intercooler tank design question

[dwilliams]

I've been looking at intercoolers lately. I thought I knew something about how air flows, but looking at the tanks of most of the "racing" and OEM intercoolers, it looks like most of them just shoot the air across four or five tubes and out, with the rest of the core in dead air. Air-as-I-understand-it doesn't naturally turn 70 or 80 degrees, jog over nine or ten inches, and then turn another 9 degrees to go through a tiny passage...

If the tanks the big enough they'd be plenums and it wouldn't matter much where you fed them or what they were shaped like. But many racing and most OEM tanks are even smaller and closer to the core than water radiator tanks.

I don't know how you'd deal with an intercooler on a flow bench. Cutting big windows in the tanks, covering them with Lexan, and using smoke trails, maybe... you could rig up a bunch of thermocouples on the core and see how it actually performed while on the road, but that would require a whole, running engine and a car to put it in. Which would be the true test of intercooler performance, but a real hassle.

Surfing the web I found a bunch about core types, but almost nothing about tanks. Though I did note the "air to water" guys usually make tanks much more in line with what I would consider reasonable.

So, if you're making intercooler tanks, is it reasonable to try to balance the air flow evenly across the core, or is it one of those things that doesn't make any difference in practice?

[j-c-c]

On my ATA intercooler, the internal ends of the cross flow air channels were basically flat 1/2" surfaces. To improve things, I epoxied varying widths of half round alum shapes to those flat surfaces to try and equalize some what the air flow thru the IC based only on intuition, thinking the inlet primarily directed most of the air flow at only the center tubes. Figured no matter what, it couldn't hurt, and I felt smart.

[SchmidtMotorWorks]

I designed a manifold with an internal intercooler.

One of the optimized parameters was the distance from the intercooler to the floor of the manifold.
The conflicting goal was to have as low overall height as possible and as much flow as possible.

I flow tested a range of distances and graphed the results.
For an intercooler about 180mm wide the flow curve leveled off (IIRC 90% of maximum) about 15mm gap.

[gunt]

I ALWAYS THOUGHT THE SAME NOW I SEE , some of the big names showing internals , with fins inside to dierect air up and down , its one of the only ways i guess

[ptuomov]

There was a long thread on a related topic where people answered many of my questions:

viewtopic.php?t=39841

[j-c-c]

There was a long thread on a related topic where people answered many of my questions:

viewtopic.php?t=39841

Yes, that was an informative discussion.

[LoganD]

Any charge air cooler designed in the last ten years for an OEM application is going to have a very large amount of CFD done on it. This does often yield counter intuitive designs, and the OEMs are generally favoring charge air cooler utilization over minimizing total pressure drop. In other words, they want maximum temperature drop across the charge air cooler rather than maximizing the charge air cooler's flow rate. This is because extra boost is easy to come by, but keeping temperatures in check is critical for detonation resistance. During this CFD process they will focus heavily on "balancing" the flow across the cooler, and generally you'll deliberately restrict flow through the tubes at the edge of the charge air cooler as they have less heat transfer capability.

Generally the charge air cooler and the corresponding ducting will have more CFD and flow bench testing than the intake manifold and cylinder head combined. Very, very large durability and performance gains can be gained through enhancing charge air cooler system performance. It's also absolutely critical to keep temperatures in check as BMEPs keep increasing in OEM engines.

Now, aftermarket charge air coolers are generally garbage. They almost always have terrible cooler utilization and are focused on maximizing size and flow rate. While some large CFM number looks great for selling stuff, the overall driving experience diminishes when you have too large of an inlet volume to pressurize. Keeping volume between the compressor outlet and the intake valve to a minimum is absolutely critical to throttle response, and it's why water to air charge air coolers now dominate the OEMs.

[ptuomov]

Any charge air cooler designed in the last ten years for an OEM application is going to have a very large amount of CFD done on it. This does often yield counter intuitive designs, and the OEMs are generally favoring charge air cooler utilization over minimizing total pressure drop. In other words, they want maximum temperature drop across the charge air cooler rather than maximizing the charge air cooler's flow rate. This is because extra boost is easy to come by, but keeping temperatures in check is critical for detonation resistance. During this CFD process they will focus heavily on "balancing" the flow across the cooler, and generally you'll deliberately restrict flow through the tubes at the edge of the charge air cooler as they have less heat transfer capability.

