Cooling System - Overflow starting at 135F, not overheating
 

2005, 106,000 miles on the chassis, 13,000 on the motor (mazda reman), OEM cooling system until recently.

About a month ago, I had my 8 on a trip for a week, and it regularly spent time driving in traffic at ambient temps >95F, with a peak of 105F. The following week upon returning from work and popping the hood to facilitate cooling as I usually do, I found that the return drive from work had dumped coolant out the overflow. The next day I pulled my thermostat, and stuck it in a pot of water along with a Mazmart 170F thermostat, and brought it up to a boil, watching the movement and measuring temps with an IR thermometer. I confirmed that the OEM thermostat did start opening at 183F as it is supposed to, but never got more than ~50% open, even at the 212F water boiling point. The Mazmart thermostat worked correctly, 170F start and 190F full, so I put that one in and put everything back together.

Ever since, I have a cooling system pressure issue apparently. Starting up the car and letting it idle up to temp, watching the overflow and the coolant temp, starting at about 135F, coolant starts dripping out of the overflow, increasing in flow rate as the temp increases, though not drastically. When it gets to the thermostat temp, it holds at 172F easily, but there is a tiny but steady stream of coolant from the overflow. I even grabbed 10ft of clear tubing to replace the overflow hose and ran that out of the engine bay to the driver's mirror so I could see the flow rate while driving. And I can empty the bottle to the point of getting the coolant level light triggering (validly) in about 10 miles of casual driving without hard throttle or high revs. Temp sensor is showing normal acceptable temps of 170-190 the entire time.

I did a full flush of the coolant system with a hose (>20 minutes of flush time) including full bleeding it, which failed to correct the problem. I just had the system pressure tested and coolant tested for combustion gasses, and the mechanic claims that it held pressure up to 20psi. The combustion gas test came back negative.

Side note:
The mechanic said that my "exhaust doesn't sound or smell right, so there is something obviously wrong with my engine, and the coolant overflow confirms that, so I must have lost an apex seal." :uhh:


So wait, my BHR midpipe is sucking coolant out of my coolant bottle overflow, which means I lost an apex seal....? :scratchhe


:rollingla :lol2:


Ok.


Not going back there again.




So about the only thing I can trust from the test is that the coolant is free of combustion gasses, so it's unlikely to be a coolant seal problem? I have sent a sample of oil off to Blackstone for testing, which should confirm it.

I guess I will take a day or two and disassemble everything I touched during the thermostat replacement, and inspect/replace. Any suggestions, tips, places to check?

Radiator cap should be checked since it controls that aspect.

Paul.

New cap from Advance Auto. The mechanic said it holds pressure to 20psi. Though I am not certain I can really trust him completely.

Part of me has been thinking I just have a coolant bottle and cap failure, and that I need to order the brand new set from you, but I'd rather make sure of that before I just toss you an order :)


(though I did JUST email you on a different topic regarding my Miata.... :lol:)

You sure the cap seals properly to the bottle...OEM cap would likely be better

No, I am not sure :)

Yeah, probably need to get an OEM cap to test with. Just shipping usually doubles the price... or the dealer doubles the price without shipping, so I just grabbed one locally. Though if it's the bottle itself that isn't allowing the cap to seal, then an OEM cap won't solve the problem (though I'd just order an entirely new bottle and cap from Mazmart at that point).


I wonder if my Miata's cap is the same size and rating.... hmmmm

It's 13psi too, and fits right on. Idling it warm now to see if I still have overflow.

Goddamn that's bold.

Isn't the RX8 is 15psi?

13 that I was aware of? :dunno:

Got a video. There was coolant left in the long tubing, so I looped it up through the hood latch, letting the coolant in the tube sit against the overflow spout. Idling warm there was a stream of air bubbles out of the overflow for about 5 minutes.

Then the bubble flow stopped, coolant level in the tube remained steady, so once the air was done escaping, coolant wasn't.

I hopped in the car and increased the revs. The coolant level in the tube dropped as soon as I picked up the revs, and dropped again at each rev increment increase. Coolant getting sucked back into the coolant bottle. I dropped the revs back to idle and the coolant level in the tube remained at the low point that it had dropped to, but then I started to get a stream of bubbles again.

I would guess that this means that if the air in the system got squeezed out during the initial idling, increasing the revs introduced more air back into the system. Possibly a coolant hose failure that is letting air in?


Video upload to youtube seems to have stalled.

And yes, it's not really 'coolant' at the moment, almost entirely water until I can get this solved. Maybe a pint of 50/50 in the system to give the green tint so I can see it more clearly. Once I have it solved, I'll reflush and increase the actual coolant %. In diagnosis it helps being able to ignore having to catch draining coolant.

