3rd new heater core leaking - caused by heat? 2001 Caravan

On a 2001 Caravan (3.3) I initially misdiagnosed a bad heater core (coolant in the air plenum). Turns out a gallon jug of coolant sitting on the floor at one time splashed up there. But, I pulled out the old core and replaced it. I tested the old core and found no leaks, but replaced it with an aftermarket anyway. Two weeks later, the new one leaked at the joint between the tank and the core (plastic to metal) even at lower pressures. Pulled it out and went to reuse the original and lost an issue with the o-rings and assembled with a non-OEM o-ring. Leaked at the o-ring. Got correct o-rings for $23 at Dodge, continued. Went to reinstall and see in my messing around I had cracked the tank flange. Got a new core. Installed, 2 weeks later it leaked (same place). Put in another - same thing. Cap is good. System pressurizes into overflow tank and sucks back in. System flushed.
So
 this time I tested the thermostat and found it to be bad - never opens. I don’t drive the car - belongs to my son - so not sure what the thermostat read on the dash but here’s my theory:
New core uses marginal parts (you can’t buy OEM cores unless they are NOS - they stopped making them in 2020). Occationally, due to the thermostat not opening, the car overheats. The higher temp causes the seal between the tank and body of the heater core to fail - a deformed gasket, or maybe they use some kind of glue that fails at a high temp.
Does that sound plausible? Naturally, I’m going to replace the thermostat, but debating if I want to pay for a NOS OEM core on this beat up 2001 Caravan.
On the upside, I’m very good at replacing these now. Book time is like 7.5 hours and I can get it done in 3.

2 Likes

Suggest to ask at your local repair shops & parts stores if anyone knows of a good radiator specialty shop. They deal with this sort of problem every day, and will know what your best options are.

Do you test the pH of the coolant regularly? If not, good idea to adopt that habit. Best to not allow the pH to go less than 7 - 7.5. The more the engine runs, the more acidic the coolant becomes. Eventually the acid will start metal corrosion and damage the seals throughout the entire cooling system.

I don’t think I’ve considered a radiator shop in ages since they all have plastic these days, but yeah I’ll check that out.

Definitely not corrosion - fluid totally clean (flushed, replaced) and this is happening within weeks. Also, no head gasket issues (no exhaust in system).

Thanks,
-Jeff

A followup - it’s been a month and a half since the first post and it’s still working ok. So I think the culprit was the bad thermostat causing increased pressure/heat in the internal cooling loop that weakened the heater cores. It doesn’t help that all the heater cores available are nowhere near the build quality of the OEM which are no longer made. But the latest core is the same model as a previous one that failed, and it’s holding up.

I don’t like how the computer handles temperature swings - it lets it go quite high before kicking on the fan. Not sure if a new relay would fix that (as in it works, but only partially) or just live with it. I’d like to get a schematic of the fan relay - I think it has more than one loop/speed. The dash thermostat is amazingly inaccurate.

Thanks,
-Jeff

Has the radiator cap been replaced?

First of all, I am incredulous that the heater core contains plastic at all. Second, many vehicles–including Dodge Caravans from the 1990s and 2000s use the heater core as the coolant bypass when the thermostat is closed. To improve long-term reliability, I always drill a small orifice in the thermostat, which helps with air bubble removal, and allows some flow at all times. Third, aftermarket parts vary widely in quality from “at least as good as OEM” all the way to “cheap knockoff”.

Radiators have plastic tanks. A heater core is essentially a small radiator. What’s the difference?

1 Like

The radiator is easy to inspect and replace on just about any vehicle. The heater core is not. And plastic end tanks become brittle over time, and the rubber gaskets which seal the plastic end tanks to the aluminum micro-channel coil will eventually wear out and leak.

1 Like

Yes, cap was tested ok at 15lbs. At one point, I replaced it anyway with a non-flip-lever release cap which also tested ok at 15lbs.

I don’t know about that. The Caravan body style used to require removing the front bumper cover to replace the radiator, many vehicles require discharging A/C and removing radiator/condenser as an assembly.

If I remember correctly, the bread-and-butter Taurus used to pay more to replace the radiator than the heater core.

As you can see in the image - it definitely has plastic. The plastic part is the end-cap with the input/output tubes. This is the OEM heater core – I cut it to see how it goes together. You can see that the copper pipes are inserted through the cooling fins, metal header and a black rubber gasket which goes across the whole end. Then it appears they flare the tubes locking them on the metal header, and then the plastic ‘tank’ assembly is crimped on top of the gasket to the main body using heavy crimping tabs all around the parameter. There appears to be no glue used here. I don’t have an non-OEM one to cut open, but they all had aluminum tubing instead of copper and a lower density of fins and tubes. This original OEM pictured here did not fail in the car – I broke it all by myself by over-tightening the bolt and had to replace it.
I can only conclude that higher heat and pressure caused the seal to fail on the gasket with the non-OEM replacements, and that these non-OEMs were more cheaply made than the OEM (since the OEM did not fail in the same conditions). I would be curious what kind of gasket material is used in the non-OEM ones and if they used any sort of glue/silicone in the process.

The OEM thermostat had an extremely small hole to allow air to escape. It also had a BB check-valve on the hole, so if running, I would expect zero flow through the hole. The non-OEM replacement thermostat did not have the check valve, just the very small hole.

I certainly would have purchased a high quality or OEM core if possible, but I was unable to find one. Mopar no longer makes them. I did purchase one for 160 at one of the big auto parts stores which showed a pic of looking like an OEM (copper tubing, high density fins) but upon opening the box, it was a Spectrum - the same one I had earlier used that failed and retails for about 40 on amazon and I returned it.

My only recollection of chronically failing heater cores was on a Dodge diesel that seemed to be cured by installing a expansion chamber on the incoming heater hose.

I doubt you’ll find any glue or adhesive used anywhere in those cores. They rely on the “rubber” gasket and a crimp seal. I’d be willing to bet any of those cores would have survived under normal conditions. The OEM units likely have more margin in the design. Perhaps the gasket material is of slightly higher quality and has a durometer more closely matched to the design and application. But could this simply be a matter of timing? The original timeline of events is complicated. If the thermostat was bad when the original core was in place, you’d expect some significant events to occur if the van was driven while overheating. At any rate, the coolant expands when heated and the weakest point will fail first. Likely something is going to fail as the engine continues to overheat and the coolant continues to expand and there is no mechanism to allow the volume to be relieved. In this case, the gasket of the aftermarket core gave out first.