I realise there are a few known intercoolers for the coupe such as pro-alloy etc and the mitsi evo unit I run. and lets say 400bhp is enough for most modded coupes

The Evo intercooler is known to become a restriction at circa 400bhp and 2 bar with most mitsi modders swapping it earlier.



Photo of mine above with a core size of 450mmx270mm and 65mm thick - this is made up of 25 of 450mm long tubes hortizontal turbulator tubes.

total frontal area of core 1215cm2

Lets compare to this one (forgetting the ugly inlet and outlet pipe angles)

Still a 65mm core but 600mm long and 170mm tall so total frontal area is 1020cm2, but it is made up of 43 of 170mm vertical long tubes - rated to nearer 600bhp+



They say nearly all the cooling in an intercooler tube happens in the first 15-20cm of the tubulars tubes so the second one should have long enough sections.

infact this treadstone unit (U.K. brand I don't know) rate their similar sized unit efficent to 1300bhp - but it is 10mm thicker at 76mm

http://www.treadstoneperformance.co...prodname=TRV25O9+and+TRV25S9+Intercooler

The second one would actually be a better fit in my coupe. I suspect as with the evo the top half is restricted for airflow by the centre section of the front bumper so very little "sees" airflow. The second one however could be ducted amore easily and with the removal of say centre 4 lower uprights (below numberplate)in the bumper and movement of the plate JDM style I think I could make an acceptable duct.

My car is a toy so not too worried about a slight change from OE and I have a spare bumper



I am struggling to understand sizing

Can anyone explain the cfm airflow relationship (maths formulae) between number of tubes, frontal area, tube depht and lenght etc, or at least some rules of thumb

I realise the second one will give up a little in cooling but pressure drop should be far lower (i have been quoted 1.5psi at 600cfm) not sure on the evo

Both are tube and fin the second being 2 Mazda Rx7 cores welded onto custom end tanks

Cheers

I have the second version on my coupe..
( up to down )

I used that, because of pipe work and as I

read too, it should be better for cooling.

Is the evo one not doing the job anymore ?

It's a juggleing act between % of air charge cooling and cfm flow. There are too many factors to give any rules of thumb, for instance two identical I/C of same construction and size but one with denser internal turbulators. This one would cool the charge air better but give a higher % of pressure drop. Just match one to your output, better to not have a massively oversized cooler for your application either.
It's all about the ducting imo, the evo one you have could be ducted properly and improve it's efficiancy by quite a lot, open out your bumper to get the full effect.
The vertical ones seem to be in fashion at the moment, and the reasons for having them are apparent but try not to forget what the I/C is there for in the first place.

I've got a sum for you

http://naca.central.cranfield.ac.uk/reports/1940/naca-tn-781.pdf

The conclusion is at the end, pretty much says most tubes possible over a not long length, getting towards 15" long looks to be dropping off.
I think unless you drive a truck your stuck with getting the most length you can into the space once the end tank is decided.

Not a new trend Richard, that's a 70 year old paper.

It's just chasing efficiencies and restrictions, I have moved to a larger (easily 400bhp capable) turbo and I have been reading books which is always a dangerous thing

After the discussion on tubular manifolds and the lack of efficency from certain designs I started looking at the coldside as well, I had been told by someone who knows there onions that vertical cores were the way to go - that very long tubulator cores start to add little cooling beyond a certain lenght - but add restriction (law of diminishing returns), thrown about that xx% on the cooling happens in the first 5-6", and it could be 70% or more from opinions but can't back it up.

I see from Nik's article 15" is more ideal so the 7" core is a bit on the short side however likely still adequate.

I have been reading A. Graham Bell "forced induction" and Corky bells book, both of which would suggest 28-32inch of tubulator tube for 400bhp in a 2.5" thick core,

The evo one at 270mm height (inlet side) is 11" for a 2.5" core, the RX7 pair are 24" (over double the area of tube flow than evo) so still a bit on the low side but better.

the question is what cooling would I be giving up for the short tubes, versus gains in compressor efficency and pressure related heating of the charge

Hard to answer I know but like I say rules of thumb on lenght

Thats a bit more than a sum ! My calculator has just crashed !
An interesting paper, it backs up what we already know about how differing dimentions affect pressure drop, but the conclusion does'nt go into enough detail with tempratures. (unless I missed that bit)
I don't think you can make an assumption on a core unless you know compressor outlet temp, which varies with boost level. I suppose it would be safe to say that when the turbo works at it's highest efficiancy point, the corresponding actual outlet temp would be the one to gauge it by. Thus a ratio of charge air cooling to pressure drop could be worked out for any size core, giving a better indication of which would be the one for you. All the core manufatures Iv'e seen just quote cfm values, which are usefull but not conclusive.

