Turbo aficionado needed (now 'turbo lexus build thread' updated 11/8)

[exdeath]

Hukt on Fizicks

[phucheneh]

jlee and Sunny - I was just saying that 30% would be a maximal number, correct me if I'm wrong, but that seems to just be based on static numbers. If faster air speed is induced, the amount of flow can't be figured on cross-sectional area alone, right?

And I think you know what I meant when I said that smaller motors ran smaller pipe without issue. You can't just tell the dude with the 300hp Honda, 'hey, you can double your airflow by going from 2" pipe to 3" pipe!'

Double the potential, yeah. The question is...well, JCH seems to be addressing that.

And when I manage to wrap my head around it, I'll reply to him, too.

[phucheneh]

New protip: Greddy everything is shit. Why the fuck do people pay 'good performance part' prices on what is essentially APC ricer garbage? Jesus these idiots fail at everything.

Test drove car for the first time since accident. BOV, now that I fixed the 'durr-hurr we can't engineer a paper sack' issue, no longer sticks shut...hangs open. Mostly an idle issue, though; I think it shuts all the way once the vacuum in the manifold is gone. Sorry I made your garbage valve not leak vacuum, you assclowns.

But also, guess what kinda coilovers are on this car? Yup, Greddy.

Complete shit. It's barely been driven, never been lowered very much, and the dampers are still about as effective as screen door cylinders. I've got it handling a LOT better, but that simply came from dropping the thing and putting some preload on the springs so it wouldn't be mushy garbage. Wish I could put the 'adjustable coilover' part on some dampers that weren't engineered by toddlers. Springs suck, too. APC may very well make a better suspension.

...overall, car's doing pretty good, though. Looks humorous with no front bumper and the big ass intercooler hanging out there. Like one of those zombies with its jaw ripped off...but a sweet set of gold teeth.

[JCH13]

EDIT: This guy seems to be assuming that Mach 1 is the same speed in free air as it is in compressed air, which is NOT TRUE. I am going to look into what he's off by when I get a minute... /EDIT

EDIT 2: Taking some WAGs at numbers I calculate the speed of sound in IC piping air to be in the neighborhood of 400m/s, which actually makes his numbers conservative.

My analysis still stands that 2.5in piping will have 2.1x the head loss of 3in piping, but 2.5in piping might not wind up in the danger zone of turbulent flow. /Edit 2

Here is some information looking at a different aspect of intercooler plumbing, the transition from laminar to turbulent flow. Very well done math and well thought-out conclusions.

http://www.dsmtuners.com/forums/152057611-post8.html

Note their example at the end of the post is for a ~450whp car. They suggest keeping the Mach number under .3 to account for imperfections, turns, etc. as these changes can increase flow velocity. This puts 450whp right on the border of 'okay' and 'probably not okay' with 2.5in piping, but well into the 'safe' range for 3in piping.

Being that you think this car is in the 500+whp range, it would land in the 'probably not okay' range of 2.5in piping and still well in the 'safe' range of 3in piping.

If you really want to reduce some of the piping I would suggest leaving all of the 'hot' piping 3in, and reducing the 'cold' piping. The velocities in the 'cold' piping will be lower because the air is cooler, and thus more dense, and thus will flow somewhat slower for the same mass flow rate (lb/min).

Note there are two ways to measure flow: volume [ft^3/min] and mass [lb/min]. The volume flow rate can change as the fluid density changes, but the mass flow rate will remain constant assuming there are no leaks.

[Sunny129]

Originally Posted by phucheneh

jlee and Sunny - I was just saying that 30% would be a maximal number, correct me if I'm wrong, but that seems to just be based on static numbers. If faster air speed is induced, the amount of flow can't be figured on cross-sectional area alone, right?

if all other factors are treated as constants, then what else is there to figure besides the cross-sectional area of the piping?

for instance, the length of piping from point A to point B remains the same regardless of its diameter (i.e. its a constant). if we take an area and multiply it by a length, we get a volume. therefore the volume of air/gas in a particular length L of 2.5" piping is still 30% less than the volume of air/gas in that same length L of 3" piping...or perhaps it'll be easier to see in math terms: V = A x L = 4.90625L for 2.5" piping, or 7.065L for 3" piping. the actual value of L does not matter - the ratio of the inner volumes of 2.5" piping to 3" piping of length L is still ~70% (2.5" piping still flows ~30% less than 3" piping)

now lets suppose the air is being compressed by a supercharger or a turbocharger, and we experiment w/ both 2.5" and 3" piping at identical levels of compression (i.e. compression level is a constant). you'll see that even though a greater mass of air can be contained in a pipe of particular length L and cross-sectional area A at that level of compression than it would at atmospheric pressure, the ratio of the quantity (mass) of compressed air/gas in a length L of 2.5" piping and 3" piping is still ~70% (so again, we see that 2.5" piping still flows ~30% less than 3" piping, under boost or not).

another factor might be the piping material's coefficient of friction, which will affect how well air/gas flows over the inner surface of the piping. but again, we can treat this value as a constant, b/c we wouldn't use one material for the 2.5" piping and another material for the 3" piping in our experiment, would we?

[phucheneh]

I'm curious as to how people are getting the 'stay below this velocity' numbers.

General book-learnin' (okay, TV) has taught me that as air approaches mach (relative to a fixed surface), something happens...what's the technical term here...uh, 'crazy shit.'

