Thanks for the feedback Padlock. I assume you already read that, but I saw 105 C when climbing the local ski hill. I was just curious to see how much of a difference, if any, the additional hood vents did. I lost the log, but from what I remember that was 2 minutes into the hill and I turned around. 30 C (80 F) ambient temps on that day, but we'll see upwards of 38 C (100 F) later in the summer. Temps did sit around 104 for a while before hitting protections at 105, so it's possible they were stabilizing.

I'll try to get some more testing in and actually get a log or two. I'm not trying to build a car that can run pikes peak, but I would like to track the car at some point and right now I think it would turn 1 lap before I'd need to cool down (which maybe isn't a huge issue).

----------

Coolant is bled and I changed the CAN stream on the Link from "Generic Dash" to "None" and suddenly my phone started showing values on both Car Scanner and Real Dash (after a few tweaks to the Real Dash settings). That's a big win, especially for checking temps when I'm putting around town and don't have the laptop handy.

 

WOOOT! I'm glad you got the OBD working. What bitrate did you end up using? Post an update after you use it a while. It has been in my list of todo's, but since I have the CANChecked it is low in said list.

 

I bought one of these exact bottles with the thought it would be great in the garage, but it has never once worked for me. It certainly doesn't stop bleeding on its own and if I do manage to get a cut to stop bleeding, it seems to dissolve away the clot on contact. And it's like superglue in that it gets everywhere you don't want it... I must be using it wrong.

 

I do plan to do a write-up once I'm happier with the install and have played around with realdash and some other apps. RealDash seems like it would be awesome to run on a tablet, but I don't have any plans for that.

The liquid bandage isn't perfect, but I've used it with success here and there. I've also painted on way more than expected at times and left myself with a glue coated finger.

 

Not out of the weeds yet on cooling. Armed with my handy dandy OBDII readout, I saw coolant temps hit 102 C today while driving home from the park. That was a 10 minute drive with mild traffic at 32C (90F) ambient temp. Not the worst case by any means. I drove like I normally would, didn't push the car too hard, but got into boost at one point or another.

I ran AC the whole time, and honestly this is the best the car has run with AC on since I went standalone. Super nice to be able to pull up to a stop light without worrying about idle dips or other weird things. That alone has me way more inclined to actually use it. In the past the car felt super unhappy when it was turned on. I have also been getting a screech whenever the compressor turns on, but after giving it some thought I realized the belt was probably too loose and, sure enough, a quick tighten seems to have fixed that issue.

So I guess I have two questions now. Am I chasing a nonexistent issue here? I haven't seen the car actually "overheat" before. I don't have any reason to believe the coolant temp was going to stabilize, but 102 doesn't seem dangerous. It wasn't climbing quickly, but I don't think I'd have any trouble raising it with a quick pull or some uphill. I can turn AC off as well, but again, what's the point of having AC if you can't even use it when it's hot outside?

My other question, is trying to figure out how much of a difference my sagging belly pan is making. When I jumped under the car the other day I realized it wasn't sealing up with the ducting I've done, so I'm assuming it's acting like a big scoop and pulling a bunch of air up into the engine bay. Not sure if that's correct, but I know that's one of the common things people mention when discussing cooling.

I have more coroplast lying around, so maybe I'll make another sheet to sandwich the belly pan and seal things up. If I'm not seeing much improvement with some more experimentation and testing I think a new radiator will be on order.

 

105C ~220F isn't actually all that hot considering.
but heat can runaway pretty damn quick.

120 is really danger range.
and 130 is pretty much HG range.


That said, I'm hitting the same temps dual driving on autox, so definitely wanting to get it NOT that hot just off a 45s run.

as for the undertray gap, definitely something I'd look at fixing, keep the rear of the tray open, all the air should go out there (and out your hood vents).

 

You're still on the stock thermostat I assume (90C/195F)?

I should mention that I'm running an 80 C (176 F) thermostat.

 

Got that belly pan gap sealed up. Probably better ways to go about this but I wanted to get it done quickly before the sun went down too far so I could go do some more testing.