An extract from a core makers site, another factor which people forget about imo.

" Vertical and horizontal flow cores that we manufacture have different internal fin
design characteristics. Horizontal flow coolers have long runner lengths which means
higher pressure drop through the core, while vertical cores have short runner lengths and
lower pressure drops, we understand this by changing the density of our internal fin
structure for any given intercooler. Many vertical cores that we manufacture are designed
for larger displacement motors that do not require high boost pressures, therefore we need
a dense internal fin design to extract all the heat we can, as the air passes through 6-10”
of core length. On the other hand many smaller displacement motors 1.6L to 3.0L will
require higher boost pressures 15psi and above, to achieve the power levels we are
looking for. With this higher boost pressure means higher compressor outlet temp, and
we need to cool this air over a longer runner length of intercooler. Making the runner
length longer means higher pressure drop through core, and this is something that robs us
of power. Manufacturing a slightly less dense core than our vertical flow cores, we can
still have low pressure drop through core, while still extracting as much heat as possible. "

I guess what I'm trying to say is one design is not simply just better than another based on one factor, there are other factors. I know when I get round to this myself I will definatly get some help from a couple of different core specialists, send them some detailed drawings of the amount of space availible, turbo size and model and see what they come up with.

Hi Richard,

Which manufacturer does that quote come from?

Treadstone.

sorry to but in on your thread but i am sure youll understand why
i am not well up on intercooler design but am very interested in whats going on here because my car is nearly ready for mapping and with an expected 600ish bhp i need to get this bit right...i was under the impression that the evo intercooler was more efficient than the pro alloy and that it would be good for the 600 or so as i will not be running hi boost..i doubt it will be as high as 1.8 more like 1.6/1.7...is this wrong thinking?
i thought lieghton was using the evo at 500 without issue?

i also would appreciate any help on this

end tank design - part it it though not the core question I started though some may find it useful, not surprising the result, discusses box type sheetmetal (and evo style) fabricated end tanks vs cast profiled types

Performed by an engineer using an industurial blower

In summary

From my testing I have found that at nonpositive boost pressures (atmospheric pressure) that there is about a 23% increase in flow between a typical box fabricated end tank, and the cast ones
presented

http://www.dvdtfab.com/intercoolertestlab.pdf

Leighton would of course know better than I do but from Evolutionm forum I have read many comments that 400bhp is the point where the pressure drop gets too great on the OE evo 4-6 intercooler due to too much restriction, Corky Bell and Graham Bell would suggest the same in their books thats why I started this thread.

Based on what I read over there and in the books 600bhp would really need something different (either longer or thicker end tanks)

Hopefully leighton will comment, but i am not in this power league so would suggest checking out evolutionM or similar

its an evo 8 intercooler i have if that makes any difference ? i think it was an fq one or some such
if i need bigger any suggestions on a good one ?

For 600 bhp that size of intercooler would hinder the intake flow, maybe it'll cool it fine but you'll have to wait longer for it to reach the engine as it all cues up waiting to get through. Mr Bell would have you believe an end tank 35" long x 3" wide would be more suitable. For a 600 bhp engine it's easy to build in lag, the turbo is an easy scape goat.

thanks guys any suggestions on a cooler then? i have looked on ebay and am not sure what shape or type i need to buy

when you say 35" end tanks do you yourself think this is the way to go? as i know you are very knowledgeable...your help and suggestions would be much appreciated..

and a happy new year! all the best to you and yours for the coming year

i take it also that a pro alloy then would not be much better?
sorry if i seem to be hijacking your thread saint it just happened to be posted at the time i was contemplating this anyway..hope you don't mind as there is not much i can contribute to it i'm afraid

To get one that size your faced with buying cores and fabricating. Nearest off the shelf is the Treadstone in Pauls link above. Mr Bell would have you believe its 500 bhp rated, maybe with good ducting it'll be ok for you.
I do believe you need that big a charge area.