That's like mach .8+. But I see numbers ranging anywhere from mach .3 - .6 when people are talking about optimal piping diameters.

With regard to velocity, flow rates, pipe size, ect...the math seems good. But all of the math is being done to make sure you stay under a limit that doesn't seem to have supporting empirical evidence. Which basically means testing in a scientific manner...you just can't judge air turbulence with math. Not without some crazy-detailed computer simulation that models every little hiccup in the piping, at least.

I'm not sure why bends are cited as a source of drag, also. I mean, everything contributes; but a smooth mandrel bend surely does a heck of a lot less to disturb flow than, say, a couple of pipe ribs and a silicone coupler.

I went ahead and swapped in a 2.5" pipe I had to try and solve my clearance issues...still can't clear the strut brace, dammit. I probably just need to make my own bracket for that side. But it's a lot closer and I can put the turbo blanket back on.

So IC > 3" outlet > two 3" 90* bends > reducer > 2.5" with a 75* bend > throttle (actually using a 3.25" coupler for it; 3" was a stretch)

...no difference noted. Maybe when this gets on the dyno I'll do some pipe swaps between pulls and report back.

But my strong suspicion is that I could put a 2" rollercoaster loop on the cold side of the IC and still not make much difference...if something's causing a pressure drop, it's the 2.5" hot side piping; anything after that is going to have a blunted impact.

What I can't figure out is why smaller hot piping is so widespread. I know I've seen factory turbos that actually has the biggest restriction right at the turbo outlet, with the piping increasing in size before it even gets to the IC. That doesn't make any sense to me.

What does make better sense now, however, is fat cold piping. Not by the IC so much as by the throttle...having an 'expansion pipe' of sorts that allows a reduction in velocity (and therefore turbulence?) before hitting the throttle plate and intake runners seems like it would be a positive thing, no?

[phucheneh]

begin picture of interest

end picture of interest

Is there some crazy WRC tech that warrants that fat charge pipe? Or is it just there to modify the way the air is flowing?

Note how it necks back down before the throttle. And has 90* transitions.

Wut?

[NutBucket]

Its a Subaru, different rules apply

[JCH13]

Actually, computer modeling just uses hand calculations to determine when turbulent flow occurs. Computers just do a bunch of calculations at once. If it can be computer modeled, it can be hand-calculated. It would be a good idea to check most of your assumptions at the door.

http://en.wikipedia.org/wiki/Reynolds_number

Look at "Flow in a Pipe" and "Transition and turbulent Flow"

About only part of the Reynolds number that we can nail down with some reasonable certainty is the average velocity and the hydraulic diameter.

The best we can reasonably do is to pick conservative numbers for the viscosity and density, without losing our minds considering every possible scenario.

Of course, working out a few calcs, I can't get a reynolds number anywhere near laminar for IC piping... methinks that someone out there on the internets is confusing turbulent flow with compressible flow. I am a little embarrassed I missed this error completely in the DSM guy's write-up, I should have remembered it from FSAE work

Anyway...

Apparently at mach 0.3 is when air transitions from incompressible flow to compressible flow. Funny things happen in the compressible flow regime that are not intuitive. Things like choked flow and really weird friction effects can happen. I have never done much work with compressible flow so I can't offer much more insight. What I do know is that choked flow sucks, and if you're not careful in your design where choked flow can happen then performance will really suffer.

[phucheneh]

I'm thankful for the links but depressed that I'm about to spend my Friday night reading them...heh.

I didn't know FSAE went that in-depth. I mean, I know; it's a little racecar built by engineers. But the ones I've seen weren't exactly the pinnacle of engineering, especially in their FI setups. But then again, I did go to a college in east Tennessee.

For a year, that is...

I do have a better idea of what you're saying now, as far as this 'critical air velocity' is concerned. Basically, even if your IC piping is made from babies' asses, there is going to be a speed at which the characteristics of the airflow change considerably.

I am dubious about taking car advice from an engineer, though. ;-*

What field are you in, JCH? If you don't mind me asking. [edit: 'engineering' is assumed, I meant in a more specific sense]

[phucheneh]

Uggghhh.

I have concluded that trying to understand fluid dynamics this deeply is painful for me.

It's like, hey, I'm a pretty coordinated guy. I can bang and stomp on things. So I'll just try and play the drums like this guy.

...wait, no, I'ma go kill myself.



Part of it is that I'm prone to, uh, 'epiphanous thinking.' I must understand an entire system on a functional level, and am completely befuddled by any gap in my knowledge. Understanding comes in like a freight train.

So I'm befuddled like a motherfucker here...I'm coming to understand the 'just tell me what to do' mentality present on a lot of 'tuner' forums...

[JCH13]

Heh, yes, we went into serious depth on car analyses. Not every team does, though. Many teams have a much longer history than ours and could do even more detailed refinements.

I am a mechanical engineer by degree and profession.

Here are a few videos to put Formula SAE/Formula Student cars in context (SAE refers to American teams, Student to our European counterparts).

http://www.youtube.com/watch?v=BLsjS5muF3I

http://www.youtube.com/watch?v=O8ti5G4jFSo

Keep an eye on the g-meters in this one, I see lateral Gs peaking over 1.5, and sustained over 1.3, acceleration Gs at close to 1, braking peaking over 1.3.

https://www.youtube.com/watch?v=EMt4QwFNS5s

And yeah... some people make their careers out of analyzing intake flow, not surprising that most people just want an answer!