Overheated on track - now what?
 

Overheated the car at Mid-Ohio during HPDE. 7th run of the weekend, 3rd of the day; ambient temperatures around 80F. Coming out of the esses (medium-speed, hard right turn, slightly uphill) in 3rd gear around 5000 RPM the low coolant buzzer went off and the temperature gauge started climbing. Let off and limped into the pits with the heater blasting but the gauge still got to "H". Spent no more than 5 minutes at that condition and did not exceed 3000 RPM during that time. Shut down and there was coolant boiling and steaming out of the overflow bottle. Once cooled off added water to the radiator - didn't take too much initially but the low coolant buzzer kept going off so kept bleeding the system and adding more water. Made the 2.5 hour drive home with no overheating issues but at lower speeds and/or tighter corners the low coolant buzzer would sound and then go off after a while. Added more water last night and drove the car to work today with no buzzer or overheating.

Some observations from before/after:
-Cooling system was flushed June 2010. 60/40 distilled water/anti-freeze mixture with a touch of Redline Water Wetter.
-Since Friday night there's been a high-pitched squealing only during low-speed deceleration. Sounds like belt squeal but goes away as soon as the clutch is pushed in or I get back on the gas. The "adjuster bolt" on the airpump sheared off so it is being held in place by the "lock bolt". Belt seems tight.
-Prior to the run was more coolant in the overflow bottle than usual.
-Prior to the run I had a hard time starting the car. Did the flooded engine start procedure and it limped to life with a little backfiring but idle was steady once warmed up.
-I changed my shifting plan that session - I only went into 4th at one point on the track, rather than three. Engine speeds were almost always above 5000 RPM, but never exceeding 7500 RPM.
-This morning it started a little harder than usual and the engine sounds a little "different" during decel. If in neutral and revved it almost sounds like a 'hissing" noise as engine speed drops but it goes away when it hits idle speed.
-No white smoke, no power loss or any other driveability issues that I can discern.

I know I need to flush and fill the system again - need to get the tap water out of there. Next will be an attempt to replace the sheared adjuster bolt in the airpump and then replace all the belts, but then what? Inspect/replace thermostat? Waterpump? Why did the low coolant buzzer keep coming on and going off? Most importantly, is the engine fried?

For the people who track their FC's is this a result of water pump cavitation? Time for an underdrive pulley of some sort? Time to upgrade the radiator?

I hate to say this, but it sounds quite similar to a soft seal letting go. You might want to let the engine sit for a day or two and check to see if you have any pooling coolant in the combustion chamber.

When mine let go I could easily start it up and run it around town, but as soon as it hit operating temperature the engine would start to run on one rotor. When temperatures got warmer the car got harder and harder to start.

Also pay attention to dropping coolant levels. You may have a link somewhere (whether it's by the water pump housing, a hose, or internal of the engine).

it was opposite for me when i had bad coolant seal. it would run on one rotor at start, but after about 15 seconds or so it would run like normal. white smoke on startup, from the coolant puddling in the chambers burning off. that and i could fill the overflow container with a couple 3rd gear pulls lol

but yeah i also have overheating problems on the track. it will start creeping past 1/2 way after about 10 minutes of track time, even in cool weather. im still running a tmic, so the radiator should be getting plenty of flow. it was suggested to me to use aluminum tape and seal off all the parts around the radiator that air could get past, but i have not tried that yet

I assume I check this by pulling the EGI fuse and a spark plug and cranking the engine to see if anything comes out? Spark plugs are also something I need to check.
I drove 2.5 hours back home yesterday at speed >70 mi/hr and I'm positive I didn't loose a rotor, or any power for that matter. If I do have a failed coolant o-ring or rotor housing, it must be very small. The weird thing about the hard starting is that it was easier to start it this morning (ambient temperatures ~65F) after sitting all night than it was yesterday afternoon (ambient temperatures ~80F) after only sitting for ~3 hours.
I have no found any large external leaks, but that doesn't mean much. When I flush the system this weekend I'll get a much better idea of what kind of shape everything is in. As far as I know the radiator, cap and water pump are all original OEM equipment. I'm pretty sure the thermostat was replaced at some point and if it was I'm almost positive it is not an OEM part.