I think it helped a decent amount, but hard to tell for sure. I went to our local ski hill road and did some sensible driving uphill with the AC on. Mostly around 3K rpm, on and off throttle, following a soccer mom at one point, etc... After about 13 minutes I was at 110 C.
Flipped AC off and continued (keep in mind that as I climb the ambient temp is dropping a good deal) temps stabilized around 100 C.
Romped on it for the last bit and was back to 106 C after about 6 minutes, which was when I got to the summit.



Next thing I'm going to go do is a climb a hill that's steady state ~55 MPH. There's one around here that's out of the way, but it's a stress test every time I climb it in the summer. Actually had a guy tell me he will take a detour to go around that hill in the summer because his Mazdaspeed3 will get too hot. Not sure if I'll hit that hill yet, but I've got another that's closer and still pretty long.
I'm also going to grab an IR temp gun to try to get an idea of oil temps via the dipstick. I was idling around 17-18 PSI of oil pressure when I got to the summit compared to ~24 when I started. AC was on when I started, so idle RPM was a bit higher, but only 50 RPM or so, so it should be negligible for oil pressure IMO.

 

Those are insane temperatures in my world. I would have pulled over at 105 and let the car cool with engine off, fans on. Do that for ~10 minutes, idle the car for 2 minutes, and do it again. Fans cool off rad, then you cycle it through the engine.

Keep in mind steady up hill climb is the WORST conditions for cooling a car. You're under a decent amount of load, gravity is fighting you, you're not going fast (less air flow), you're not sucking in a lot of air through the engine, and it takes a long time since you're just going 55, all while trying to target what, 15:1 afr? I'm fairly certain if you repeated that climb in a competition scenario, you would run cooler. More air into the rad, more air pumping through the engine, richer AFRs, less time, etc etc. When I do partial load tuning on the dyno, there's no fan in the world that would keep a car cool for long periods of time, I always have to take a break.

THAT rant over, why, if you have a 80c thermostat, are you never actually at 80c? I typically hover within 5f degrees of my thermo temp, track sessions are 5-10 over. You start that log 15c over, then get 30c over.

You took out your rad that's protected by the AC condenser, what's the AC condenser look like? I'd seriously consider at least testing without it. Maybe have it drained at a shop, remove it, do some testing, and if you can't live without it, put a new one in.

Your IATs are also really high, when I did a track session in 95f ambient, I was running 205-210f ECT, 93-103f IAT. You're at similar ambient temps with 10f degrees hotter AITs.

 

Stupid question Dept.

Have you allowed a big enough exit behind the radiator undertray? There's a lot of air that has to get out, and while bonnet vents help, there needs to be space there and I am not sure how big the opening is behind your coroplast.

 

I'll jump into this..

What world is there in which 105C is scary? That's barely over the boiling point of water at atmospheric pressure. If he was 5C lower, one could argue he almost wouldn't need a pressurized/sealed system.

Common misconception, but tstats don't directly control max operating temp of the system. The radiator does as its the only component that can remove heat. Tstats only function as a minimum temp and warmup control device. Lower temp tstats may indirectly lower your observed max CLT temperature just because they open earlier, make the system warm up slower, and allow coolant to get to a colder temp in low load sectors of a track therefore the highest temps in the high load sectors aren't as hot, but the tstat isn't controlling that max as it is open the entire time regardless. It's just a heat in from engine vs heat out from radiator with min temp fixed by tstat calculation.

In the OE performance world, it's quite rare that cooling systems can shed heat faster than engines can generate it in WOT conditions. If you are hovering near your thermo temp, that just means you have a cooling system that is able to very efficiently dissipate heat (or oversized) or you aren't generating that much heat to begin with from engine load / hp (again pointing towards cooling system being oversized). There's nothing particularly wrong with being oversized on the system (other than higher costs in doing so), but I wouldn't call hovering at tstat temp a requirement. It's far from it. I can name a large pile of OEM systems validated to comfortably operate with 185F tstats where typical operating temp is 205F-230F.

"really high" feels very exaggerated here in my experiences. Seeing a 5-7F increase in ambient to IAT temp is quite good if you are achieving that, but 10-20F increases over ambient is well within reasonable values where you still shouldn't require MAT corrections. I'd actually kind of expect that 10-20F IAT temp increase over ambient if utilizing budget friendly components with a turbo that is operating at higher than OEM boost levels (12psi) where the compressor map isn't quite as efficient.