Chris (My Miata man) will respond to that email shortly. He will need to know things like how recently the swap was done, what coil type is being used etc.

Paul.

Yeah you're right... 13psi I just checked mine.

The fish in your local creeks will thank you ;)

:lol:

And my neighbors won't get upset about me dumping coolant on the ground.

Update:

I left the radiator cap swap as is, and drove the 74 miles to work in the Miata using the radiator cap that was on the 8. Zero coolant loss, temps remained steady, even in stop and go traffic that extended my usual commute by an hour. So it's not the cap.

I also kept testing the 8 into last night periodically, including disconnecting all the hoses involved in a thermostat removal and reconnecting them, refilling and re-bleeding the coolant again, and the problem remains.

You might try twisting the oil fill cap open to make sure the crankcase isn't pressurizing, thus potentially forcing gas (not necessarily combustion gas) past a coolant seal. Another test would be to measure the pressure in the coolant overflow bottle. Is there any water in the oil?

Hmm. I was about to say "I did that", but then I realized that it's slightly different. You are saying "determine what the pressure is in the coolant bottle", not "find out what pressure it can hold" ?

Will test the oil cap off to see when I get home.

I'm guessing the pressure is really climbing to 13 psi and the cap is opening properly. In the absence of boiling, it seems that blowby would be the only cause of that. You might also take your IR thermometer and "Survey" the inlet, outlet, heater hoses and radiator to see if the temp distribution makes sense, in case the water circulation is messed up somehow. I'd still bet on blowbuy though and a 50/50 change the bubbling stops or slows when the oil fill cap is slightly opened.

Hmm, yeah, making sense. The air getting pushed out during warm up as it comes up to pressure, then getting re-introduced with climbing revs, and then getting pushed out again once it returns to a stable idle has been making me thing that this isn't a bottle/cap issue.

Sigh. I really hope it's not a coolant seal. Although if it is you all might get to see a 13,000 mile reman tear-down....

Just so you are aware, the OE Coolant Radiator Cap N3H1-15-205 is specific ONLY to the RX-8, Mazda does not use it on any other car, I know it is not the same PN as my NC Miata, could be label, have not compared...Just throwing this one out there.

And Yes, I have seen this with incorrect 'style' of radiator cap, is the non genuine one 100% identical to OEM N3H1-15-205?

The only other thing I can think of and have seen it before even though system pressure test is OK, and that is Coolant Entry (Coolant Seal) into Engine when engine is hot at it's 'very early' stages of deterioration...sorry.

Can you remove cap (hot engine) and see IF you are getting a very fine 'string/strand' of bubbles?

BTW, You need someone else on the Throttle while You observe coolant bubbles, hold revs at around 2500-3200 RPM for a short while and look...Good luck...I don't like the sound of this though Pearl..

Idle only, usually won't work, unless you have REALLY bad Coolant Entry...but then you would have start issues.

Agreed that the caps aren't the same part number. The top part with the "handles" is a different size of gripping surface. The have the same gripping diameter and depth, and the pressure part appears to be the same. I don't think that it's a cap issue, even if I had the OEM one, since if it was a cap issue then I'd expect I'd get a failure on the Miata (NB '99).

Most of my observations right now are with the impact on coolant in the vertical tube, which I can see from the driver's seat. There IS air getting into the system when I increase the revs, though it only gets out into the tube via the overflow spout once I let off the gas.

I tried with the cap off, and I literally just get overflow out the top filler hole at idle. And most of the contents of the coolant bottle disappear into the coolant system, triggering my low coolant level light, when the revs increase. Not sure how to figure out how to watch a stream of bubbles like this.



On the other hand, I was answering a question in the Dumb Question thread, and realized that I am at 106,452 miles (wife checked for me), and I had my engine replaced at 95,935, which puts my 12,000 mile service warranty good until 107,935, so I'm still under that service warranty for another 1,483 miles :D:

If I have indeed lost a coolant seal, I need to prove it to myself first, and then figure out how to prove it to the dealer. I seriously doubt I'd have sufficient compression loss this early in a coolant seal failure.

You need to remove water in CB to half full, otherwise it will overflow out of top of CB.

Obviously you have to be careful when removing hot cap.

You need 2 people, one on throttle the other looking (inside CB) once cap is off for 'a' bubble string...
You have to do all this relatively fast or pressure builds up and overflows, idle is no good, keeping revs up 3200 keeps water down.

The level varies even more than that Ash. Not something that I've seen an 8 do before, and I think it another indicator of the issue.

