Reroutes, and gaskets, and data...
 

The recent thread about the new Qmax coolant reroute package has re-awoken an old thought scratching around in the back of my head. It's about the differences in the stock head gasket between the 01-05 engines and all the other NA/NB engines.

There was a big thread about this back in 2010 or so on M.net. In was mostly Jason_C and myself documenting the locations of the coolant passages on the blocks, heads and gaskets of various BP-series engines. The short version is that most of us are familiar with the shortcomings of the cooling-system design in the NA and NB engines, stemming from the fact that when Mazda took what was originally a transverse FWD engine and turned it 90° to stuff it into the Miata, they wound up relocating them thermostat and radiator feed from the "back" (flywheel side) of the engine to the "front" (pulley side). This design deprives the rear of the engine of coolant flow when the thermostat is open, thus causing the back of the head to run much hotter than the front, and decreases the efficiency of the system as a whole by allowing significant amounts of coolant exiting the pump to travel straight up and then right out the front of the head again, bypassing most of the block and head, and not picking up nearly as much heat in the process as it otherwise could.

What few people (myself included) had noticed until Jason brought it up was that in 2001, Mazda apparently acknowledged this shortcoming, and re-designed the head gasket to block off quite a lot of coolant passages, with the effect that the coolant was now forced to take the long way around the engine, thus permitting it to actually do its job.

I wound up creating the following illustration which depicts the active coolant passages between the block and the head, taking the gasket into account, for a '90-'00 engine vs. an '01-'05 engine. In these images, we are using our X-ray vision to look straight down through the top of the head:

https://cimg7.ibsrv.net/gimg/www.mia...1990a9dc18.png



For anyone who has spent a significant amount of time pondering the design of the cooling system, and the various aftermarket solutions to it, the ramifications of this should be pretty obvious. Assuming a totally stock configuration, the '01-'05 gasket design means that the coolant can no longer just flow right up the front of the engine and then out the thermostat. It's forced to travel back through the block into the #2-3 area, then flow upwards, and then flow out both the front (thermostat side) and rear (heater core side) of the head.

Seems like a pretty clever solution which, from Mazda's point of view, incurred zero additional cost and required changing only one part; the head gasket. (It should be noted that this head gasket design is totally unique to the '01-'05 Miata, whereas the '90-'00 head gaskets are the same parts shipped on all of the various FWD implementations of the B engine.)



For those who have NOT spent a significant amount of time pondering the design of the cooling system, the ramifications, and indeed the very core of the problem, may be non-obvious or even counter-intuitive. I'm not going to offer a primer here, merely an apology: if you don't already know what we're talking about, then sorry, but you're in the wrong thread.




And this got me to thinking...


1: Does the stock design in the '01-'05 engine actually suffer from the same basic inadequacy as its predecessor? Was Mazda's solution vis-a-vis the new head gasket design every bit as effective as a rear-thermostat reroute in the earlier engines? Or do these engines still have the potential to benefit from the relocation of the thermostat back to where it belongs?

2: On the obverse, can doing a rear-therm reroute on an '01-'05 engine (with the stock head gasket) actually do more harm than good?



My reasoning is as follows: With regard to the image above, on the side labeled '03, visualize the flow of coolant as it exits the pump and is forced into the front of the block, around the #1 cylinder. We can see that the coolant is being encouraged to travel down the block until it reaches the #2/3 cylinders, then turn upwards and flow into the head. Now, were the thermostat still at the front of the head, the coolant would take two diverging paths. Some of it would flow towards the rear and exit through the heater-core feed, picking up heat from the #3 and 4 combustion chambers. And some of it would turn towards the front and exit the thermostat housing, picking up heat from the #1 and 2 chambers.

But imagine that we have relocated the thermostat to the rear, and completely blocked off the front coolant port. Now, all of that coolant flowing upwards around the #2 and 3 cylinders is going to turn towards the back, picking up heat from the # 2-4 chambers on its way out the exit. But look at the #1 chamber; its basically a dead zone, particularly at the front and intake sides.


To my overly analytical and yet data-deprived mind, this would seem to actually create a problem exactly opposite that which a rear-therm reroute solves in the earlier engines. We're now depriving the #1 combustion chamber of coolant, and forcing it to run hotter than the rest.


But of course I have no data to either prove or disprove this theory.


I've read a number of anecdotal accounts, in which people have said "I did a rear-them reroute on my '01-05 engine, with the stock gasket, and it didn't blow up." Ok, I'm totally on-board with this. We know that rear-therm rerouting on an '01-'05 engine doesn't destroy it.

But does it actually do any good? And, more specifically, does it, in fact, cause the situation which I've postulated above, to wit the temperature gradient across the head actually becomes more uneven than it was before?


