INFMS

I am an ex- manthey engineer, we were building 3.9's back in 2001, the 4.1l was built at the start of 2010, 4.3's have been around all this year 4.0l racing engines were built in 2007, porsche 4.0l racing engines were built at the start of 2008, the first one by the #1 Manthey engine builder and me....
and yes what I tell you about the GT2RS is true, what porsche wanted to know was how we could get such big rear tyres on the car, not about the power, but they copied it anyway, olaf would get Marcel Tiemann to test the car,
Olaf didn't want the most powerful, he did that with 996's early on and found that it was too much work for what its worth, this is why he capped the M600 at 625hp

INFMS

I do know that the M600 is all that, ask Marcel Tiemann, he got out of it shaking!

avader906

hi - can you please confirm who designed and manufactured cores (not the end tanks) ? last time i was there we joked THEY WERE the 997.2/gt2rs cores.

avader906

Hi Marcus - great post. It is all about compromises. But I would like to note that once you start to move to race durability and performance, numbers change fast - be it suspension, bodywork etc. You cant compare the cost of race parts vs. street parts. Hence "you get wat you pay for".

I had 997.2 coolers in my car for the street use - and if I am honest - for over 20degC climate and agressive / long distance driving they are almost but not sufficient enough (with modded VTGs). Now that that car is undergowing conversion for track only I am moving to secans without further questions.

15k$ on a street only car would buy you a lot of goodies - but you wont have a lot of change if you do buy the best (branded) exhaust out there (cats + ti construction). Similarly, a custom inconel complete race system would match that budget but likely wont last full season.

In (porsche) racing world 15k$ wont buy you anything pertaining to custom design race parts. If you run an entry level P-car series - even something like Porsche Cup, it would be couple races budget.

I do think they (secans) are inexpensive in racing terms, even taking into account that a shunt could render them useless. Surprisingly i've spend 15k$ just on few kevlar/carbon fiber molds.

TB993tt

I'm referring to the engine. I'm sure the chassis set up is awesome and a hot lap would leave anyone shaking but I am focusing solely on the topic of this thread (intercoolers) and what they can do for the VTG 997GT2 engine.

If you read around you will see that I have commented that the limit of VTGs is around 700PS (engine dyno) with 920NM from 3000rpm to 5000rpm..... I can assure you (from reading the acceleration tests of the M600) that this provides a lot more acceleration and VTG fun than the M600 can muster but requires expensive race quality intercoolers (see avader's comments) and I do not believe the M600 intercoolers are of this quality but would love to see some real data as is the way of this thread

avader906

Most of the intercooler "marketing" material I've seen around involves around statements (myths to put it unnecessarily harshly) which I (to my best knowledge) would try to clarify.



1) Our cores are bar & plate - implying it is the best design around and everything else is inferior
2) Our cores have least amount of pressure drop - implying that it is the only desired property
3) Our cores have 30% better flow - implying, again, that it makes these particular intercoolers most efficient and of the best design
4) Our cores are designed for 1,200hp engines - implying that all things equal, that design would equally benefit 600hp version of the same engine
5) Our end tanks are metal (and not plastic) and polished - implying again that those 2 properties alone make the intercooler superior
6) Our chassis dyno shows 40hp gain across RPM range - implying that a)chassis dyno is somehow important and b)that intercooler alone "makes" additional power.
7) Our product is tested to 100psi - implying somehow that this is a necessary property for 30psi flow end-user application.
8) Our cores have double the core density - implying linear relationship to resulting increase in power

First of all you need to consider air-to-air or air-to-water design. Both have different applications - below I will consider air-to-air, more in line with street and endurance racing applications.

As mentioned in another thread, proprietary intercooler product involves choosing the correct core for the application (if not custom manufacturing one); calculating flow, packaging and aero requirements, designing end-tanks for the application and then manufacturing. Most tuners really will spec/buy cores and "design" end-tanks. So far this requires a CAD program and a bit of basic knowledge - "manufacturing" would require additional welding skills. In terms of engineering - this is what you pay for when you buy them. Hence the need to be transparent about what the product actually is and the amount of bespoke effort that made it into the product.

There are few manufactures around who can supply (to spec) cores for the end-user for a given application. To list a few well-known: Bell/Garrett/Marston/Honeywell-Secan/Denso etc (as well as "cheap cores" from China). There are several different cores designs - each with a long list of +/-, so just choosing the type is already solving for few design constraints (out of 300+ variables in a basic CFD model). These are weight/space (packaging) requirements/strength-durability/importance of external air penetration vs importance of charge air flow disruption. Once the design is chosen, you need to spec the alloy (and fin type / production process) requirements accounting for most of the cores cost. You then need to spec the layout, thickness, tube length and end-tanks.