In the meantime I guess I'll just keep an eye out on the coolant level. The coolant level sensor seems overly sensitive to me, but it hasn't come back on in a while. If nothing changes I can only conclude the conditions simply outmatched the car's 22 year-old cooling system and some upgrades are in order.
Would the white coolant-smoke and smell make it past two catalytic converters? Aside from Magnecor spark plug wires and a K&N drop-in filter my engine is all stock. The exhaust was replaced after the manifold at some point but it looks to be a stock replacement, not a performance modification. I only ask because I see no smoke and don't smell coolant at the tailpipes.
I was about 20 minutes into the session - literally the last lap before cool-down. As far as I know, this is the only time the gauge crept past 1/2 way during all of the sessions. When I heard the buzzer I looked at the gauge and it was just then starting to move from 1/2 way to "H" - it's not like I ran a full lap at WOT with the gauge pegged at "H".

If you were closer I'd lend you my bore scope (makes checking things really easy). But yes, you can do it that way too.
Which makes sense if you have a casting give way.
Sounds like a plan.

This should tell you if the engine is bad or not.
No sense guessing everything else unless you get confirmation...


-Ted

When it stops raining this is my plan - I'll even post pictures when I'm done. I suppose with coolant/water I'm looking for a white/yellow fouled plug?

The car started and ran just fine after work yesterday. No buzzer, no overheating, no weird noise (other than loose belts). Letting is sit for today and I'll see what it does tomorrow morning (pending plugs).

Is it possible the slightly loose airpump belt was not providing enough friction on the water pump and it was slipping at higher RPM?

Not necessarily...

Typically, if coolant is leaking into the chamber, the spark plug is going to come out clean - it's like if you're running water injection.
If you have the leading versus trailing plugs next to each other, it'll be obvious.


-Ted

Well I'm pretty sure the engine needs a rebuild, but more on that later. First, spark plug pictures. Rotor 1 first, then rotor 2. Leading is on left, trailing is on right.

https://lh4.googleusercontent.com/_7...-41-58_163.jpg

https://lh4.googleusercontent.com/_7...-50-17_210.jpg

So I'm not sure how to read those but I don't think it matters. They weren't wet or anything but you can see the white/brown build up.

Per some suggestions I pulled the EGI fuse, removed the radiator cap and had my wife crank the car with the throttle open. After a few bubbles the fluid level began to rise. After putting everything back together and driving around the block for a bit I noticed that I'm getting a good amount of bubbles coming out of the overflow tube into the overflow bottle. Occasionally there would be periods where some amount of coolant would come out of this tube. Additionally, the upper radiator hose is "hard" when the car is running. Right before I turned the car off the low coolant buzzer came on again. Correct me if I'm wrong, but this all seems to indicate at least a failed water seal and the combustion gases are partially pressurizing the cooling system.

The spark plugs don't look like there's any obvious sign the engine is burning coolant.

With that said, if it's a small coolant leak into the engine, it will not show on the plugs.
Also, failure of the water jacket o-ring could make it possible it won't show on the spark plugs.

The low coolant buzzer triggering could be just because you opened the cap and compromise the cooling system - allowing some air to stay trapped and then trigger the buzzer.

I'm trying to be optimistic at this point. :)

The only way to tell is to drive it a significant distance and then check the coolant level to see if it significantly dropped (or puked out the overflow).


-Ted

I agree with Ted on this. The coolant level sensor is pretty sensitive, and when you flush & fill the cooling system, it takes lots of "burping" to get all the trapped air out of the system.

If you can get your hands on a cooling system pressure tester, you can do a quick pressure test to verify if you have a leak or not and find it. Basically this is a quick pressure leak down test - remove the radiator cap, replace it with pressure tester's cap, pump it up to the recommended pressure and watch its gauge for 10~15 minutes. If the pressure drops, you've got a leak - external leaks will be found wherever you see coolant peeing out; internal leaks may not be as easy to find, but if the pressure drops you've got a problem.

Check with your local auto parts chain shop or tool rental shop - you can probably rent a pressure tester for a reasonable cost; just have to make sure they have one with the right cap to fit your radiator.