I'd love to see this data. Has anyone actually instrumented an '01-'05 head with multiple temperature sensors, and done an A:B comparison before and after a rear-them reroute, to actually determine whether we're making the situation better or worse?




And we finally come to the post which triggered this whole line of intellectual gluttony. It's copied from a different thread, so paging @yossi126:

Ok, but was this with or without a rear-them reroute?

If without, then when you are saying makes total sense, and supports my presumption.

If with, then this totally throws my theory on its head, and also contradicts much of what has been said on the matter by knowledgeable persons, to wit (paraphrased): "The '94-'00 gasket with a reroute is best, but the '01-'05 gasket with a reroute is better than nothing."

Joe, I too was wondering about this. I actually, once remember explaining to someone asking about this, exactly what you just posted. This is why you don't reroute with a 01+ headgasket.

It is my belief as well, that this would stagnate coolant flow in the exact same places you listed, effectively duplicating the problem we had in the 94-00 cars in the cylinder 4 area, just on the opposite side of the engine.

One glaring exception, the reroutes do not relocate the CLT sensor to cyl 1. If on the 94-00 vehicles with no reroute, stagnant coolant flow to cyl 4 began to be a problem, it would reflect in the CLT reading. Not so in this case. I feel the line of thought that has appeared (Read: A reroute is ALWAYS better no matter what headgasket) may not be well supported. I myself have not seen validation of CLT temps at the front of cylinder 1 post reroute on a 01-05 gasket.

How effective is the 01-05 gasket? I don't know. I have not done any track days on my built engine, so I can't tell you that perspective. But my N=1 experience, hooning my car around in 105 degree heat, then idling for 30 minutes on searing hot pavement in essentially grid lock with one cooling fan down did not result in a CLT reading over 205* My eyes are always on my shadow dash in those situations.

That being said, I still believe the 94-00 gasket with a reroute to be the optimal cooling solution for the BP engines.

WHAT ASSHAT JUST QUOTED ALL THAT. I DONT EVEN HAVE THE TIME TO SCROLL UP AND FIND OUT. if I still had mod powers...there's a warning for being a retard!

u mad brah :giggle:

"Ok, but was this with or without a rear-them reroute?"

Yes, both cars are running rear-therm reroute.

Am I understanding correctly that Yossi's sample of 1 event, 2 cars, tends to indicate that the best choice is NB2 gasket with re-route? If, of course, one looks at Temp at back of head only.

One other approach is to use NB2 gasket, without re-route, on all NA and NB's. EDIT: Inthink it has already been shown that, in all cases, a reroute is better than none.

Those two cars had different radiators, fans and no mention of how good the ducting is....... That isn't a data point.

The easy test here is to add a temp sensor to the front of the engine and compare temps pre and post reroute.

This is what I suspect. And yet I can't prove it.




I actually believe that the situation may be worse in the case of the '01-'05 engine.

In the early cars, there wasn't a lot of coolant flow around the #4 combustion chamber when the thermostat was open, but at least there was some. The rear outlet to the heater core, while narrow and restrictive, was at least always open. By comparison, when you rear-them an '01-'05 engine, there is literally no laminar flow whatsoever around the front and intake side of the #1 chamber, merely whatever convective and turbulent transfer might happen to occur by chance.

Is this causing '01-'05 engines to be destroyed en masse? Clearly not. We have enough anecdotal evidence to reach this conclusion.

I merely ask whether the rear-therm reroute has any benefit at all on the '01-'05 design, and posit that it may actually cause some small amount of harm.




And that's why I'm leery of all of the "data" which to data supports this configuration. It would not be especially hard to settle the question; just tap a hole in the blockoff spacer which is normally installed at the front of the head when a rear-therm reroute is done (assuming the water neck is being removed entirely) and stick a temp sensor into it. Log this sensor alongside the rear sensor, and post a graph showing how much the two spots deviate from one another, how much the temp at one location lags / leads the other as overall temp changes (eg: due to movement of the thermostat), and so on. Repeat for an engine with an '01-'05 gasket (and an otherwise identical configuration), and that'd pretty much settle the issue.

Sadly, I lack both the physical health and the number of Miatas required to perform this test at the moment.





Huh, turns out you're right. :giggle:




I believe you are, given that:

1. There's obviously a big difference in radiators between the two cars,
2. "Overheating" is not defined in this context,
3. Yeah, we're probably talking just about the temp behind #4, which totally ignores the problem which I suspect may be occurring, to wit; the temp at the front of #1, and
4. I'm not entirely sure how much I trust one anecdotal datapoint supplied by yossi126.

