I agree with that. As an illustration, if the air spent zero time in the intercooler it wouldn't be cooled at all.

I don't know if the EVO Race intercooler described above is any good, but from the dimensions given the air passing through it will spend three times as long in the EVO Race intercooler as air passing through the standard intercooler. The mass of air passing through each tube will be ony 81% of the standard intercooler.

With low air demand from the engine and on low boost, the difference in performance would be minimal, but as the intercooler is expected to cool more mass of air from a higher temperature, the larger (slower?) cooler should be significantly better at rejecting heat.

With the vertical flow RX7 cooler, the air will only spend 90% of the time in the tubes compared to the standard cooler, but the mass of air flowing through each tube is on 31% of that flowing through the standard cooler tubes. For that reason it too should be very effective.

The improvement from spending longer time in the cooler tubes is not linear but exponential, so the law of diminishing returns applies, just as it applies to increasing the tube length or total cooler area.

Out of interest has anyone come up with any novel charge cooling ideas?

For example, would increasing the surface area (not necessarily the volume) of the connecting pipework between the turbo and throttle body be of benefit? I.e finned piping. Or what about a refrigerant system surrounding a section of piping? You'd have a series of venturies encompassing the piping, you'd then accelerate a fluid through said venturies to pull the heat from the charge air.

It just seems to me that unless somebody spends a large sum on R&D on intercoolers then they're probably as good as they're getting. The desirable characteristics are at odds with each other. On the one hand you want great flow, low pressure drop (part and parcel of the same thing) which requires minimum restrictions and a mirrored surfaces to reduce impingement. But that then reduces the coolers ability to desipate heat as what you need is a high surface area to volume ratio which means rough surfaces and multiple narrow tubes.

Martin, I presume by your lack of reply that you aren't sure?

No. It meant I couldn't be bothered to answer. I am certain you know that a flow bench measures differential pressures and derives a flow value, so what's the point in answering?

* I should add, the flow benches that I know use two manometers. Other designs may use other methods, but the 'two pressure reading' types are perfectly valid in establishing flow rates.

I'm afraid that's completly wrong, as is a lot of which you have written in this thread. I'm sure it's very impressive for the majority to read your numbers but most people don't know any better. There is no way to calculate numbers from the infomation given, not real "honest" numbers, there is no calibration shown so your numbers are null and void I'm afraid.

I will repeat again, this is a simplistic test that shows retriction, nothing more, nothing less.

Re: Flow Bench:

OK - in that case I would be keen to learn how a flow bench measures flow - please tell us.

Re: The Tests:

I agree that the flow tests conducted doesn't show much - they are however interesting and even thought provoking. They show the relative back pressures, but the percentages mean nothing as they don't relate to total flow ability of any of the intercoolers - only the comparative difference - hence my offering of what 15cm water head means in terms of "real world" back pressure, which in this case is the same as pressure loss across the core at around 380CFM.

Re: My Numbers:

As for comparing the intercoolers described, it is quite straightforward to give 'ball park' accurate comparative figures for the heat exchangers, though I would not pretend that I could predict/calculate actual operational results such as what the output temperature would be given x input temperature, y ambient temperature and z cooling airflow mass.

If asked, however, I could predict the order of performance of the four designs - and would be happy to be proved wrong. It would, though, require someone to conduct the empirical testing.

Summing Up:

I am am guessing that you feel yourself mathematically challenged?

p.s. I'm not here to impress anyone - just to dicuss and contribute to the discussion.....

It's obvious you have no idea how a flow bench works, this I can see plainly. I'm not here to explain to you, that's for you to learn on your own. What I will say though, they have a calibrated orifice, that calibration is done at a quoted pressure, do you see that in the test? I know you don't! If it was as simple as the test to make a flow bench I wouldn't of spanked the amount of money and effort I have to make one.

Having spoken with Bruce who runs PTS flowbench parts/forum and shown him this thread you can be sure your wrong in your opinion. He confirms the test only shows a mind squint at restriction like I keep telling you, nothing more nothing less. Don't make me quote him or rant on any longer.

Sure, I'm not the best at maths but spotting BS I'm top of the class ;-)

I think you just like to be Mr Angry for some reason!

