IW STi Forum banner

T-Gs Air to Water IC Build

242K views 424 replies 67 participants last post by  BooSTi.D 
#1 · (Edited)
These are my own conclusions. No water was harmed in the making of this movie. Do not eat. Dispose of properly. Any resemblance to persons or places real or imagined is purely coincidence. Enter at your own risk. Beware of dog.

So I figured since all the info on my AWIC build is scattered through out my journal I'd make a seperate thread specifically on this project for those interested in how I did it along with the thought process and research that went into it before hand.

First some background. The idea came as a result of periodically seeing random threads asking about it on here and that other site we don't speak of and just a general curiosity about why they aren't common in our world. I started picking up on the fact that there is alot of misinformation and misunderstanding about these setups and set about researching the true benefits and drawbacks. Basically the general opinion is that they're only good on a drag car with ice water running through it or on a mid engine car for packaging reasons.

The biggest misconception of an AWIC is that you have to run ice water in order for it to work. This is used as a reason that, for a daily driven street car, they won't work. That view point is based on the fact that drag cars do use ice but it usually is only good for one or two passes and then has to be replaced and there is no heat exchanger used. Repeated pulls will cause the water to heat up and the intercooler to become ineffective with higher and higher IC outlet temps. Without a heat exchanger the water has no way of shedding the absorbed heat. A properly sized heat exchanger is absolutely critical to make it work properly on a DD. Many of the DIY attempts I found had either an undersized or more commonly no heat exchanger. Typically a motorcycle radiator or oil cooler core was used if one was in the loop. I call this pre-meditated failure.

The next issue is location of the reservoir if there even is one. Nearly all of the ones I found have located the reservoir under the hood subject to alot of heat. The idea that you would subject your primary cooling medium to the hottest temps anywhere in the car baffles me. Pre-meditated failure.

Next on the list is the argument of an extra layer of thermal transfer. I have no counter argument for this one since it's basic physics. In an air to air the heat transfers from the charge air through the intercooler fins to the ambient air-a three step process. In an air to water the heat goes from the charge air through the intercooler fins to the water, from the water through the heat exchanger fins, to the ambient air-a five step process. Every step has it's inefficiencies that add up. However with a properly sized core, a properly sized exchanger and a reasonable reservoir located outside the engine bay the effects can be minimized.

I'm going to touch on reservoirs again. While not required I believe they are a benefit on a street car. There are several reason for this. It allows a place to keep the majority of the water out of the engine bay in a much cooler location and results in less heat having to be removed by the exchanger. The other valid reason to have one is the added total thermal mass of the system. A larger thermal mass means smaller temperature variances which equates to much more consistent IC outlet temps. The down side of a larger thermal mass is the longer it takes to heat up the longer it takes to cool down. The fix for this is again a properly sized heat exchanger. Seeing a trend yet?

Pumps. If there's any place to be a tight wad this isn't it. A high quality purpose made pump with extra flow capacity is a must. The norm is to use plastic bilge pumps that have an open flow rate of 5-10gpm or so. While that may be sufficient the part they don't tell you is that in this application they cannot be run continuously. They have to be boost activated and only come on under positive manifold pressure. What this means is that while your cruising the core and the water in it are heatsoaking to a pretty high temperature. As soon as you step on it the water starts flowing but the temps are still very high. The temps won't come down as fast as one might think either due to the core being hot as well. The water has to cool the core and the charge air at the same time. If the core is at a cool temp with a constant flow system the only heat that has to be removed is from the charge air. I liken it to using a fire extinguisher after the fire is already out. A step in the right direction is the Bosch pumps that are OEM on the Ford Lightning, Cobra and the GT. These are good pumps that are readily available, relatively inexpensive and move about 12gpm free flow. They will stand up to constant operation as long as they are down stream of the exchanger not subjected to high fluid temps. They have to be mounted at the low point of the system to maintain a prime. The best option and the one I went with is what alot of the Lightning and Cobra guys upgrade they're pumps to which is the Meziere WP136S pump. Constant flow, 20gpm, full ball bearing, billet impeller, rebuildable, 3000+ hour service life, set up for AN fittings and not very cheap.

So on to some pics and more info specific to my setup.