 

And I fully expected you to come in here to contradict me a bit, which is understood. We could go on a deep dive of this, I'm fairly certain all the "that's way too hot" culture came from when cooling systems and engines were simply way too inefficient, so the thought was start your 20 minute sprint race without a thermostat at 150f, finish at 210f, rather than a 210-270 race.

From what I've seen, cars are just built better these days, and I'd expect OP's to perform a bit better too. Even modern GT4 factory built cars race at ~205-210, with little to no cooling mods from the street car you can buy from the dealership. While 105c isn't too hot, my experience with turbo miatas says you want to build a system that can maintain a cooler temp than that on the street, so during track use it'll have some head room before warping heads at 240f. And hopefully I didn't say they make less heat at WOT, I didn't mean to. Simply that it's less time and less heat soak. Track time also involves off throttle deceleration zones, which usually result in a dip in temps.

In my experience, it's not too hard to make miatas hover around thermostat temp, and their AC condensers act like a brick wall. As long as it hovers around 80c at idle and knows that his AC condenser is probably a lot of his problem, that was the gist of my post.

 

Curly, I appreciate the feedback and comments. The whole reason I let the car get this hot is because I wanted to stress test the system. I've seen people say that their system seems to stabilize between 210-220 F (or whatever), so I was curious to see if the coolant temp increases would level off at some point. I also wanted to get a feel for how long it would take to get that hot because I expect to take the car up Lolo pass again (long, steep, 55 MPH, 100 F Ambient) this year and I remember watching coolant temps slowly rise on that climb last year with next to nowhere to pull over.

This is the first time I've "built" a car, so I'm still learning a lot and experimenting to see how things behave.

• In that datalog I wasn't holding steady state, it was on and off throttle because the road is pretty windy. I plan to do steady state 55 MPH at some point as I think that will give a better idea of how the cooling system is shedding heat.
• Fans kick on at 85 and 92 (or with AC). I think I was at 95 C because the car simply gets that hot when putting around with AC. I have a switch to force them both on regardless of temps, so I usually flip that on when I'm going to be beating on the car.
• I did look at the condenser when the radiator was out and hit it with the shopvac brush. I didn't have a good way to hit it with the hose, so I'm sure it could be cleaner, but it didn't seem excessively dirty and most of the fins looked fine.
• I'm going to blame IATs on the sensor placement behind the radiator. Next time it's convenient I plan to move that to the intercooler end tank.

Basically I'm trying to get an idea of what the system is capable of and how it behaves. In my eyes the system as it sits now is inadequate, but I also don't want to throw money at a problem unnecessarily.

Gee, the opening at the back of the belly pan is unmodified from stock. The belly pan is just hanging down towards the back of the subframe. I haven't modified that at all.

Part of this is also because I have plans to run more power in the future. I expect moving to a turbo that's more efficient at higher pressure ratios will help with heat as well. I think the 2560 is pretty choked on my setup around 200 KPA. Regardless, I still expect similar heat from a bigger turbo. Definitely worst case or stress test conditions, but I'd rather not have to sweat when I see a big hill.

 

Aligned to this. The "too hot" culture really stems from how much cooling systems (specifically radiators and controlling their inlet/outlet airflow) have improved over the years... super thick, inefficient, low fin density cores of the "old days" that had minimal inlet ducting were meeting street uses, but not up to the task of additional heat input when on track so people will install lower temp (or remove) tstats to bandaid the problem.

Bold part is important and very real... OEM's have much better CFD modeling software now than we ever have.. OEM's are also starting to build performance cars for more than just straight line performance, so optimizing cooling circuits to perform under those demanding use cases is something more common now that they frankly just didn't do or have capability of doing well 10-15 years ago. The c7 vette is a notorious modern example of poor system execution (fine for the street, but too hot on track).

The intercooler you select can be an equally restrictive brick wall if focusing on coolant temps. Optimizing the core stack of condenser, IC, and rad isn't a trivial matter as they all influence each other. Starting with ducting (which OP is doing) is the best first step. It's feasible that the Mishimoto radiator or intercooler selected are not jiving well with each other. The condenser is surely doing no favors, but I can respect not wanting to remove it.

