CarTalk.com Blogs Car Info Our Show Deals Mechanics Files Vehicle Donation

Radiator fluid not being sucked back into the engine when it cools

99 Toyota Avalon, 3.6 L engine, 160k miles
The engine sends radiator fluid to the overflow reservoir but doesn’t suck it back into the engine when it cools.
My mechanic replaced the radiator, cap, and hoses…no change. So. he put a chemical in the radiator fluid that is supposed to form a plug to stop a leak. After ~5 months of waiting, still no plug…no change.
Tomorrow (Tuesday, Aug 14), I’m bring it back to him. If he says it needs a new head gasket, do you agree? If so, what would be a reasonable cost? Also, this timing belt has 68,000 miles on, do you think I should have him replace it, too (since he is doing work on the engine)?
Do you think I am foolish to put the cost of repairs into this vehicle? The rest of the vehicle is good. We’ve had it since 2000 and have kept up with maintenance and services. We want this to be our son’s college car but only if it safe and dependable.
Your advice would be greatly appreciated–especially if it arrives before Tuesday, Aug 1th.

Is coolant being expelled from the radiator to the point that the recovery tank overflows and the engine overheats?

If you replace the head gasket, you should take that opportunity to replace the timing belt, although it is early.

BUT, if you have no evidence of water leaking outward or into a cylinder from the head gasket, there is almost no chance that you are drawing in air through the head gasket. A pressure test of the cooling system should confirm this before you start doing open-head surgery on the car.

When the mechanic replaced the radiator and hoses, did he replace the overflow tank? From the image in the parts book, it looks like two or three hoses run to your overflow tank. A problem with any of those hoses, or an internal crack in the overflow tank, can cause the problem you are experiencing. On a 13 year-old car, a crack in the overflow tank would be my first suspect.

Is your car overheating or showing any signs of overheating? If not, at this point I would just keep an eye on the antifreeze level in the radiator itself, and ignore the overflow bottle completely… Cars went decades with out one, and in most (not all) cars is more to catch antifreeze from leaking onto the ground then anything else any way. SOME overflows are pressurized (the radiator cap is actually the cap to the overflow), if this is the case on your car… Ignore my advice above.

I had a similar problem with a 1993 Oldsmobile 88 that I owned. The hose from the overflow bottle to the radiator was limp and would collapse when the car cooled and the radiator tried to suck coolant back from the bottle. Easy fix.

I had the same issue with an '03 Honda Civic. First I replaced the radiator cap, not the one on the overflow tank, rather the pressurizing cap on the top of the radiator. Didn’t help. Then a pressure test when the car was warm showed no loss of pressure. After sitting overnight when the car was cold then the pressure test showed a very slow leak at a hose connection to the radiator. The clamp was loose.

The reason the pressure test failed when the engine bay was hot is due to expansion and contraction of different materials. Before spending more money and replacing more parts I’d have another shop have a go at finding your leak.

There needs to be more solid evidence of a head gasket problem than you have now. Without coolant polluting the oil, or oil polluting the coolant I’d do more pressure and leak down tests before thinking blown head gasket.

Yep, do NOT get a head gasket until it’s proven to be the problem. There are a number of tests needed before you think about spending that kind of money. Did your mechanic actually do a pressure test?

As I suggested to someone else here about a week ago, get some type of suction bulb, an antifreeze tester would be just fine, pull the small hose off the radiator, just below the radiator cap and try to suck fluid up that hose with the bulb. If that doesn’t work, the problem is in the overflow tank.

There are two basic styles of overflow tank, one has the overflow hose connected to the cap of the overflow tank. On this type, there is a hose on the underside of the cap that hangs down into the coolant to suck it back on cool down. If this hose is soft or missing, that is the problem.

The other type, the overflow hose goes to a fitting on the tank itself. The “hose” that goes into the coolant is molded into the side of the tank. These can develop cracks on the inside of the tank, so there is no external leakage, but it won’t suck the coolant back into he engine. The tank has to be replaced.

The third possibility is the overflow hose itself. It may have gotten stiff and brittle and is not making a good seal at one end or the other. It may not leak the coolant out, but it might not be sealing when it is supposed to draw it back into the engine. The hose itself is cheap.

