Hi All, (THIS IS WHAT I SHOWED TO NATHAN AT MITM/MTTS 2010) With the our latest R56 Intercooler development now completed, we've turned our efforts back toward the R53. What you see in the picture above is the product of realizing that: 1) Water-to-Air intercoolers (W2A-IC's) don't have super-stellar recovery time after heat soak (parked for an hour after a hard run) on street driven R53's. W2A-IC's are great for track driven R53's, but you need to allow a slightly longer time for recovery when on the street. 2) The R56 is a much better candidate for a W2A-IC on the street and at the track. (We used our R53 research as our baseline and the R56's recovery numbers kicked its butt at the track and on the street -- Sad but true.) 3) We always knew that the R53's intercooler is in a crappy location, but we needed a way to cool the charge air effectively while creating a completely reversible, easier-to-do mod. 4) The GP intercooler works great, but we've always thought that there was a way to use a smaller volume to do the same amount (or more) of charge cooling. 5) Removing your intercooler to do minor maintenance (like changing your plugs) is (more than) annoying at times. With these factors in mind we put aside all our R53 W2A-IC fun and embarked on the path of Air-to-Air intercooling (A2A-IC). This took-up the better part of this summer, but it was well worth it. We ended up building an A2A-IC that gave us the performance that we were looking-for with quick recovery on the street and track. How did we do it? A) We did a TON of materials and suppliers research and found that copper has a thermal conductivity of 231 Btu/hr/ft. Aluminum has a thermal conductivity of only 136 Btu/hr/ft. This means that copper can conduct heat 59% more efficiently than aluminum. B) We fabricated several core designs using copper fins, copper turbulators, and brass tubes with a very high copper content (85%). The space available for the IC limited us slightly, but we created a core that outflowed and has less pressure drop than the stock A2A intercooler and the GP intercooler. C)We did a major study of the stock A2A intercooler and GP intercooler to see why they both worked fairly well given their constraints. Matt Richter ("Dr. O" of MC2 fame) always likes to point out how the intercooler tubes are crimped in a way so that the air is ramped to minimize disturbance. However, I noticed there are 2 major areas of disturbance at the corners of the endtanks where the air does not flow smoothly (to ease manufacturing). With the factors above in mind, DoS designed special endtanks that have airfoils at the entrance and exit to each tube to promote laminar airflow. We also made the end tanks out of a high performance, high-temp resistant, engineering plastic to create contoured airflow surfaces and minimize heat soak whererever possible. In 100 degrees F conditions, initial datalogging is showing 5-10 degrees F above ambient at cruise and 20-30 degrees F above ambient under full load/boost. (Better than what we expected). As an added bonus, our copper A2A-IC recovers from parked heat soak & after heavy load/high boost conditions 60-to-75% faster than DoS's Gen III R53 W2A-IC kit and 35-to-45% faster than a stock A2A-IC. Projected retail pricing is ~=$900.00-to-$1000.00 I'll keep you posted on the details. These are projected to start shipping in October. -Clint www.defendersofspeed.com | [email protected] | (404) 433-2452
Very interesting and I'm glad to see someone working on solutions for the IC in the stock location. One question I have is that in reading about copper and aluminum it was noted in some of the literature that while copper does conduct heat better than aluminum, aluminum dissipates heat faster/better. I was reading about the materials for making heat sinks and in some of those applications folks would use copper as a base with the fins made of aluminum... Not trying to start anything at all, I'm excited about the prospects of your IC and wondered if you had found that to be true? Nice work! PS: Love to test one for you at an upcoming HPDE!
how are the plastic endtanks connected to the core? the stock IC on the '05+ Subaru Legacy GT and '08+ WRX use plastic end-tanks and they're known to 'pop off'. here's a pic to show you that you guys are in good company (Subaru's are known for efficient TMIC designs with positive pressure from the hood scoop across the core)
Never mind if it works--my god that thing looks nice! I never thought I could get slightly aroused by an IC, but if that thing works as good as it looks, I might be in trouble.... Do you have direct comparisons to the GP intercooler--ie how do the temps at cruise and under full load under the same conditions compare to the GP IC, and how is recovery compared to the GP IC? Does this one use a air diverter like the GP IC, and does that projected price include all the parts and hardware? The DoS IC is not cheap; you can get the entire GP setup for at least couple hundred bucks less than that, but if it significantly outperforms the GP intercooler, it deserves a look.