OP do you have any photos of what your inlet ducting setup looks like with the front bumper off?

 

Padlock, my old ECU had a weird issue where it would read some crazy coolant temp spikes during autocross, so that was part of the reason I opted for the 80 C thermostat. I wanted more headroom before getting to a dangerous temperature.

Ducting photos here. Not the best but I think it's decent. The mouth of the bumper slots into the opening pretty well.


Ignore the horrible bumper mounting. On the list to fix... again.

Driver side IC

Note the notch in the back is for the radiator drain plug. The coroplast goes right up to the bottom of the radiator.

Passenger side IC

 

Is there any gap between the IC and the ducting or is all the air being forced through the IC before it can enter diffused section where it starts widening out to the condenser and radiator?

 

I think the fitment to the top of the IC is pretty tight, but it's not fastened or attached to the top of the IC.

 

Forcing all (or at least most) of your flow through the IC first may be restricting the total flow rate you are getting into your core stack of heat exchangers.

Moving the IC back to open up the space is hard due to the tow hook mounting that you already have... as an alternative, moving toward a slimmer support bar and cutting the OE cross bar out really can free up space for better ducting. Not sure if curly has experience in this arena with turbo miatas, but the theory behind has proven to be pretty sound.. I'm thinking something like this


Pardon my quick sketch, but the left is what I am assuming you have now. 2 large blocks are rad and condenser, the small block is IC, the ducting is pink, the OE cross bar shown darkened in... with removing cross bar (as shown on right sketch), the top panel of your duct can be more of a straight shot to the top of the condenser and free up an airflow path above the IC that you currently don't have. This right sketch setup is more or less exactly what's shown in the image above.


Just another idea in the list of ideas. This idea doesn't force you to buy different heat exchangers but does require an angle grinder and new ducting panel(s). You also run the risk of some amount of airflow bypassing the IC, which could negatively affect intake temps as a tradeoff for more net airflow reaching your radiator for improved coolant temps.

 

Not a bad thought. The coroplast would probably be pretty easy to modify to add an inch of clearance or so above the intercooler. I've had some quick release pins I've been meaning to add to the bumper anyway, so maybe some experimentation will happen this weekend. It's also oddly cool this weekend, so the garage won't be an oven.

 

Haven't done any additional testing/stressing of the coolant system. The past 2-3 days have been way cooler (today's high is 25C/75F, while last week we were closer to 35C/95F).

I did finally install some quick release pins on the front bumper though. I've never had a good solution for mounting this bumper, but I think this is going to work for me. When I got the car there were 5-6 self tapping screws holding each side on, I moved over to some bolts and wingnuts which were an improvement but it's always been cobbled together.

The biggest improvement for me with the quick release pins is that if something goes wrong like and intercooler hose blowing off, I will actually be able to service it. Previously I would have to get the car in the air and remove the front wheels to get access to the bolts to remove the bumper, which always made me a bit anxious, fearing that I'd need to service something in the front of the car unexpectedly. It also made me reluctant to do any jobs that required the front bumper to be removed (IE anything with the intercooler or ducting).

So, this was actually way easier than I expected and probably took an hour or less.

Here, you can see my old setup. A bolt and wing nut. This worked, but was far from ideal and was a PITA to reinstall.




I grabbed these generic quick release buttons/pins on Amazon a while back but haven't pulled the bumper since I ordered them. Purely by luck, they were the same thread pitch as the captive nuts that hold the bumper and fender together. One of those bolts was already missing on my car, so I just had to thread the pin in.



I didn't take very good photos, so for anyone unfamiliar with this style of latch, here's a stock photo of the components.

After threading the pin into the chassis it was just marking the bumper and drilling out a hole with a step bit. I didn't do as good of a job as I could've (made one of the holes a bit too big), but these seem to be pretty forgiving.






Removing the bumper is now just 2 10mm bolts (used to be 4 up by the radiator but I ended up removing 2) and 2 pins. Takes a minute or two and the bumper is off.

I did a bit of sealing around the intercooler ducting but nothing crazy. I'll probably look into it later this week now that the bumper is quick to pull off.