Yes indeed, a Happy New Year to you all.

Saint the treadstone one is pretty much like the one I experimented with years ago,
[IMG]http://img.photobucket.com/albums/v74/sediciRich/carpics131.jpg[/IMG]
although like AG Bell said I had an internal baffle on the inside to split the flow and encourage use of the first 1/3 of the cooler closest to a entrance. Bing made from OE cores it stalled the cooling face air massively, if it had an air tight duct of smaller entrance apature then it could potentially have significant heat rejection ability. On my home made mock up it stalled the incoming air so much that any string tassles I put on the rear failed to more at all. When a duct was made it improved significantly, and in a moving vehicle it would be better still. It did highlight the effect of fin density on the cooling face in relation to through flow. A lower fin count would flow more cooling face air, no doubt in the NACA doc it must relate the balance between fin density and heat reject versus change in air velocity across the core and heat rejection. Clearly if the cooling air does not penetrate the core then it will not reject heat from the charge in the core, or if the core has such a low fin count that flow is very high then rejection would be low due to little surface area for heat exchange. The OE core might be small but it is well ducted and I would estimate pretty good for its size espeically with moulded end tanks - for the standard vehicle it was well designed, obviously most cars are not far from that spec.

A number of know (esp in this thread I'm sure) that similar things apply to radiators, the extra cooling abiltiy of an Alloy (aluminium) radiator over an identical size Copper rad comes from the ability to construc the fins differently or rather to a different size an pitch. despite Al's lower thermal conductivity its greater strength allows larger fin gaps (lower density) meaing air flow through the radiator is increased and a greater rejection ability whilst not creating an item that collapses when it is taken out of the box.

i have mailed the us company but if anyone spots a potential candidate in uk could you please let me know..
do they have a particular name so i can do a search for the design
and thanks again

Thanks All

Rich, I plan to change the end tank design on the RX7 fabricated IC I have in the pic to 3inch inlet and outlet but with half the pipe feeding the first half of the IC (first RX7 core) and the second half the second core as it runs on the angle with an internal baffle.

I am a little concerned with through flow but as the evo IC is also OE hope it isn't any worse - something to test as I don't want to cause overheating issues at track days etc.

this article is interesing also

http://www.are.com.au/feat/techtalk/intecoolersMR.htm#Bar%20and%20Plate%20or%20Tube%20and%20Fin

But especially this info on the siphon effect of end tank where inlet and outlet are above each other rather than inlet and outlet at opposite ends



from the article above "In our initial testing, 80% of charge air used end 25% of core at 264 cfm @ 25" water. When we doubled the length of the tank, 55% of the intake charge flowed through the end 25% of tubes & overall flow dropped 9%, indicating that this design has too have very long tanks, but should be able to be avoided anyway. We also discovered flow 'syphoning' for the first time during this testing. The intercooler was making a strange noise, & with the aid of the second manometer probe, & then removing the tank & using the old tissue streamer on a stick trick, found that air was actually being sucked back up the tubes next too the inlet end, by the same principal as a spray gun operation. This means that some of the already cooled intake charge was being recirculated through the core again. This creates a double loss as not only does the extra syphoned air create more pressure drop, it has already been cooled so takes away core space needed by the hot air. I didn't waste time measuring the % recirculation as we rarely use this design. Please note that this paragraph refers to intercoolers with the inlet/outlet above each other (on the same side). It is reduced noticeably if one tank is reversed"


This is why I am going to try and make the RX7 one work, as even though it adds approx 25-50cm of intercooling tubing by going both side of the raditor like the evo rather than only the gearbox end, I hope the better design will be of benefit over an slightly increase in IC tubing

that is an interesting effect I never considered but perfectly plausible.

The most basic equation I remember from Heat exchange design school was

H=UA(t1-t2)

OR heat transfer = heat transfer co efficient x area x temperature difference.

Thats pretty self explanatry except for the co efficient which is very complex. The main thing to remember is there are trade offs with design.

Its all way too long winded to explain this stuff on a forum, but if you're really interested I can explan the basics next time I see you. I think I even have my old text books on the subject.