "hard" (well, harder) and didn't use AC. The inverter loop definitely needs bleeding, so I'm on it.

Lwerewolf

"Progress" pics:
Before (24th of August):


After some time (today, 9th October):


Today's shot was a night shot, hence the difference in lighting in the areas not well lit by the camera's built-in frontal LEDs.
At any rate, no "progression" of the "potential leak", so I'll just forget about it for now. I'll pull the VVT-i OCVs for inspection tomorrow, resistance across the terminals is between 10-11 on all when the engine is hot after some highway... we'll see about that when they're cold.

Lwerewolf

OCVs do not stick, resistance is the same all across (8.2ohms when "cold").
Swapped coil packs between banks, didn't change a thing.
Compression test when cooled down somewhat (~40deg. coolant temp) gives between 135 and 145psi on all cylinders, so that appears to be fine...
Borescope down the combustion chambers reveals nothing unusual.
Intake valves are almost pristine, as expected given port injection + a 1:500 tc-w3 mix (ravenol watercraft vollsynth).

I'll play around with the injection way next, if I manage to force it.

Lwerewolf

Couldn't force the injection way for some reason (via techstream), followed prerequisites and still no go. At any rate, more observations from today.

First a reminder - this is a gs450h with the hybrid transmission - the engine is practically always mechanically connected to the transmission, differential, axles, etc.
Here is a good overview of the mechanicals:
https://gr8autotech.wordpress.com/20...ower-splitter/
Main differences in the gs450h (and ls600h):
-MG2 is connected to the output shaft (the one going to the differential) via a ravigneux gearset - basically a two-speed automatic.
-The differential is connected via the external driveshaft, which just so happens to have a bushing.
-The PSD is basically like an open diff - the engine's power goes wherever there's least resistance - either to spin MG1 or the wheels (and consequently, MG2, since they're locked unless MG2 is changing gears). MG1 modulates the engine's RPM by "pushing back" on the engine, either by generating energy from it, or by actually pushing on the engine. MG2 pushes the wheels with energy generated from MG1, or regenerates energy so that MG1 can "push down" on the engine, forcing its power to go to the wheels (incidentally, it also propels the car forward this way as well). The battery can supply extra power (up to 30kw - 40hp), recover energy that would be... "wasted" otherwise, or just power the car when the engine would have to idle otherwise. You should now be able to understand why "total system power" in these vehicles is not the engine + the two electric motors combined. Very detailed info here: http://prius.ecrostech.com/original/...OnAsIDrive.htm

So now...
-Commanding an AFR change at both ends (-12.5% to 25%) doesn't change a thing - in fact, the engine idles like velvet (albeit very briefly) at 700rpm while the idle speed adjusts. Same thing at 1250-1500. The "rattle" occurs even at forced 2500 (full throttle in park).
-The "occasional rattle" does occur in maintenance mode with a fully warmed up engine... while in park.
I then noticed that the clunking noises sounded kinda like they were coming from the transmission, and then it hit me that it might be internal "looseness"... or actually the bushing between the transmission and the diff being worn.
I then put the car in reverse with the handbrake still on, and pushed on the accelerator as much as I could (to "preload" the bushing). I let off the gas, held the brakes, switched to park... almost undetectable occasional shudders or whatever. Preloading it forward had a similar... "change" in effect.
This would explain why the car would not do this under any sort of real load - even if it would just be charging the engine via MG1.

Next step - looking at the driveshaft while the car's in the air, attempting to reproduce this by holding it in place/preloading it one way/letting it go. Another thing to think about - maybe it's just gear slap while the transmission is not warm... thought that'd be excessive. At any rate, food for thought.

RamAirRckt

Fantastic post! I have a similar cold start misfire. I gave a leak in my drivers side LHD muffler and it’s pretty noticeable at cold start. Now that I’ve patched up the exhaust leak it’s harder to hear but still present.

I haven’t done my first oil analysis yet and I have a sample that was in the car when I purchased it to analyze.

I’ll be likely taking the intake off this spring to inspect for coolant leaks. I have a NA Spec 2008 GS450h.

no leaks not frothy coolant or oil. Good fuel economy. But a slight misfire on cold startup.
I haven’t heard hissing yet or not noticed it at least.

thanks!