After the first "event" I filled up the radiator and managed to drive 2.5 hours home averaging 70 mi/hr with no overheating, but the buzzer would come on occasionally during tighter turns and lower speed driving. Once I got home and the car cooled off I topped off the radiator and bled the system the best I could. At this point the radiator was completely full and the overflow bottle was at the full level as well. I was able to drive to and from work (only 8 miles) with no issues on Monday and I let the car sit most of the day yesterday. Yesterday afternoon, after driving more recklessly than I'd like on the street, I came home and revved the engine in neutral in the driveway. No buzzer while doing all this so I turned it off and let it sit for about 30 minutes. Before starting it up again I noticed the overflow bottle was now ~1" above the "full" mark. I made it to the end of the driveway when the buzzer came on. Since it was still fully warmed up I just pulled back up and turned it off for good. The buzzer did go off as soon as I started to move back up the driveway.

What has me most worried is the constant bubbling I'm getting in the overflow bottle. To ensure it wasn't be caused by engine vibrations I removed the overflow cap and tube and placed it in a cup of clear water. With the car fully warmed up there was a constant stream of tiny bubbles coming out of the pickup tube. Only once did a moderately large amount of coolant get burped into the cup. After that it was back to the bubbles.

Pressure testing is my next step before throwing new parts at it (rad. cap, thermostat, water pump). The plan is to swing by an auto parts store after work to see if they have anything I can rent.

The one thing I can't seem to explain is the original overheating event. IIRC, the cooling system holds like 2 gallons of water and I know I didn't add nearly that much after it overheated. The fact the buzzer came on just as the temperature gauge started to rise also has me stumped considering it has come on many times since that and the temperature gauge hasn't moved.
I am a self proclaimed pessimist and worry constantly, but I'm trying!

The 1" extra coolant in the overflow bottle 30 minutes after a hot shutdown sounds about normal. When you shut down the engine the radiator cap allows pressurized/expanding coolant to flow into the overflow tank, raising its level. When the system cools down enough, the resulting vacuum should suck the coolant back into the radiator, leaving the overflow bottle around the full mark.


If air is getting into the system, from either a leaky radiator cap or anywhere else in the system, it could explain the bubbles you're seeing. And it would explain the added difficulty in getting the system fully "burped" and keeping the buzzer quiet.


If you're going to throw parts at it, I'd suggest getting a new Mazda radiator cap & thermostat regardless, especially if you don't know what's in there now. Don't bother buying a Stant or aftermarket thermostat - for some reason the ones the make for rotaries are crap.


If your system is all stock, the cooling system is barely sufficient for track duty - it has very little excess cooling capacity, and a just a little bit of pump cavitation due to the prolonged high revs can easily boil things over. The buzzer sensor needs to be immersed in coolant at all times to keep it quiet - just one air bubble under it and BUZZZZZZZZZ. The temperature of the coolant has no effect on the buzzer - it's a level sensor.

When I get home I'll check it out again to see if it went down. I know the radiator pulled a good amount back in at the track as it cooled down, which isn't looking too good for the "there's a leak somewhere" theory.
I have noticed there is a tiny tear in one of the rubber seals in the radiator cap. I though it too small to make a difference but who knows. I always figured with the car running the pressure generated by the water pump would be more than sufficient to keep air from being sucked in, even at idle. If air is getting into the system with the car off, I'd expect to see the bubbles stop at some point as well.
This I know. The rad. cap is the original I believe, but I'm almost certain the thermostat was replaced at one point (I have the records from the previous owners) and it's likely it's not an OEM Mazda unit. I'll do the pressure test and if that passes, change the radiator cap and go from there. If it fails, well I've got some exploring to do...
Yep, cooling system is all stock. I had no issues during the other runs over the weekend but I did stay in 3rd gear a lot more for this particular run so perhaps that was the proverbial straw. Aside from an upgraded radiator what other modifications would bring the cooling system up to track-ready levels? It's not a track-only car and I maybe drive it 25% of the time. I've seen underdrive main pullies but those make me nervous with the stock alternator's output.

Is there any way to "officially" bleed the S5 cooling systems? I open up the heater, turn the car off, then fill the radiator and overflow bottles. I squeeze the upper radiator hose to get as many air bubbles out as possible.

Definitely replace the cap with a new Mazda cap then. If it's not sealing properly, it won't hold its rated pressure, and this in turn will make it more likely to boil over.