1) There are few core designs - to list few - plate tube & fin, extruded tube (ex: folded tube core), bar & plate and tube & fin (most common). So is bar&plate necessarily better ? On the 997tt/GT2 platform intercoolers are located on the sides of the vehicle and have separate airflow ducts. So external disturbance is not an issue (it would be if the intercooler was on top of another radiator - or intake). So little static pressure drop (charge air penetration), BTUs dissipation per square centimetre, core strength, weight and packaging are important. If packaging was not an issue (you could adapt ducts to whatever frontal area/size you want to run) you would still consider tube & fin. If you needed most consistent method of heat transfer (least heat soak) with heavy duty use requiring strongest core design you would go for extruded tube. Physically weakest core, least charge air flow disruption but second best in BTUs dissipation would be plate tube & fin. For our particular application plate tube & fin and bar & plate are almost equally attractive without bringing in additional factors - such as cost. You could trade off durability vs weight/size, consistency vs heat dissipation and end up with bar&plate, addressing heat soak with better external aero flow and the inferior charge air flow disruption with thoughtful end-tank design.

2) There are 2 kinds of pressure drop. Static and dynamic. We are interesting in best [dynamic pressure drop - static pressure drop]. To argue about static pressure drop is reverting to the right core / core design argument with other tradeoffs. What is the marginal charge temp rise for 1psi of additional boost in order to overcome for the static pressure drop and how much of it would be taken out by the more efficient core ? The more the charge is cooled the higher the dynamic pressure drop because cooler air is denser and occupies less volume. Holding all things constant once the core design is chosen, the more the surface area, the higher heat transfer but also the greater pressure drop. So strictly speaking dynamic pressure drop is a desired property. You would want the most heat transfer (lowest charge temp) for the given target cfm.

3) What is important is not the absolute flow (cfm). Will get to the surface area of tubes in the next point (more open flow design - higher the flow). But it’s sufficient to say that 20% gain in air flow through a component would not increase the engine power by 20% - it's 5% or less. And obviously removing the intercooler will flow more but is it better ?

4) The artsy bit in the intercooler design is sizing the surface area of tubes where the charge air enters - charge air window. Too small for the cfm flow, the pressure drop increases and air speed through the core will be too high. This will not allow enough time to get the heat transfer out of the charge. If the window is too large (which is obviously the case with 1,200hp designed intercooler used in 600hp application) then the pressure drop will increase for little extra cooling (requiring you to re-design the intercooler with shorter tube length). You need to size the intake area for the cfm requirement of the engine and then the length of the tube for the heat dissipation. The more surface area, the higher heat transfer but also great pressure drop. More surface area = more weight. So such intercooler will be suboptimal in engineering terms.

5) Polished is easiest to address. Smooth surfaces cause capillary tension increasing drag and pressure drop. End tanks warrant a separate thread because the optimum design dictates the airflow shape to the charge air window. Let's say that it is on par in important as choosing the core itself, so not easily overlooked by P/oem supplier. In terms of plastic vs metal - with thin material heat soak is negligible but it's least durable. Thick material will be detrimental and you also need to consider the design of where the end tanks are and heat soak from surrounding elements.

6) Chassis dyno does not allow for the necessary precision in testing incremental design features. Standard variation of runs is simply too large and will make any "testing" of components prohibitively slow and expensive - as you need to account for so many performance impacting variables outside intercoolers which you can not hold constant. More efficient intercooler allows you to re-map to more aggressive timing maps - and would ultimately require a precision retune of existing maps. And once we go into additional ECU tweaks you cant compare as results would vary widely depending on how aggressive remap is. In addition, most of the design decisions are done using rather expensive software and tested on even more expensive flow bench - so discussion of what models vs what tests been performed are more relevant for performance evaluation. For end-product testing, evaluation and tuning engine dyno would be more usefull in terms of running an exact cfm programme, simulating thermal load - as well as calculating Q values from taking direct temp measurements - if dyno is supplemented with special intercooler stand (fluid filled).

7) So thick tubes and/or durable cores are used withstanding higher pressurization than other products. Tradeoffs ? Weight comes up immediately. Are they then for the correct application or over-engineered in durability resulting in higher heat soak, pressure drop and weight ?