Of course, maybe there's some external factor that I'm completely blind to which, in the case of the VVT engine, makes none of this matter. Maybe there's a lot of thermal conduction happening through the MLS gasket, maybe the #1 cylinder in these engines just naturally wants to run cool. Maybe Kanye West actually does posses some modicum of talent. I can't prove any of these things, but I also can't disprove them.


Replacing the head gasket is a lot of work. Bolting on a reroute kit is relatively quick and easy.

I wouldn't posit that the NB2 gasket with front-therm is superior to the earlier gasket with rear-therm. I'd accept that the two configurations might be close in both overall system performance and linearity of temperature across the head. But, again, I have no data to support any of this, nor does anyone else appear to have any data to refute it. And that's just frustrating as hell...

^^ Oh, I suspect that Andrew and Emilio will be on shortly (California time) to set this all into perspective.

Just so my reference for "overheating" is clear, the point which we both set a rev limiter on is 225f or 107c, at which point we backed off. I was never going to go into the science bit of things, as there are too many variables between us. And to the guy who wondered, no ducting or hood vents on neither of the cars.

I wish someone would. Emilio & I have discussed this concept offline some time ago, but so far as I'm aware, no one has yet gathered any hard data on the matter.





So, as per DNMakinson's remarks, we're looking at the temp at the back of the head only.

This is interesting.

I'm not ready to accept that a rear-therm '01-'05 config is superior to a rear-therm '94-'00 config in any way. In fact, I flatly reject this hypothesis.

I can buy that the 949 radiator is probably the best unit out there, and that numerous other tedious differences likely existed between the two vehicles.

And I'd have really, really loved to have had some data as to the temps up front at the #1 end of both of these engines while the #4 end was heading towards the top of your scale.


But alas...

Don't you have an 01+ Miata? Install qmax reroute (when the lung reinflates) and tap a few holes for thermocouples in the 4 corners of the head.

You seem to underestimate both my cheapness and my apathy.

For what it's worth, which may not be much, I've done some combustion chamber cooling design work, and at least intuitively, I agree with Joe. The NB2 head gasket with a coolant reroute seems likely to result in dramatically reduced flow around cylinder 1. According to the nifty x-ray diagram, only one corner of the cylinder would see any coolant, which would result in a very poorly and unevenly cooled combustion chamber. I would think such a non-uniform cylinder temp could potentially cause issues for the ring pack too. In fact, this seems to be a flaw with the NB2 gasket block-off trick in general, since 3 out of 4 cylinders only see coolant on one side. That being said, my (somewhat limited) experience in power cylinder design comes from industrial diesel and natural gas engines that are required to last 80k hours, so "little" things like uneven cooling are a big deal. Nonetheless, for that reason, I would say the NA/NB1 gasket with the coolant reroute would be the most effective.

I would also say the yossi data point is worthless because of the different radiator setups, unless the delta T across each radiators could be known with some accuracy and taken into account. Even then, though, the cyl 4 temp sensor will only tell part of the story, and it's the cyl 1 temperature that's concerning.

*edit*
When I say "see coolant" I should be saying "see coolant flow." I understand there will still be coolant there, but if it's not flowing, it's not doing much cooling.

If cyl#1 in the 01-05 config is seeing less flow and the pressure differential in both setups is similar, would you expect to see more flow reverted past cyl#4? Could that then result in lower temp measurements at the sensor?

I'm wondering if the front cylinders are seeing a sort of "heat engine" effect. The pump itself may not be producing much flow, but the temperature differential helps. There could also be some pressure differential induced flow as well. Water "exiting" the system leaves a lower pressure area and brings some of the high pressure (hotter) water into the flow with it. Its been a long time since I've studying thermodynamics, so I could just be pulling random bits out of my head.

you bought a keyboard and you use it. that would be the extent?

I had a 2001 with no reroute and it overheated. Mostly because the water turned into liquid iron and started turning freeze plugs into weep plugs.

Prior to that clt temps were pretty solidly in the 220-230F range when it was running hot. It never asplode.

Lower than what?

If you mean "lower than the temp around cyl #1," then yes, that's exactly what we're talking about.

Liquid Iron? I have no porpoises.

But I have been keeping a close eye on CLT in my stock '04 since noticing the leak (which I still haven't fixed) a few weeks ago. As reported by OBD2, seems to hover around 210-215° on the street at 80-90° ambient with the aircon running, which is pretty much ideal. (This is mostly freeway driving, not stop-n-go.)

I find myself being annoyed by the relatively greater complexity of removing and installing the cam cover on the VVT engine as compared to the earlier models...

It is unbolting the hard oil line at the back of the head, and then unplugging the VVT pigtail. Then no harder then any other VC on a miata. Well maybe having to remove the stock COPS.