I can top off the system completely cool, and it will push enough water out by the time that it's warm that when reving in neutral to 2,500-3,000 rpm the level of water in the CB drops below the low level sensor in the bottle. I suppose I could have my wife hold it there with the cap off and me re-filling it to keep it up high enough to check.

Jeez!

The more I'm thinking about this, the more I think it IS a coolant seal failure.

Nothing else makes enough sense. I think the key point is that if all the air in the bottle gets pushed out in idle, where is more air coming from when I increase the revs?

Logically, it would have to either be A) a gas that is under a higher pressure than the coolant system, or B) there is a localized pressure drop in the coolant system to lower than ambient pressure, and it's coming in through a crack or other failure point (and doesn't spew out coolant when under pressure).

I don't see how B could happen, but someone else might. I can see a few ways A could happen, and none of them good.

To me if that much 'cold' water is disappearing it is like you have large air pockets in system.

I know this sounds a little contradictory but your Cooling System needs to be 'as normal as possible' as in no air pockets, it needs to be bled first, then has to be engine hot and TS open...to do bubble test..seal leak test.

Return water with compression bubbles comes back via Top Radiator hose and CB return hoses.

And you are not having cold or hot start issues?

The coolant seal test is no different than checking for banger Head Gasket Coolant Leak, it's just not so violent (usually) with Rotaries, unless the Coolant Seal has a hole in it about half an inch!...then it won't start anyway (usually).

Note, the coolant doesn't just disappear and never come back. If I let off the gas and it returns to idle, the coolant level returns to where it was at idle before I increased the revs.

Where it goes? :dunno: I don't have the answer for that. It would imply that there is alot of air somewhere in the system that is getting displaced by the coolant getting pushed through the block, but when it returns to idle the air goes back to wherever it was, pushing the coolant level back up.

I would also guess that if there IS a chunk of air somewhere, this air heating up is over-pressurizing the coolant system causing the overflow.

But then the question gets back to "where is it coming from?" since I've bled the system several times, including popping the throttle body lines with the car on an incline, with the heater on max. Unless I'm just not ever getting it bled completely? :dunno:

Agreed about the importance of bleeding, see my last post. I'm following the instructions, I don't know if I'm actually getting it completely bled.

No, I have no starting issues at all. Fires right up in under 1 second of cranking, hot or cold, even if it's sat without starting for a week (as usually happens now, haven't driven it more than ~20 miles in the past 3 weeks while trying to solve this) And no smoke on startup either. So suggesting that I don't have puddles of coolant/water sitting in the housing accumulating until I try to start.

If you are running your car with cap off then YES any car will overflow coolant at idle when hot as system is trying to pressurise...without correct pressure cap it will overflow on any car.

Understood.

As I have said before....the seal bubble test has to be done quickly, No mucking around.

It is cap off, revs up and look inside...all in 10-15 seconds...with ALL the Hot Coolant Cautions removing Cap.

1.Apart from that, I would have said wrong Cap...but your Miata was OK
2.Air In system.
3.Coolant Seal.

Perhaps someone else can chime in? :)

I would still go for a Genuine Cap N3H1-15-205....for the $15 it is worth it.

Anyone here with an old Coolant Bottle (sensor replaced) with a 'good cap', or someone that could lend Pearl an RX-8 'Cap'.?

Could the cap be good but the bottom isn't sealing properly? This could lead to either the spring being useless or not taking as much pressure to allow coolant past.

:lol:

I hear you loud and clear Ash. :) Never fear, I respect the pressure and heat of a hot coolant system. Never know who else might be lurking that doesn't though :suspect:

I sent a PM to CRH asking him to take a look at this thread when he has time, since his experience with both engines and cooling systems should be able to assist with a diagnosis as well. I'll probably be "mucking about" with the problem all weekend, and using as much of the public water supply as I need to do it. Might even temporarily replace some of the CB hoses with clear ones just to see if I can see air movement through those. Or make a BIG loop of clear hose to the throttle body to see if I can get air to get trapped up there :D:


If I can find any conclusive evidence that it's the bottle or cap, I'm going to be ordering a brand new set from Mazmart. I really don't think it's the cap, but I can't rule out that the bottle itself isn't allowing the cap to seal. Which is why I'm not inclined to just keep throwing money at caps. If I replace the bottle, Mazmart's listing adds the OEM cap in as well so I'd be back to OEM at that point.

The cap/bottle point doesn't explain how this air is getting in the system though, so I am not confident that the problem is there.

OK, just logged on to tell you...this has been bugging me...so I doubled checked my new parts..

NC Miata/MX-5 Radiator/CB Cap is 16PSI.