Generally the charge air cooler and the corresponding ducting will have more CFD and flow bench testing than the intake manifold and cylinder head combined. Very, very large durability and performance gains can be gained through enhancing charge air cooler system performance. It's also absolutely critical to keep temperatures in check as BMEPs keep increasing in OEM engines.

Now, aftermarket charge air coolers are generally garbage. They almost always have terrible cooler utilization and are focused on maximizing size and flow rate. While some large CFM number looks great for selling stuff, the overall driving experience diminishes when you have too large of an inlet volume to pressurize. Keeping volume between the compressor outlet and the intake valve to a minimum is absolutely critical to throttle response, and it's why water to air charge air coolers now dominate the OEMs.

While I agree that new intercoolers in cars like BMW M5 (TU) are great, I wouldn't dismiss all the aftermarket or home fabricated coolers. Some of them are good.

Furthermore, my (limited) personal experience is that every time we make the boost pipes bigger, the throttle response improves despite the increased volume. By taking away restriction in the intake pipes, the turbo doesn't have to be working as hard for the intake valve to see the same pressure. I'm not saying that no extreme size would be too large, but my (limited) experience is that in practice all the intake pipes should be as large as what you can fit under the hood.

[volodkovich]

I've done CFD on standard style aftermarket IC end tanks and they are indeed horrible. I suspect they reason they work is because the extra mass of core conducts heat away from the cores that are actually passing the air. The end tanks with a side entry I have designed used a guide vane to split the flow into each row on the core (as most flow to the outside row and bypass the inner) and had a nice diffused entry to utilize the whole width of the core.

[4vpc]

Corky Bell's Maximum Boost is well worth a read for the design and maths, Forced Induction Performance Tuning by AG Bell is a close second.
Most production cars and indeed aftermarket run air-air coolers, water cooled have always been less popular due to complexity, price and weight.

[ptuomov]

Corky Bell's Maximum Boost is well worth a read for the design and maths, Forced Induction Performance Tuning by AG Bell is a close second.
Most production cars and indeed aftermarket run air-air coolers, water cooled have always been less popular due to complexity, price and weight.

Almost all the new pricey cars I look at have Behr water to air to water intercoolers and a low temperature coolant circuit. Sick design in that counterflow cooler.

[LoganD]

Any charge air cooler designed in the last ten years for an OEM application is going to have a very large amount of CFD done on it. This does often yield counter intuitive designs, and the OEMs are generally favoring charge air cooler utilization over minimizing total pressure drop. In other words, they want maximum temperature drop across the charge air cooler rather than maximizing the charge air cooler's flow rate. This is because extra boost is easy to come by, but keeping temperatures in check is critical for detonation resistance. During this CFD process they will focus heavily on "balancing" the flow across the cooler, and generally you'll deliberately restrict flow through the tubes at the edge of the charge air cooler as they have less heat transfer capability.

Generally the charge air cooler and the corresponding ducting will have more CFD and flow bench testing than the intake manifold and cylinder head combined. Very, very large durability and performance gains can be gained through enhancing charge air cooler system performance. It's also absolutely critical to keep temperatures in check as BMEPs keep increasing in OEM engines.

Now, aftermarket charge air coolers are generally garbage. They almost always have terrible cooler utilization and are focused on maximizing size and flow rate. While some large CFM number looks great for selling stuff, the overall driving experience diminishes when you have too large of an inlet volume to pressurize. Keeping volume between the compressor outlet and the intake valve to a minimum is absolutely critical to throttle response, and it's why water to air charge air coolers now dominate the OEMs.

While I agree that new intercoolers in cars like BMW M5 (TU) are great, I wouldn't dismiss all the aftermarket or home fabricated coolers. Some of them are good.

Furthermore, my (limited) personal experience is that every time we make the boost pipes bigger, the throttle response improves despite the increased volume. By taking away restriction in the intake pipes, the turbo doesn't have to be working as hard for the intake valve to see the same pressure. I'm not saying that no extreme size would be too large, but my (limited) experience is that in practice all the intake pipes should be as large as what you can fit under the hood.