First would be the core. The core I ended up using is a dedicated air to water core. They are built differently than an air to air core with larger air passages and smaller passages for the water. The reason for this is it takes much less water to remove alot of heat than air so the charge air passages can be opened up and flow very well while the water passages can be made to flow well and still have sufficient surface area for heat transfer. This is the reason for such low pressure drop through air to water cores. Many so called air to water cores are just air to air cores that are jacketed to flow water. This core flows 745cfm. My Dom3 has a 52lb wheel which equates to 644.36cfm. Plenty of headroom.



The core compared to my Spearco.




Inside shot of the core showing the wide air passages and thinner water passages.



My exchanger. It's a 26x7x2 two pass bar and plate core. Both in and out are on the same end so the water does a U turn hence the two pass description.




Next is the reservoir. Total capacity is 2.6 gallons. The rest of the system holds about another 2 gallons but thats just an estimate on my part. The pump is mounted to the reservoir so thats covered here as well.



Here's the pump.




This is an underhood shot to show the layout.






Those are the major components. For the charge pipes I had Mark at P&L fab up those. The one from the snail is 2" diameter by 11" length aluminum and bead rolled at both ends. The other section along the side of the core is 2.5" diameter by 6" length, bead rolled, and has a 1" nipple for the BPV.

The majority of the components I used are from Frozenboost.com. Here's a laundry list:

Type 14 core
Type 118 exchanger
Ice box reservoir
6 NPT-AN adapters -10
2 90* -10 fittings
4 45* -10 fittings
2 straight -10 fittings
~40' of -10 braided line
2" straight coupler
2"-2.5" 90* reducer
2.5"-3" 90* reducer
2.5"-3" straight reducing coupler
3 2" t-bolt clamps
3 2.5" t-bolt clamps
2 3" t-bolt clamps

The pump is controlled with a switch thats installed in a Subtle Solutions ashtray delete plate along with the temp meters.




The temp meters are from Mach V. They're made by Davtron Avionics and are designed to measure the inlet and outlet temps on the intercoolers of aircraft. I use one for that and the other reads the water temp in the reservoir and a direct ambient air temp at the fog light bezel in front of my APS cold air.

Total cost for the whole setup ran about $1200 give or take. The temp meters added another $500 but I wanted those so I could have some hard data to work with and wouldn't be necessary for a normal setup.

Here's some numbers from some preliminary testing. Subambient IC outlet temps are the result of running this system in conjunction with my Aquamist water/meth at a 50/50 ratio by volume. I'm using a 1mm nozzle which is around 15% or so of total fuel flow.

CRUISE:
Ambient 93*
intake temp steady at 97*
Turbo outlet temp ~115*
IC outlet temp steady at 2-3* above ambient
ON BOOST:
Boost up to 20psi
Ambient 93*
Intake 97*
Turbo outlet 145-150*
IC outlet 85-86*

I have since done alot more testing trying to find the limits of the system and so far haven't done so. Multiple full boost pulls at 22psi from a stop through 4th gear have not given any indication of heatsoaking the water even with turbo outlet temps near 180*. With outlets temps that high with multiple pulls the water temp has never climbed more than 6-8* over ambient and outlet temps remained below ambient. I believe a true test would be on a road course with sustained full boost but I don't do HPDE events or anything similar in order to test it there. I believe if the exchanger is able to shed heat at a faster rate than it's absorbed heatsoak will not ever happen but thats just my own theory at this point based on my feeling that the exchanger I have is more than capable of doing so.

So far the project has exceeded my expectations and appears to have all the benifits of a top mount with throttle response, post turbo volume and such with the cooling advantages of a front mount. The downside is obviously the added complexity albeit only a small amount and the small amount of fab work involved but I feel those are far outweighed by the advantages I've seen thus far. There is more info in my journal but it's somewhat scattered about hence the reason for this single thread for it.

Ask away if there's any questions. Sorry there aren't more pics but it's my daily and it has to come apart and go back together in a timely fasion and I tend to be on a mission to get stuff done usually.

T-G
 
See less See more
9
#4 · (Edited)
Ah yes, I've been waiting for this. A cheat sheet :)

1. Have you ever thought about coating/wrapping the core, like the gold foil that reflects radiant heat? That way it wouldn't heat soak as much, not that it does now.
2. What do you think of the type 101 heat exchanger, which is 1.5" thicker than the one you have? Is there room down there? That means more water, so you could make the reservoir smaller.
3. Is an overflow necessary for the change in temperature, ie 20*F in the winter and 95*F in the summer? Is it not a big deal and I'm over-thinking it?