Thanks for all the comments–I couldn’t answer because I was at work. Here’s more information and answers to your posts:

Over several days of driving, additional fluid is added to the reservoir until the reservoir overflows. Even when this happens–and the remaining fluid level inside the engine is very low–it has NEVER caused the engine to overheat. The gauge has never exceeded half way between C & H, and there are NO symptoms of the engine overheating (smell, pinging, etc.).

I inspected the original hose from the engine filler neck to the reservoir, and it looked fine. But since the problem wasn’t getting any better, I replaced it (not with a Toyota hose, but with a generic hose).

The overflow reservoir has the integrated tube molded along the outside of the reservoir to which the hose from the engine filler neck connects–that’s the only hose connected to it. I have not replaced the reservoir.

I don’t know if the mechanic has actually done a pressure test of the system.

At this point I am suspecting the overflow reservoir based on your inputs. Althought this doesn’t explain one other thing that I didn’t mention: even when the engine is cold (like the next morning), when I remove the radiator cap from the engine filler neck (the pressurized cap), I can hear a suction noise. That mystifies me because it “sounds” like it has a suction, but it is not capable of pulling the fluid from the reservoir back into the engine.

If you have any other ideas based on my new information, please let me know.

Alan

You shouldn’t hear a ‘suction’ noise when you remove the cap, it has a one-way valve that should allow coolant to flow in as the engine’s cooling. To me that says radiator cap or collapsed hose to the reservoir problem.

Radiator cap replacement comes to mind as the first attempt at throwing parts at the problem.

The overflow not working properly isn’t a reason to get a new head gasket. You may need a new head gasket, who knows, but not for that reason. Good mechanics know exactly how to diagnose a bad head gasket. If you don’t trust your mechanic, seek out some recommendations for other mechanics.

What does your owner’s manual say about the timing belt? The safest thing to do is replace the timing belt at the intervals suggested by the owner’s manual. If the belt were to break at high speed, even with a non-interference engine which yours may well be, it could do considerable damage to the engine. That said, on my early 90’s Toyota Corolla, I replaced the timing belt the first time at 120K and noted it still had a lot of miles of wear left on it. How much lee-way you have depends upon the car, the climate where the car is driven, and the way the car is driven. But if your owner’s manual says replace at 60K, then the safest thing is to do just that. You might well want to bring all the suggested maintenance intervals up to date at this same time if you plan to keep the car. There’s no fundamental reason why a 2000 Avalon shouldn’t provide a lot more miles for your family. In the current economy, a lot of families are keeping their cars 10-20 years.

I’m assuming the overflow tank is separate from the radiator and removeable from the car. If so, what I’d do if this were my car is remove it and the hoses connected to it. I expect the hoses may be plugged up or have air leaks or are collapsing. You could try to diagnose each of the hoses. Me, I’d simply replace all the hoses and associated clamps with new ones. If that didn’t fix it, then I’d replace the overflow tank too.

The mechanic removed the front 3 sparkplugs and put a scope in there. 2 of the 3 were dark and had carbon deposits consistent with a 13 year old engine with 160k miles on it. The third was shiny clean. He said that was proof that radiator fluid was in seeping in there either through the head gasket or through a cracked head. So, they’re pulling it apart, sending the heads to a machine shop, and we will see what’s going on. I sure hope he knows what he’s doing.

Do any of you have experience with the scope, and does his explanation sound reasonable?

Coolant entering a combustion chamber can certainly make things shiny clean in there and a borescope can show this.
However, if this is the case then the lowly spark plug should also show signs of coolant contamination.

To each his own I suppose, but there are various tests which could verify a head gasket problem and probing with a borescope would not be something that I would do in a case like this.
Compression test, cooling system pressure test, hydrocarbon test, and vacuum test are examples and all could be done in 15 minutes or so.

Hopefully they’re correct.

Too late now but no way would I be putting $1-2000 in a Toyota that old with no other symptoms. That would have been a perfect case for Tester’s head gasket fix.

Since your overflow tank has the molded tube, and you detected a vacuum at the radiator cap, this could be a real cheap fix. You have debris in the bottom of the overflow tank that is acting like a one way valve. Remove the tank and flush it out. Be sure to force some water into the molded tube while the tank is upside down. You should be good to go after that.

You could also use compressed air to clear the molded tube while the tank is upside down if you have it available.

What is “Tester’s head gasket fix?”