All RX-8's N3H1-15-205 is 13 PSI...NOT 16 PSI ...:)..as I said this cap is ONLY used on the RX-8 CB.

I would not trust a non genuine (Advance Auto) Rad Cap...have you had it tested @ 13 PSI, is the spring identical to OE, seal, etc?

I think i have a new cap somewhere in my garage, if u need it i can sell it yo u, thats if i can find it tho

Or i can just lend u my cap to test

Oh have u check the bottle's tab to see if its holding the cap properly?

Ash, the Miata cap I tested is for the 1999 NB, not the NC, and it is 13psi, just like the RX-8.

However, no, the caps are not the same. The generic one has a spring external to the center "shaft", the Miata cap apparently has it's spring internal. The handles for gripping are different dimensions, however they have the same size and type sealing right, retaining cap, and gripping flanges. I have them sitting side by side in front of me as I type this.

If I had any suspicion at all that it was the cap, I'd have already ordered an OE one from Mazmart. The local dealer here is a tiny shared dealer, and they don't really stock Mazda parts that I can find.

With the same expected rating between the two cars I have and each cap fits each car just as snuggly, I would expect that a cap failure problem would have just transferred the problem to the Miata when I swapped caps. However the Miata did perfectly fine for 150 miles of driving, including slogging through I-80 traffic into the office yesterday morning. And the 8 continues to push coolant and air out of the overflow.


IF it's a failure at the cap/bottle end, then I would suspect the bottle itself is not allowing the cap to seal. But that doesn't explain the reaction of the coolant level when I increase revs, it doesn't explain how after reving I have more air in the system that gets pushed out at idle that wasn't before.

This is why I'm going to substitute two of the normal lines with clear lines today to test with, to see if there is actually air getting pushed around the system, and when. (No, I'm not going to keep it that way, I know this tubing isn't designed for the heat of the engine bay, and if it fails while I'm standing there it's easy to instantly shut off the engine, not going to drive with it like that)


I'll take video of behavior and post it. Have a few errands to run with the wife over the next couple hours, so I'll start posting updates and results this afternoon.

Ok, thanks Ray. It's a better / more detailed diagnosis of what I was headed toward anyway.

Since I still have 1,400 miles left on the 12k service warranty from my replacement at 96k, I have a chance at getting a replacement covered. Any recommendation in how I approach the dealer for a replacement? I don't think a compression test would fail at this point, or within the next 1,400 miles?

:lol:

Sounds good.

Fortunately the infamous (famous?) Wayne Mazda is 45 minutes away, just got to find out if the service warranty is good between dealers or just the issuing dealer.

If I get turned down or not viable, I'll be chatting engine rebuilds with you :)

Yeah, I expect that to be the case. Just want to have Mazda Corp validate it :)



I'd just grab a personal loan to cover it in the short term and start pulling this engine. I've said the remans are good for 10,000 to 100,000 miles, just didn't expect to be on the short end of that stick :icon_no2:



Shrug. Not terribly concerned about it.

Did you ever send off the oil to be tested? I am pretty sure it is a coolant seal same as my last motor but the oil test confirmed it for me

Yes, last week.

Posted this elsewhere earlier:

So did you get the results back from blackstone?

Haven't been contacted by them yet. Not sure how long the usual turn-around is.

It's quick, shouldn't take long at all before you get a call or email.

Well, sent Thursday, so "not long" would be last Friday? Today? Tomorrow?

I would expect you to hear from them today or tomorrow. Depending how the mail. You sent it USPS?

Ash, Ray,

I think this tells the story. This is after the engine is up to temp, thermostat open, and after about 3 minutes of my wife holding the RPM at 2,500. The video is at idle, then back up to 2,500rpm.

The tube is a section of clear tubing I added between the upper coolant bottle return hose and the bottle specifically for this test.

Aerated water.

It is when your wife raised the revs is when you check for a bubble string, (hard to see then, vid finishes).

The rotaries I have viewed (Coolant Seal failure, Coolant entry) for a minor blow back it is a constant fine bead around the size of a 'pin' head.

A small match head size bubble string and larger you would have major starting issues...Coolant use/ entry.

Repeat...you have to keep revs high 2500-3000 and constant.

Yup Ash. We did that before I grabbed the video. It basically looks like aerated water, pinprick bubbles. You can see them gushing through the tube in the video as a gush or whiter water. All bubbles. You can also see farther to the right along the tube where they were collecting on the inside of the tube and releasing as they accumulated enough. When the revs first come up in the video you can see the entire tube water flow turning white, and it's all air bubbles. (or some other gaseous substance)

Is basically a solid stream, though the volume changes at different times, idle OR higher revs.


It's definitely there, no doubt about it.