If it's not a Mazda T-stat, I'd just replace it. The Mazda T-stat & gasket for it is pretty cheap, like under $15

Besides an upgraded radiator, just make sure the fan shroud and all the ducting is in place & functional, including the plastic undertray. Key point here is to make sure that air only flows thru the radiator & oil cooler, and not around them. As for the fan, it only matters when you're at idle or driving at low speeds, but just check it per the FSM procedure to verify the clutch on it works - when it's hot, fan should turn but resist freewheeling when you try to spin it by hand. The underdrive pulleys will help reduce waterpump cavitation at high RPMs on your track days, and if paired with a similarly sized underdrive alternator pulley (i.e., to keep the drive ratios the same/close to stock pulleys) would solve any alternator output issues. Anything else is just maintenance - flush out your system regularly, use fresh coolant & distilled water at proper mix for your climate, check/replace all hoses as required.

Not that I can find in the FSM. What I do whenever I flush & fill the system is make sure the heater is set to full heat before starting, and after I do the initial filling of the radiator with coolant, I'll let it idle with the rad cap off. When the engine reaches operating temp and the upper hose is hot to the touch, I'll run the engine in neutral at a constant 2~3K RPMs for a couple of minutes, then I'll quickly rev the engine up past 5K & back down to idle a few times, still with the cap off. When the level in the radiator drops (it will as trapped air pockets escape), add more coolant till its full. While you're doing this drill, the coolant level buzzer will sound off from time to time - it's normal. I keep repeating this process until I can't add any more coolant to the radiator. Then I top off the overflow tank with coolant to the full line, put the radiator cap back on, and take it for a drive. Let it cool down fully, check the level in the overflow tank, add coolant if needed to get it back to the full mark.

I'll take a picture of the tear and post, but you're right, might as well replace it as they're only like $10. It looks like regardless of the condition of the engine I'll be needing a new thermostat so might as well pick up one of those as well.

That all makes sense, but the problem I'm having is with the car on and the radiator cap off I get a lot of coolant coming up and spilling out. Maybe there was just a large air bubble pushing it all out but the pessimist in me says this is the pressure from the combustion chamber. Maybe another picture will suffice...

I forgot to mention that I remove the radiator cap after the coolant is at operating temp and circulating thru the engine - this helps avoid most of the spilling, since once it's all circulating, the level in the rad should drop a bit. But yeah, if you remove the cap while its still cold and run the engine with the cap off after just filling the radiator, lots of coolant will barf out. Just use a thick rag to keep from burning yourself, crack open the cap to the first stop to release the pressure, let any excess coolant bleed out then remove it all the way.

I did this at the track before heading home and a good amount of coolant came spilling out. Kind of like a slow geyser. I didn't leave the cap off long enough to see if it was going to stop though so maybe that's normal?

So more updates. First a picture of the radiator cap. The tear is hard to see but it's "inside" the "ear" of the cap toward the bottom:


https://lh4.googleusercontent.com/_7...-16-50_616.jpg


I managed to rent a cooling system pressure tester from the local Autozone as well. First the good news. I checked the radiator cap - it opened at the appropriate 0.9 bar (about 13 psi) but it would not hold any pressure above ~9 psi and even then it still fell slowly. Outright, this makes me believe the radiator cap is dead so there's at least one thing.

Next, the not-so-good news. I attached the pressure tester to the radiator and pressurized the system up to 15 psi. Within 10-20 minutes the pressure had dropped to 14 psi or so. I brought the pressure up to 20 psi next to see if it fell off any faster but it dropped ~1 psi in about the same amount of time. I did not see any external leaks, so something's got to be off internally. Then again, perhaps the seal on the tester wasn't perfect and the engine's fine as I'm not detecting any gas or exhaust fumes in the overflow bottle or radiator

I see the tear - it's trash; replace it.
The system has to be airtight - a tiny leak like that will compromise it's effectiveness.

15psi is about as high as I would go.
20psi is way too much pressure - in fact, most cooling system will start to leak all over the place.
Remember, the stock radiator cap is only rated to 0.9 bar(?), which is like 13psi.

15psi drops to 14psi is okay in my book.
This pretty much eliminate a major water jacket o-ring failure.

Replace the cap...
Burp all the air out of the system...
I think your engine should be okay...
Cross fingers!


