Vapor under the Hood?

2004 HONDA ACCORD
190,099 MILES • 3.0L • V6 • 2WD • AUTOMATIC
Avatar
NAED61
  • MEMBER
  • 6 POSTS
When it comes to cars, I know how to drive and that's about it so please forgive my ignorance.

I noticed there is white smoke under the hood, which seems to be coming from somewhere near the valve cover. I can't see the vapor with the hood down, and there is oily residue coating the surface facing front. I thought it might be the head gasket (see my previous posts). However: There doesn't seem to be any oil loss and the color is fine. The coolant level and color in the overflow tank has been the same (perfect) every week. The exhaust from the tailpipe is fine. Can't test compression, but acceleration is great. I ran the engine with the radiator cap off- zero bubbles, looks perfect. Temperature gauge perfect forever, no variation. No visible leaks of any kind making it to the ground. Considering age and mileage, car is driving great.

I am including a photo of the oily buildup and three videos of the smoking in action. Thanks in advance.

Dean
Jun 11, 2024 at 12:08 PM
Repair Safety Notice: This information is for general instructional purposes only. Vehicle repair can be dangerous. Verify all information, follow manufacturer service procedures, use proper tools and safety equipment, and consult a qualified repair shop when needed.
Advertisement
Avatar
JACOBANDNICKOLAS
  • AUTOMOTIVE REPAIR CONTRIBUTOR
  • 110,190 POSTS
Hi,

There is an oil leak that is getting on the catalytic converter and exhaust. That is smoke from the oil burning off.

Have you checked to see if the engine oil or transmission fluid is low?

Joe
Jun 11, 2024 at 7:26 PM
Advertisement
Avatar
NAED61
  • MEMBER
  • 6 POSTS
Hi,

Both the oil and transmission fluid levels are reading full. Also, the transmission seems to be shifting fine. I am only putting 15-20 miles per week on the car.
Dean
Jun 12, 2024 at 6:19 AM
Avatar
JACOBANDNICKOLAS
  • AUTOMOTIVE REPAIR CONTRIBUTOR
  • 110,190 POSTS
Hi Dean,

We need to determine what fluid is on the exhaust. Is the power steering fluid full? Is the engine coolant low? Something is leaking. Based on the location, I would have thought a valve cover was leaking oil onto the exhaust.

Let me know.

Joe
Jun 12, 2024 at 8:45 PM
Avatar
NAED61
  • MEMBER
  • 6 POSTS
Hi Joe,

Valve cover gasket sounds right to me. I think it's a very slow oil leak from a very lightly driven car. Even though the stick reads full, it easily could have been more full just after my latest oil change last Nov.

I fed the same info I initially gave you (including the pic and several stills from my vids) to GBT-4o, and asked for its most probable cause... It replied valve cover gasket. If an expert mechanic and cutting edge AI agree, that's good enough for me.

Do you think it's safe to run a few errands before I bring it in for repair next week?

Dean


Jun 15, 2024 at 2:28 PM
Avatar
JACOBANDNICKOLAS
  • AUTOMOTIVE REPAIR CONTRIBUTOR
  • 110,190 POSTS
Hi,

If that is what is leaking, it should be okay to drive. Just remember, oil can burn, so don't run it very hard. You don't want it to get really hot if the oil is reaching the exhaust manifold.

Also, be careful using AI. I've done a lot of tests and found AI to provide a universal answer for multiple things.

Let me know how things turn out for you. I'm interested in knowing.

Take care,

Joe
Jun 15, 2024 at 6:31 PM
Avatar
NAED61
  • MEMBER
  • 6 POSTS
Hi Joe,

Just wanted to let you know I brought my accord in to my mechanic- it was the valve cover gasket. Had both of them replaced; $438 + tax.

From your last reply:
"Also, be careful using AI. I've done a lot of tests and found AI to provide a universal answer for multiple things".

I'm curious which version of chatgbt you were testing. I've been playing around with OMNI, and I find the difference in both context and accuracy compared to 3.5 remarkable. I went back and prompted it with only the photo and 5 secs of video with smoke rising, and asked it to diagnose the problem with only one answer... it had no problem identifying it. True, this is a fairly obvious problem to ID. I wonder how good a model could become, trained exclusively on auto mechanical data? My guess would be really, really good.
Jun 29, 2024 at 10:32 AM