I know one of the downsides is the extra weight. That's mainly an argument for the suspension gentlemen who drive their cars to the limit in the corners. For a street driven car with enough suspension mods, I don't think it's as noticeable.
 
#5 ·
Ah yes, I've been waiting for this. A cheat sheet :)

1. Have you ever thought about coating/wrapping the core, like the gold foil that reflects radiant heat? That way it wouldn't heat soak as much, not that it does now.
2. What do you think of the type 101 heat exchanger, which is 1.5" thicker than the one you have? Is there room down there? That means more water, so you could make the reservoir smaller.
3. Is an overflow necessary for the change in temperature, ie 20*F in the winter and 95*F in the summer? Is it not a big deal and I'm over-thinking it?

I know one of the downsides is the extra weight. That's mainly an argument for the suspension gentlemen who drive their cars to the limit in the corners. For a street driven car with enough suspension mods, I don't think it's as noticeable.
I have given some thought to insulating the core but it hasn't happened yet. It's in the future cards. The thicker exchanger was the one I was originally planning to use but wanted to make sure I had enough room behind it for some fans. As it turns out even with the thicker one they would still fit depending on how the exchanger is mounted. That said I've not seen anything that would require the thicker core up to this point. The reservoir is the overflow. The system isn't pressurized and whatever thermal expansion there is has turned out to be minimal. All I did was drill a fairly small hole in the big black filler cap on the reservoir and put a short piece of small tubing pointing straight up to prevent any leaking from it and it's been fine since.
 
#6 ·
Added another underhood pic from a different angle.
 
#8 ·
Funny you should ask. I actually did quite a while ago when I ported it. Used some high temp ceramic paint and did the manifold and TGV housings. I cleaned em up and just painted over the factory red and it just so happens to match nearly perfect.
 
#10 ·
Another pic added showing the inside of the core comparing the air and water passage sizes.
 
#13 ·
I can't believe there isn't more interest in AWICs for our cars. My last car was a Tiburon, and pretty much everyone who was supercharged used an AWIC to cool the intake charge. Anyway, just a couple questions about your setup, since I might do something like this myself eventually:

- What supports the intercooler? Did you fab up some brackets, or is it only supported by the hoses?
- How did you mate the BPV? It looks like you are using the stock one, but what kind of adapter was necessary?
- Did you have to remove the front bumper beam to fit the heat exchanger?

Thanks in advance!
 
#14 ·
I can't believe there isn't more interest in AWICs for our cars. My last car was a Tiburon, and pretty much everyone who was supercharged used an AWIC to cool the intake charge. Anyway, just a couple questions about your setup, since I might do something like this myself eventually:

- What supports the intercooler? Did you fab up some brackets, or is it only supported by the hoses?
- How did you mate the BPV? It looks like you are using the stock one, but what kind of adapter was necessary?
- Did you have to remove the front bumper beam to fit the heat exchanger?

Thanks in advance!
As of right now the core is strictly supported by the hoses and has been fine but I am actually in the process of fabbing a bracket for it as we speak. The BPV is the stocker. The charge pipe has a 1" nipple on it and a flange with the same nipple and the BPV is bolted to the flange. There's a short coupler for the nipples. I had to rotate the outlet for the BPV 180* and trim the length of the return hose but it's basically the same setup used on a front mount. Currently I don't have a bumper beam in place and it's mounted with a fabbed bracket but since this is basically a prototype it's fine for me at the moment. I did find that there is plenty of room to mount the exchanger with either a JDM beam which I have or even the stocker with some clever trimming and a little planning ahead though I don't plan to go that route. I'll end up reinstalling the JDM beam and adjust the mounting to fit that and call it good for me. The other option is to use a different sized heat exchanger but how small is acceptable is yet to be established and probably depends on the application be it a DD, track, drag etc.

T-G
 
#15 ·
COPIED FROM MY JOURNAL:

Since it's been a while I thought I'd update all interested parties on my air to water intercooler project.

My initial impressions after I got it up and running were all positive. The system runs flawlessly so far other than a small seep on a fitting on the heat exchanger which isn't a fault of the components but a fault of mine since I forgot to use the teflon tape on that particular one.