Lwerewolf

Hissing is from the AC in hot weather, specifically when you turn it off. Haven't changed the AC moisture filter yet, tech has enough work on his hands to bother with such things on my vehicle... hopefully this summer?

Just in case somebody's still landing here and getting all scared, update (tl:dr don't be scared lol):

Car is now >400,000km. I had coolant loss... well, very slow evaporation from the expansion tank... due to a bad radiator cap (of the disintegrating variety, it was OEM ^_^). Symptoms are - if the expansion tank levels fluctuate >2-3cm between hot and cold, you might want to check it. Other than that and the very annoying part of fishing it out (had to remove the thermostat housing to get to the last small part since I decided to be cool and didn't use a magnet from the get go), nothing different.

Cold start or hot start, "idle" (when forced) can be shaky or silky smooth. Fuel trims are combined <5% across, except in a very narrow load range in which bank 2 is slightly higher (7%-ish, saw it once). Compression test that I did very recently showed between 125 and 150 across the board, every cylinder could hit 150 but due to the way that the car "cranks" (revs for less than 2sec effectively), you may have to try multiple times in order to get the highest number - and by multiple tries I mean releasing the pressure after every try - it's just a gamble on how many times it'll compress the cylinder in question. TLR - no issues there whatsoever.

Re: how to do the compression test - I took out the fuel pump fuse, ran the engine until it went dead by itself, then tested the "hold the gas pedal to the metal while starting the car" method - it'd "crank" (sounded just like the engine idling) for a very long while this way. Once I disconnected the throttle, coils and all the other things that you have to remove in order to do a comp test properly - the car would "crank" for maybe 2 seconds tops - hence why you need to try several times in order for the cylinder you're testing to compress the "right" number of times. In my particular case - all cylinders hit 150 cold, didn't bother with hot because 1) it takes quite a bit of time to get to a state where you can test compression and 2) since the cold numbers are this high and consistent as hell... no point. Bonus points for not having to wait for the heads to really cool down before retorquing the spark plugs properly.

Re: the famous in some parts of the world 5th cylinder wear question that might be part of my worries, I don't have a borescope with a mirror that can fit through the spark plug holes, but this should be enough for you to decide whether it may or may not happen with bad maintenance and/or a leaky filter:

That might also explain why the engine is consuming a little over a liter per 10,000km (my OCI) - this is in the summer (we see 30+deg) with Ravenol SFE 5w20. No, it doesn't consume (noticeably?) less, yes, it does seem to be quieter when cold on cold days here (say, -18c). And yes, the clean runner is for the 5th cylinder - if you look at the plenum, you'll figure out why all the "stuff" goes there. If that "stuff" includes particulates that got in the airstream past the air filter somehow - the 5th cylinder's bore and valves will probably be polished very nicely, and the compression will very likely suffer. I run mahle filters, since their air filter for this application does have a rubber seal around it, unlike the OEM denso one. Why mahle - they're simply widely available here and a reputable brand. Any other good filter with a rubber gasket at the borders should be more than fine.

Now, as to why the idle might be randomly shaky or smooth as hell, my stupid idea beyond the injectors being finicky is:
1) I've seen no non-hybrid 2GR-FSE do this on videos
2) Just about every 3rd gen gs450h "cold start" video exhibits a bit of this - the slight burple that you might hear occasionally
3) The PCV valve is unique to the 3rd gen gs450h gws191 and the crown hybrid gws204, the only other notable difference that I can think of are the intake vvt-i actuators (3deg. more retarded at full advance, 60deg. range vs 40deg. on non-hybrid)
4) The shaky idle somehow seems to sometimes depend on how the car is tilted - oil pooling or something?!?!?!
5) When it's the most exaggerated, the techstream cylinder misfire counters tend to increment for the 5th only, or at least predominantly.
6) A fuel cut test to the 5th cylinder produces a few orders of magnitude more noticeable "misfiring"

So the next step in my hypochondriasis, if it triggers again, is a catch can for the occasionally shaky idle - and very likely chains/guides/vvt-i actuators/cogs, since the car is, well, of age. Stupidly reliable vehicle if you don't mind this occasional BS that nobody but me seems to notice. Hybrid battery is at less than 0.2v difference between modules when discharged with AC/lights/everything in neutral past 40% SoC - rebuilt ~160,000km ago with practically brand new gen4 cells out of C-HRs. Might swap the brake actuator preventatively just in case.