The second manometer on a flow bench measures the pressure in the metering plenum which, as you say, vents through a calibrated orifice. When trying to measure flow, that setup is essential as you need the two pressures to calculate the flow through the test piece. What the OP did was not trying to measure flow, only to measure back pressure, which he achieved with one manomenter. Not an opinion - fact.

Not Mr Angry, just Mr Likes it to be right.

Best check with Bruce again then. Mr Vizard would seem to concur with me on this one......

If anyone would like to check my numbers (apart from 1NRO who would be totally lost) just ask by PM - I'm happy for anyone to verify them.

Me too. I don't post what I'm not sure of - unless I am asking a question. Perhaps, with that in mind, you would try to untangle for me this mess of a post you made earlier?




As to me it is confusing and meaningless....

p.s. what is a 'mind squint'?

Throw the spade out Martin

I never was any good at cryptic clues. Please explain?

The hole your in is deep enough

I'll agree to differ - I hope you do the same. It's the adult way....

how a thread that looks interesting then turns into a cat / pussy fight.

I recon it's ben tried somewhere, by somebody. Problem is (or would be) the massive variation in airflow seen in any engine, any kind of system that uses refrigerant basicly 'pumps' heat from one place to another rather than physicly cooling it by normal heat transfer. On something like a fridge or aircon unit where the cooled air is basicly low flow or static there isn't an issue but on the high air speeds seen in engines there's no way to directly effect it. You could add a heat exchanger that the charge passes through, then refrigerate the otherside I guess, but isn't that the same restriction as an intercooler :-)

Iced up charge coolers are as good as close as it gets, but pretty useless on anything other than a 10 second drag car.




Yeah the R&D is done for intercoolers or should I say aftercoolers maybe ? For all intents and purposes they work just fine but you can't change physics, no such thing as a free lunch !

C'mon now, Paul's done nothing wrong

- Just going back to this post ^, I used to think along the same lines but in reality you should look at it from the other side of the coin.

Choosing a core based on efficiency alone is easy but exceptionaly bad for restriction - just pick the longest tubes with the most dense turbulators. Choosing a core based on an acceptable flow loss, total engine cfm and pressure can be calculated and from that inlet temprature can be defered from the compressors efficiancy point. We know ambient temps and therfore the temprature difference, so the only un-known is the individual core's eff. % And as Nik says all intercoolers cool air, thats what they do.

As promised here are the pics of the internal tubes of the HDI cooler, as the only aftermarket unit - it's the third type that Kayjay mentioned with internal "fins"



Branded - and their blurb

The gt2 offers a strong extruded tube and fin core with 16 rows for maximum flow and cooling efficiency. our latest design incorporates tado fins, Transverse aero dynamic offset fins. This allows air to directionally pass between the fins to improve cooling plus allows plenty of air to reach the vehicles radiator. The tubes also incorporate internal fins to further improve cooling efficiency. It can handle 800 horsepower,so is suitable for many motor sport applications.



Original coupe



As I say all the cores I tested were of similar internal design rather than being the first and second type in Kayjay's pics, the others aren't here to photograph but as I say being OE are the same type also

Cheers

Looks slightly different to the oem to me. The oem are more 'chaotic' inside, that first picture of yours still provides for straight paths for the air. I think the intercoolers with fins that are in short-distance direct contact with the tubes' 'skin' are better at cooling.

I'd still say the cooling efficiency is very important as long as a certain 'minimum' flow is achieved, which I'm sure all intercoolers provide. Tricky, completely get what your saying, but I'd LOVE to see actual figures of intake temps comparing intercoolers. If there'd be 10 degrees difference that's a lot of potential poser that can be used. Pretty sure we're not going to see those differences, maybe 5 or 6 degrees would be big, but still, would like to see what these babies actualy do, or whether you can just pick on based on price or ease of installation.

NOTE: about my 'behaviour' post, this was a warning before things got personal. Let's keep the emotions out of this thread and just keep a healthy discussion, whether it is based on physics of gases, aerodynamics or experimnts at home, it's not about the blahblah, it's about the conclusions. We all have to start somewhere. Everyone is encouraged to share info, whatever it is. If anyone wants to prove any info wrong, back it up and don't take it to personal levels. It's what THIS community is about AND how it came to life.

I did wonder about air speed, but thought a suitably cold bit of piping should do the job.