Water temps have never gotten higher than 6-8* over ambient no matter how hard I flog it and with the water/meth the intercooler outlet temps consistently go down to 5-10* below ambient depending on turbo outlet temps. I turned the boost up a little to 23spi and had turbo outlet temps up 193* and with the ambient temp at 95* the intercooler outlet temp was 86*. The core is always cool to the touch other than when I let it sit and idle with the hood closed for long periods with the water pump turned off.

I initially had a full bottle of Water Wetter in the system but most of that went away when I partially drained the system to fix the leaky fitting. I now have what I would guess is about a 1/4 bottle of the Water Wetter and there has been a slight change in the temps but looks to be only about 1*. I think the optimum mix will be about 1/2 a bottle for the volume of water I have.

I had originally debated on whether it was worth it to drop the coin on the Meziere pump at ~$230 with fittings and almost got the Bosche pump at $119 with hose barbs instead but I've come to the conclusion that the Meziere pump was well worth the money and I would absolutely go the same route again. It moves water with authority and seems to flow right around the optimal amount of water to keep things cool without flowing so fast the water doesn't get a chance to hang around and move some heat in the core and exchanger.

At this point I intend to clean up the install gradually, mostly with the heat exchanger. At the moment it's mounted using a bracket I fabbed up that goes in place of the bumper beam but I would really like to mount it with my JDM beam or even better an Oswald aluminum beam. I'm waiting on Killer B to finish the grill insert for the 06-07s and that'll go in place in front of the exchanger. Once I have that I'll start the process of moving the exchanger so the beam can go back in. I also had plans to add a fan behind the exchanger but at this point I haven't seen any need for it. The water never heats up to the point of needing to use a fan and even when it creeps up a little bit from sitting in traffic it comes back down very quickly once the exchanger gets some flow.

The other bit thats going in soon is my super secret cooling idea that may keep the water below ambient by up to 15* any time the pump is running. Works off of 12v and has one moving part in the form of a CPU type cooling fan. Not quite as good as ice in the reservoir but perfect for a DD with an AWIC. Hypothetically if the water is 10* below ambient and another 10* is removed by the water/meth and then you adjust for some fudge factor I could have 70-75* intercooler outlet temps on a 90* day.

So with all that said I'd like to find all those people who said these things don't work on a DD and that I'm wasting my time and money and have them look at my results. I've got numbers-they've got opinions.
 
#333 ·
So with all that said I'd like to find all those people who said these things don't work on a DD and that I'm wasting my time and money and have them look at my results. I've got numbers-they've got opinions.
man, that's some genius shit right there.
 
#16 ·
id imagine that your super secret thing is a peilter? id be a little sceptical sbout that but its awesome that your trying new things. I am curious though did you use your super awesome temp sensor with your A2A intercooler? I have been putting in serious research into this, and my first thought about this was water can put 5x the heat out of your body than air can, so why not an intercooler? Your results are promising. What is your water temp in the resivour, and after the intercooler?
 
#17 · (Edited)
i guess what im asking what is water temp before the radiator and what is it after? what is your relative humidity? Thanks! if you think about it a awic makes sense for a dd, bc how long are long are you actually at 20 psi? redlight idling is just cooling down the res again. and dd you dont really care about +5 - +10 lbs increase either.. thumbs up
 
#23 ·
Where is your meth being injected? Post turbo pre awic? Or pre manifold post turbo? I figured you'd drill the flange where the silicone couple rides either tb or ic and you could cover it with an undrilled coupler.. what stopped you from exceeding 24 psi? Do you feel that as you flog it the resivor builds heat as you described, does it level off at the 10 degree mark or have you backed down before it could get any hotter? I'm wondering if the system has a heat capcity (and will continue to build heat after the thresold) or if the water has a max heat increase of 10d no matter how much you throw at it.
 
#25 ·
Meth nozzle is in the IC outlet side end tank. If you look at my journal post #215 you can see it on the passenger side next to the outlet. I limit boost pressure for several reasons. The stock MAP sensor stops reading a little over 23 psi but looses resolution the closer you get to it's limit. I'm also at the limit of the intake diameter on the 65mm APS and this is manifested in nearly maxed out MAF voltage at full boost. The other reason is due to still running the stock pistons. My tune is pretty conservative especially with the timing and I'm not worried about any knock issues but I still keep power on the safe side for reliability reasons. My boost limits have nothing to do with keeping the water temps in check.