8) For the same frontal area, doubling the core density would raise efficiency around 5% but almost double the weight of the intercooler.

Finally, one would have some cash left by buying 997.2/gt2.rs intercoolers over secans. There are some solutions out there for technically inclined to experiment requiring some patience and time. For instance there is heat dissipation coating available to increase the efficiency at expense of maintenance. You could also spray them. You can improve ducting and pressure difference between in and out ports of intercooler ducts. Fully enclosed intercooler with no ambient flow will still offer around 20% efficiency. One could experiment with increasing this differential through better aero design (a la GT2 RS rear bumper) or even mounting ventilators (then you would need to calculate the optimum pressurized flow and find a middle ground). Better heat shielding, better shape of intake ports should also be looked at. I honestly do think that 997.2/gt2.rs cores are (almost) sufficient up to for the 600-620 PS street applications (as have the best performace vs cost) and residual cash could be spent to improve mentioned areas. As for motorsport applications, same would be done but using secan cores as starting point.

INFMS

I don't know who made the cores, Olaf had his contacts, these were made in 2007-8 well before the GT2RS. And if you want the info on them ask Olaf for it (good luck...lol), he is someone who does not give out info easily, when I started there I had signed confidentiality papers in regards to engine performance and set up

avader906

Correct i thought you would be inclined to speculate.... i wont comment on general availability of technical info out of any racing team and/or engine shop.... hence this thread....

INFMS

One of the things we did try was what you suggested avader906 and that was to get more air through the intercoolers, 1st thing we tried was changing the inlet duct design on the outer edge, giving it a more of a scoop like duct this did not work, we know that the exit needed to be more efficient. I had come up with another solution which was never used on that car. Olaf never knew what i had come up with.

avader906

Maybe to list few potential improvements that are not really relevant for street use.

For instance I am looking at wheel liners and underbody/diffuser (extra channels) to take into account the pressure differential for the intercooler ducts. Ducts themselves need to be resized and redesigned - and I agree with INFMS that it's the outlay duct that needs it most. If you can get center muffler-free exhaust (by itself a major weight and heat component), it opens a lot of possibilities - including (as mentioned) fans and better airflow through the engine bay itself.

Another basic improvements would be the cooling of turbocharges (raising their efficiency) (a lot of things could be done here including basics such as oil lines and better heat shields), eliminate few bends with one-piece airpipe (ie. getting rid of boost hoses - having one cf piece from compressor outlet to throttle body).

If you have the luxery of time and money the list is rather long....Gotta run

Nordschleife

I have a simple test for intercooler adequacy.

Mid/late Summer, leave Stuttgart or München early in the morning (how early - early enough to get through Austria before the cops get out of bed, ie later on a Sunday) and drive south towards Nardo in the heel of Italy. Note the point where the car feels completely clapped out.

Drive back and repeat exercise - if you can make it all the way to Nardo, you might set some speed/enduro records.!

I've never had a turbo car which hasn't had its turbos replaced under warranty (even when modified). Even modern ECUs don't seem to be able to offset the bad effects of a poor fuel load. Normally I'm religious about putting in V-Power. As a result of conversations with a variety of engineers, it appears that the closer the fuel station is to the refinery, the more likely it is that the fuel will be good. For the record there is a big refinery (more than one I think) at Ingolstadt. Much to my surprise, I have had more fuel quality problems in Switzerland than one might reasonably expect.

Finally, Pankl are capable of excellent quality engineering, at a price, well worth investigating for limited runs of kit.

R+C

TB993tt

So what is it ? understand what you say about fuel quality but am struggling to connect it to intercooler adequacy

Nordschleife

If you make it to Nardo, the intercoolers are OK, otherwise note the distance managed - via Florence (Firenze) its just over 1500 km - just compute how far you get before the car feels gutless and use that distance as a relative measure of adequacy. You must press on the whole time (its Italy the police will encourage you) and don't go down the Adriatic coast (not enough hills).


Adequacy
100% - 1520 Km - Nardo
72% - 1100 Km - Naples
63% - 930 Km - Rome

etc

This is a lot more fun and not much more expensive than some of the other alternatives and has the merit of being 'real world'

R+C

911SLOW

Great test; simple, inexpensive, without many variables and safe.

avader906

In your opinion, would a go fast run alone London > Munchen > Davos > Turino > Nice > Barcelona and back count towards that particular test or is essential to limit the run to 1,500 km ? I'm really afraid some doggy mates in Naples would jack the load.