Have you actually measured the boost response? I suspect this is the placebo effect, increasing volume is a direct boost response loss. The turbocharger has a set mass flow rate relative to wheel speed, you aren't affecting the exhaust pressure drop across the turbine because the intercooler piping doesn't affect the exhaust, and you're increasing the volume that needs to pressurized. Now, if you're installing a higher flowing exhaust system post-turbine along with these larger intercooler pipes then I believe it, but it's the exhaust that's helping response not the induction piping.

You can improve boost response by increasing the radius of the bends in the induction system or reducing the number of bends, but increasing the diameter will definitely make response worse. If you've got 80mm intercooler piping and you increase that to 90mm, you've increased the intercooler piping internal area by over 26%. That's huge. You should be very, very careful about increasing piping size.

[BLSTIC]

It could also be because if you get rid of 2psi of intercooler restriction the turbo is spinning significantly slower to produce the same manifold pressure, so it takes less time to accelerate the turbo to that speed. That would cause a throttle response improvement completely independent of intake system volume.

[ptuomov]

Have you actually measured the boost response? I suspect this is the placebo effect, increasing volume is a direct boost response loss. The turbocharger has a set mass flow rate relative to wheel speed, you aren't affecting the exhaust pressure drop across the turbine because the intercooler piping doesn't affect the exhaust, and you're increasing the volume that needs to pressurized. Now, if you're installing a higher flowing exhaust system post-turbine along with these larger intercooler pipes then I believe it, but it's the exhaust that's helping response not the induction piping.

You can improve boost response by increasing the radius of the bends in the induction system or reducing the number of bends, but increasing the diameter will definitely make response worse. If you've got 80mm intercooler piping and you increase that to 90mm, you've increased the intercooler piping internal area by over 26%. That's huge. You should be very, very careful about increasing piping size.

Not rigorously, so it could be just that the bigger pipes make it sound louder and by human perception louder means faster!

I don't even have a rigorous definition boost response. What I mean by boost response is that if I'm leaving a stop light how quickly can I get to 60 mph or if I'm cruising at 60 mph how quickly can I get to 90 mph without downshifting.

If you put a straw between the compressor outlet and the intake valve, I believe that by my definition "boost response" is bad. The car might never make it to 60mph! Making the straw larger makes boost response better, again by my definition.

Now, the question is how big is too big. If a five-liter motor displaces about (say 100% VE) 2.5L per revolution, at 1500 rpm it displaces 10x that per second. Make the pipe shorter, sure, it's better. How about for diameter, what's the optimal diameter?

[LoganD]

It could also be because if you get rid of 2psi of intercooler restriction the turbo is spinning significantly slower to produce the same manifold pressure, so it takes less time to accelerate the turbo to that speed. That would cause a throttle response improvement completely independent of intake system volume.

That's not how it works. "Boost" response and turbine response are not the same thing, modern turbines have excellent RPM response. In other words, they'll gain RPM very quickly, particularly when clever use of the wastegate is employed. You also can't think of the compressor in terms of boost, the compressor supplies a mass flow rate and boost is merely a consequence of that flow rate.

Think of of it like this: if you have two air mattresses and one compressor to blow them up with, but one mattress is 25% larger than the other, which will fill up quickest?

As long as you are keeping the velocity in the intercooler piping at a reasonable level you will not have large pressure loss in the system. For example, a common goal for a complete compressor outlet to intake valve system would be 15 kPa on an OEM project. The charge air cooler would be 6-8 kPa of that depending on the design. Changing to larger piping post compressor might get you from 15 kPa pressure drop down to 13 kPa pressure drop. That is not even remotely worth a 5% reduction in boost response, particularly if you have some surge margin in your compressor. The entire turbo system becomes MORE efficient if you can run with less exhaust flow going through the wastegate, so if you have surge margin you simply up the pre-charge air cooler boost level (2 kPa increase is nothing anyway) to achieve the same boost level in the manifold. This closing of the wastegate actually helps turbine response.