The system naturally has a limit to the amount of heat it can manage but so far I haven't found it. I believe the limit is quite a bit higher than what will be experienced on a street car at what would be considered normal power levels on a setup like mine. Once the heat capacity is exceeded the the system would continue to heat up if more heat is continuously added. However the larger the temps delta between ambient and the water the more will be transferred by the exchanger.
 
#27 · (Edited)
TG, this build up is unique and could change some things for us subaru guys. thx for taking the time and using your knowledge for awesomeness. LOL

keep up the good work!

subscribed
When your bored and refuse to believe everything the internet says projects like this are the result. Thanks for the props. As for changing the Suby world in some way that remains to be seen. I'm still waiting for two more willing victims to build one of these since three data points will constitute a trend.
 
#29 · (Edited)
Well, I will be doing the same, although with a very high hp platform. But we will see how it does. The intercooler is a bit larger...not sure if the heat exchanger is the same size or larger. I am going to see how the lightning OEM pump works. Still trying to decide how to wire up the pump. I can't seem to find a thermal switch in the correct temp range. Looking for something around 85-90*F. Everything I am finding seems to be much higher. Worst case I just do an ignition based switch.

As of right now I only have one intake air temp sensor. I am going to try to get another so I have pre and post intercooler.
 
#30 ·
Have you thought about using a pressure activated switch? I've got an adjustable one thats used with my JDM auto IC spray button and it works like a charm.

I do have a suggestion though. I think you would be better off with a pump that can be run continously. During cruise and light boost the core and the water in it gets heat saturated to a varying degree. My turbo outlet temps during cruise are always 115-125* depending on ambient temp. As soon as you hit high boost with it's associated high turbo outlet temps your sending very hot air through a hot core with hot water in it. It takes a little time to move all the hot water out of the core and the cooler water coming in has to not only remove the heat from the charge air it has to remove it from the core as well and the core is a pretty big chunk of aluminum to have to cool down in a hurry. This also leaves open a window of time where you have higher charge temps with the associated issues that come with it until the temps come down.
 
#31 ·
The lightning pump can be used continuously. In fact I believe it is also switchable from low to high speed. I am not sure how this worked from the factory (guessing varying voltage) but my partner Andy has a lot of experience with the Cobras and Lightnings. Ideally it would be nice to have it at low speed for continuous use then switch to high flow when temps climb.
 
#33 ·
The lightning pump can be used continuously. In fact I believe it is also switchable from low to high speed. I am not sure how this worked from the factory (guessing varying voltage) but my partner Andy has a lot of experience with the Cobras and Lightnings. Ideally it would be nice to have it at low speed for continuous use then switch to high flow when temps climb.
That would work fine I'm sure. As long as there's some water moving the core won't heat soak. Didn't know the Bosch pump could run continuously. Handy to know.

Wirelessly postedI'm considering it, but I want to make sure the core I get has enough flow for whatever setup I decide to go to. There are some vvts that are gt37 and larger sized that would work ok with a 2.7L setup.
I've got 100cfm of headroom on my setup with a 52lb wheel. The core I used won't flow enough for a wheel much bigger though. A 65lb wheel on a 35r is 850cfm so a bigger core would be needed. Problem with that is fitment. The one I used is the biggest one that'll fit in the space between the TB and firewall in that particular configuration. I'm sure with some different routing and a different shaped core a bigger one would fit though. Just not sure how to go about it ATM.
 
#37 ·
Any plans to take it to the track and give it a good 30 minute flogging just to humor us track guys to see if this is a viable option? :)

Did you block off the scoop? I really want to ditch the scoop.
 
#41 ·
Any plans to take it to the track and give it a good 30 minute flogging just to humor us track guys to see if this is a viable option? :)

Did you block off the scoop? I really want to ditch the scoop.
Possible track flogging in the spring. I have a friend who is an HPDE instructor in Houston.

I kept the scoop and have no plans to get rid of it. There's no reason to keep it though. I just think it's part of the character of the car. I do plan to add one of these though:

WRX/STI Hood Scoop Air Diverter Plate

How much did this set up cost you in just parts.
About $1200 after it was all said and done. However that doesn't include the temp meters from Mach V. Those were an extra $500 for the pair but aren't needed for the whole thing to work.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top