The way I see it you would have a larger diameter of pipe in line that the charge air would enter into. This would slow the air sufficiently to allow the air to transfer its heat via the refrigerant system encompassing it. The then cooled charge air would then continue on its way through a slightly reduced diameter pipe to maintain net flow. You'd pump the heat out to a radiator pack in much the same way as aircon. The beauty of this system is that you could pump the refrigerant using an electric motor and then control that to maintain inlet temperature.

This all could be the ramblings of a mad man, but there has to be a better way of cooling and maintaining flow.

The beauty of the current system is that it is rather passive and efficient. The large diameter of the pipe would be quite bad for cooling though, you'd better have constant diameter and long surface area. And even so, with just one big tube you'd be looking at very negative degrees to get good cooling of the air rushing through it. And active system needs energy input, which has to come from the engine.

Did someone ever try putting dry ice in the air box? Or is that just silly because you're also introducing CO2 rather than oxygen? Of course, something like that wouldn't be permanent, like wetting the intercooler with icecold water isn't something you can do permanently.

Here is another aftermarket core, but bar and plate I have on the shelf - meant to be of resonable quality

Bit more chaotic, as each internal fin is "stepped"



And inside



For my interest has anyone got a pic of the inside of the proalloy unit that seems to be the most popular upgrade?

Again though if I use the core above to build a vertical flow I would expect to see a big drop in resistance based on end tank area alone. The internal design is similar to a chaotic OE core - in this case the tube lenght is 175mm

Yes, so that last picture looks indeed like what I meant. I'd imagine THAT would be a bit restrictive for flow as the air is hitting fin sides all the time?

Didn't dare to put no question mark behind that sentence because of all the pro's on here, as you certainly understand...

In the most simple heat transfer equation

H=UA(Delta T)

Yes Delta T is just as important as U (heat transfer coefficient) and A (area) but we can’t change it. The ambient air is what it is and the charge out of the turbo is what it is. All we can do is change U and A to get the desired air temp out the other end.

U for a given flow channel design is predominantly influenced by velocity. Sure velocity is speed.

However a higher speed (short residence time) will give better heat transfer than a lower speed (longer residence time)

These rules are for fluid flow in a heat exchanger. A heat exchanger operates under a flow regime with a high Nusselt number. I.e. the convection of the fluid (heat transfer within the fluid) is magnitudes higher than conduction between the fluid and the boundary to the heat exchanger wall. Heating a large pot of fluid for example is a different regime (very low Nusselt) and the convection within the fluid is significant. This regime will require a long time to heat or cool.

I'm sorry, I can't let that go and say nothing. Please refer to my post above.

That is the point of this test, the RX7 has that same type of core, and cores I have here on the shelf which I plan to use have that same chaotic core.

I guess this test comes down to endtanks, and the size of tank areas available with vertical flow IC's. The resistance test I posted speaks for itself. 2.5cm deflection for the twin RX7 vs 14cm for the Evo 6.

I've taken this below from Scuderia, who I hope will still be speaking to me

Chaotic tubes increase turbulence for higher efficiency – this maybe desirable because the flow rate is low. i.e. The system maybe optimised for general road driving, low RPM and road speed etc. However as road speed and flow rate increase they get way too restrictive. You can make the intercooler far less efficient and it will still be adequate at cooling because of 2 reasons.

1. RPM is higher so flow rate is higher and turbulence is now higher anyway.

2. Because higher road speed gives more air to cool the charge (higher U) and top it off you can simply subject the charge to more surface area (bigger intercooler, longer tubes) it’s going to do more cooling.

Just keep in mind if you make the tubes less chaotic for a higher flow rate and you want to retain the same cooling capacity the only option in the real world is to add more surface area. i.e. longer tubes.

If you add more tubes with the same chaotic fins you will achieve basically the same thing. As the flow rate decreases through each tube, each tube becomes less efficient (less cooling per tube) but there are more of them to make up the difference.

I am hoping these cores I have will make a very efficient FMIC, and with a high number of tubes will have very low flow resistance.

As Scuderia says the other option is longer tubes with less choatic internals for the same cooling, as per the HDI core but the flow resistance was higher, and was no where near on par even with the internal design as per the photo. I appreciate it is a mix of tube lenght, internal turbulators,and number of tubes, end tank area but for my build, the space available and the endtank area the optimum as I see it will be a vertical.

I personnally am sold

Hope it is of interest

Cheers