-Ted

New cap already ordered!
Yeah, I was a little nervous about the 20 psi. My hope was I'd expose any small leaks by really forcing the pressure. I did not start the ar yesterday, but I will tonight and I'll be sure to look for any white smoke from the exhaust indicating I managed to force water into the combustion chamber.

Yep, it was a VERY slow, seemingly unsteady leak. Not ideal, but like you said, certainly not a major o-ring failure. It probably would have continued to drop some, but I was running out of daylight...
This is the plan, along with a new thermostat. I'm considering a new water pump and belts as well because I'm pretty sure they're the original parts.

-Ted[/QUOTE]

Well, the new radiator cap didn't seem to do anything. Somehow there is still a constant stream of air/gas going into the overflow bottle. Every now and then some water comes with it so pressure is being built up somewhere. Additionally, when I remove the radiator cap with the car running and warmed up it just keeps bubbling like the system never seems to be fully bled. If I add more water while it's running it seems to just want to push that and more out.

I have a new radiator, coolant hoses and water pump on the way (they are probably do to be replaced anyway) but I'm not thinking any of those are failed. Time to start pulling things apart now?

Come to the car show...
Borrow tools...
?????????
Profit!

Something tells me a rebuild will take more than a weekend and besides, The Mitty > CCVT car show.

I'm going to try one more bleeding technique as suggested by the Haynes manual, which involves running without the radiator cap with the car cold. If that doesn't do anything then I guess it's time for a rebuild. I can't think of any other reason why I'm getting a constant stream of gases, occasionally accompanied with coolant, flowing into the overflow bottle other than a water jacket o-ring or housing failure.

Bottom line: is there any way this could be attributed to a bad thermostat or leaking radiator/water hoses? If not, like mentioned above, it looks like rebuild time.

Not the thermostat, since it's an internal part. But in theory, if water can leak out of an extrenal part exposed to air such as a hose or radiator, then air can get in too. But you tested the system with the pressure tester and it passed, most likely you would have found a leak. Only exception to that I can think of is since the pressure test is done under static conditions (engine not running, at whatever temperature it was at when you tested it), the test could miss a leak that only occurs under some combinations of dynamic conditions (i.e. temperature cycling, vibration, etc.).

The only other possibility that I could think of would be that the water pump housing is starting to leak. Whether it's from the impeller or from mating surface would be difficult to say. It wouldn't have to be a large leak. When I did my rebuild that was an issue I had to overcome.

Other possibilities might be cracked coolant lines that run into the heater housing (let alone the heater might be leaking as well). Basically you'll need to track down to see if there's a pin hole leak anywhere in the car.

Or you could just shove a borescope (or a mirror) into the exhaust ports and see if there's any coolant pooling.

Both valid points. What seems to be missing though is any sign of an external leak. I'm not seeing anything from any of the lines; nothing from the radiator and nothing coming out of the "weep hole" on the water pump housing. By "signs" I mean no leaking fluid, no evidence of dried fluid, no steam while running, no coolant smell, etc. Also, if there were a small leak allowing air into the system wouldn't that air just bleed off and stop at some point? The system is pressurized when the car is running so wouldn't that prevent additional air from entering?

I'm not sure if I'd say the pressure test "passed". I brought the system up to 15 psi and within 10-20 minutes it had dropped to 14 psi. You would think a perfect system would hold pressure indefinitely.

I still think I'm missing something about the bubbling in the overflow bottle. Something is causing a pressure build-up in excess of 0.9 bar but it's not a constant stream of fluid that's coming out like what would happen with an over-filled system. It's a mostly steady stream of gas, whether it be escaping air pockets (but they never stop), boiled fluid (possibly a result of a stuck thermostat) or exhaust gases I'm not sure. I suppose the only way to tell is to have the system checked for the presence of exhaust hydrocarbons but I don't know where I'd go for that.

I have a new thermostat that is not yet installed and a new radiator, water pump and full set of hoses on the way. Is there any hope that installing all those would rectify the issue? If not, I guess I'm just going to have to tear the engine apart.

Sorry to hear it's not going too well...

Have you tried to smell the coolant (bubbling)?
If it's combustion gases getting into the cooling system, the smell will give it away.
Normal coolant smell is a lot diffferent from combustion gases in the coolant.

There is a chance that you have a very tiny break in the water jacket o-ring on the combustion / exhaust cycle of the rotor housing.
The combustion pressures can hit around 1,000psi with the exhaust cycle pressures several times higher than the typical 15psi, and this would explain what you're experiencing.