Q&A: If we were comparing two imaginary intercooler cores made to identical specifications (tube wall thickness, turbulator thickness & spec, and fin thickness & spec), but one was made of copper and the other was made of aluminum, chances are the copper intercooler would take longer to cool down after being heat soaked. However, in terms of steady state heat transfer (like at the track) the copper unit would outperform the imaginary copper intercooler mentioned above, trumping its identically spec'd imaginary aluminum counterpart. All that being said, it's not an 'apples to apples' situation (like the imaginary intercoolers example above) when comparing the copper intercooler that we've designed to the IC units that BEHR makes for the R53 MCS and the GP. The walls on the brass IC tubes, the turbulator fins, and the cooling fins are significantly thinner than the BEHR IC units used on MINIs. This allows for better heat transfer than aluminum yet comparable heat dissipation. We know if the 'pop-off' issue well. That's why we bond the tanks to the headerplates with a high-temp, oil resistant silicone. As a backup, there's a series of torx screws holding the end tanks in place. As a second backup, tanks are also firmly bolted to the factory mounting points. During the early R53 W2A-IC testing . . . When we were comparing the stock IC to the GP IC, it excelled at recovery. Those 2 extra rows made a huge difference in that department, but did little to get temps below the same cruising temps as the stock IC and resulted in greater pressure drop. We haven't done a complete recovery comparison for the GP IC -vs- the new DoS unit, but we will post it once we do same day/conditions steady state and high load tests on both units. The price includes all parts and hardware, but the new DoS unit does not use a special air diverter on top of the IC. You only need to change the foam seal (on the plastic duct panel that mounts to the hood) to a new configuration that we've spec'd. There will be some slight cosmetic changes on the production units that will have proper mating surfaces for the new foam seal configuration, but all else remains unchanged. Thanks, All. -Clint
Thanks! Eventually, I want to swap out the intercooler, and I was pretty much set for the GP IC, but at this point the DoS IC is looking even better.
When I tested a GP... I measured the same thing... It was interesting to me, but I didn't know if it was unit to unit variations or something in the internal flow.... Anyway, it's good to know that I wasn't crazy.... I did a thread on that over on NAM.... I'll see if I can find it... Matt
So, in rough numbers, and I realize your testing is complete, please share some pressure drop numbers starting with the standard S IC as the baseline. Max Boost Measurement Standard S = 15psi GP IC = ? DOS = ? From what you have stated above I would expect the GP IC to be lower than 15psi, and the DOS unit to actually be higher. And great development work here, thanks for not forgetting about us 1st Gen guys. On my '06 JCW car my IATs are 30F plus Ta for the street and 60+ on the track and climbing each lap. Do you have some track data to share? Do the IATs continue to rise lap after lap as with the stock IC?
It is great to see a new IC coming on line. However, since you clearly designed this, rather than took something off the shelf, why not angle the chambers forward -- to 45deg would be ideal -- so as to allow for a better air flow from the scoop to the rear of the engine bay and exit down past the header? I provided drawings of this some years ago, and posted them on NAM and my website. I even tried to contact some IC core manufacturers, but there was not interest. I suspect you could improve the performance of your copper IC (and I like the plastic end tanks very much) with better airflow of the cooling air, while still maintaining good air flow of the charge air to the intake manifold. Good luck!
I don't know that it would be that much of a winner remember the turbulators on the fins create a bunch of pressure drop, so that the flow through them isn't the same as running air through louvers. The reason that non of the IC companies would touch this one is that it would require a whole different way of tooling up to produce the core for what would be an uncertain benefit. An easier way to achive the same effect would be to have baffles that would re-direct the air into the core, kind of like the two or three fins that are on most diverters. Anyway, I'm glad Clint took the time to do some really bitchen' end tanks, and to make smooth flow into the core for the intake air. This is where a ton of aftermarket ICs drop the ball and create a performance deficit that takes a bunch of other work to overcome. Also, a note on testing ICs and how to measure what they do. Looking at just boost numbers is misleading. You have to look at CFM @ a given air temp through them and then look at the pressure drop. Reason I say this is that the cooling of the IC when run in anger reduces air volume and therefor peak boost, even though it may be delivering a denser charge. The way I got around this in testing was to look at pressure drop vs RPM at no load (cause of the limitations of the testing equipment that I had). I ended up with graphs like this: BTW, I did my TE measurements at redline in 2nd gear from power runs where the car was well conditioned before the run. Even 30 seconds of idling before a run will skew the numbers. Matt ps, this is old data, and the "flow through" was a prototype of the Alta flow through. The Alta TMIC was one with very high turbulator density. They changed the design a bit over the life of the product. And the DFIC was a v1. M7 eventually changed the design to lower pressure drop as well.
Nice dataset Dr O, thanks for sharing that. This seems to be a nicely controlled experiment. It is definitely not simple to keep all of the other variable constant and ensure you are only varying the IC design.
That chart was taken with the car at rest, holding revs for each datapoint until stable. For power runs, I'd take the car onto a freeway loop, hold the car at 60 in 5th gear for about 7 miles, then come to a stop and do the power run with less than 15 seconds of rest time. If it was more than 30 seconds, the heat soaking would skew the TE numbers. To characterize 6 different set ups took about 50 power runs. Good testing is a real bitcch. There is no two ways about it. Matt