-Ted

No. There is no check valve in a pinhole leak if it's present. If water gets out when at pressure, air will come in when there isn't. It's basic buoyancy and pressure differentials. The system will suck in air wherever possible as it's less dense and requires less energy to reach equilibrium pressure.
Coolant rarely ever is a constant stream coming out. You'll have some gas, but this is usually just gaseous coolant vapor. If however the system has air when it's coming back in from the reserve tank you have issues.
No. What you might be able to do to confirm where your issue is; would be to break the coolant system into sections. Separate the radiator from the engine section and test each section. For instance you can block off one outlet on the water pump and pressurize the other. Check pressure after a few hours. Do the same for the radiator and you'll know which one's holding pressure and which one is not.

If your radiator isn't holding pressure because the brazing is falling apart that might be your issue, but if you haven't noticed leaks with it.... then well... That's what I'd do at least.

You forget you're talking to another engineer here (and a VT AOE graduate to boot), but that is what I was getting at. If air entered through some small tear with the car off (under vacuum) it would be pushed out with the car on. The air would be pushed through the path of least resistance - the open radiator cap. However, because the car is on no additional air can enter so at some point the bubbling would stop. What I have is a constant stream (continues for more than 10 minutes) so this leads me to believe it is not the issue.
I'd love to pressure test every component but I just don't have the means.

I'm inclined to believe ReTed that there is a small failure point likely on the compression side of the rotor housing. This would explain why I don't have a continuous cloud of white smoke because the pressure is keeping the coolant out, but would explain the little bit I get after starting the car after it's sat for a while. It's always done that but I just assumed it was condensation. I can think of nothing else that would cause the continuous build up of pressure. I guess the HPDE just pushed it over the edge. I was really hoping the stock engine would last more than 105k miles...

It was more for other individuals reading than for you :p. Do the bubbles correlate to the heat increase? Or is it independent? IE if you run the engine; reach operating temperature; bubbles stop or bubbles continue but operating temperature is maintained.
And you call yourself an engineer :p. Hit up Lowes/Home Depot and build your own for less than $50.00.
If it's how I imagine it is working than that small failure point is an indication of the casting failing... which 'shouldn't' happen on an S5.

So I went out and fiddled with things again; this time I took some video. I don't know how to embed these so direct links will have to suffice.

First, here's the behavior when cold:
https://picasaweb.google.com/lh/phot...eat=directlink

Next, fully warmed up:
https://picasaweb.google.com/lh/phot...eat=directlink

Last, here's after revving the piss out of the engine. Notice the coolant level:
https://picasaweb.google.com/lh/phot...eat=directlink

Just for giggles, I think I may have found an exhaust leak. Yes, I know I should not be revving the engine like this:
https://picasaweb.google.com/lh/phot...eat=directlink

I don't know... It really seems like a fire ring has decided to let go.

Yeah, it wasn't too happy starting this morning. Basically, I've already started looking for engine stands, hoists and other tools needed to pull the engine. Likely going to start the teardown this weekend.

Ouch, after looking at the videos, I have to agree it looks like a teardown is needed.

I wish i had time this weekend to come down, I have an engine hoist i could bring with me.

Well any doubts the engine is fried have now been eliminated. Started the car after work and even though it started right up it sounded like it was only running on one rotor. Eventually the engine caught itself and I was able to drive home. Got home, shut it down and changed, came out to start it to depressurize the fuel system and it absolutely would not start. It took me 30 minutes to start it again and the same thing happened - started on one rotor but after a few revs and about 30 seconds the second rotor kicked in. Pulled the fuel pump connection and let it stall out and started the teardown. Can't wait to see what exactly happened.

Started the teardown last night. Found something interesting. Note the line for the oil injector on the primary port for the front rotor. That couldn't've been good for the engine!

https://lh4.googleusercontent.com/_7...6-29-59_30.jpg

Eh, that's not as bad as it could be. I'd be a little more concerned if the primary oil line was broken (the one that sits directly on the rotor housing).

well do you premix?

I haven't thoroughly inspected the other lines but they seem brittle as hell. What are these things made out of? What should I use to replace them?
Nope, never saw a reason as the engine was stock and I assumed the oil injection system was working just fine.