[s-cars] Re: RS2 TURBOCHARGER CALCULATIONS - LONG

[QSHIPQ at aol.com]

A call to give away the store?
Minhea, comments inserted
SJ 

In a message dated 10/15/2004 5:39:38 PM Central Daylight Time, 
mihnea.cotet at easynet.be writes:
MC>>Scott, in my first post, I was talking about 28 psi (3.0 PR) at 6000 RPM 
>>and 25 psi (2.7PR) at the redline (7000 RPM). Plot the engine flow values 
>>on the compressor map assuming 85% VE for 6k RPM and 80% VE for 7k RPM (I 
>>quite frankly don't think a standard 20vt has more efficiency than that, 
>>even the RS2) and see the results for yourself.

Ok, I'm game.  I do share your VE opinions Minhea
For 6000rpm
135.8ci * 6000rpm * .5 * .85VE/1728
200cfm with no turbo (PR = 1.0)
Assuming a Density Ratio of 1 (DR = 1) and no restrictions elsewhere, you 
multiply 
200cfm * 2bar Pressure (29psi or 3.0PR)
400cfm

Now, over to the RS2 MAP (2672 cold side)
.20 m^3 (* 35.315 * 60) = 423 cfm
.22 m^3 = 466cfm
.25 m^3 = 529cfm (the absolule peak of the RS2 at 70% Turbo Efficiency)

Following the PR line to 3.0 I see we are right at the edge of 72% TE
with an output of ~440cfm and a turbo speed of ~140krpm (150krpm max).  Looks 
ok to me, I doubt one is going to see this as time on this boost increases 
(read: heat soak), but the turbo and the engine seem well matched so far.  In 
fact, I'd even say the wastegate is still opening very slightly...  and you are 
about to shift anyhow :).

For 7000rpm
135.8 * 7000 * .80/1728
220cfm * 1.7bar Pressure (25psi or 2.7PR) = 374cfm (DR  = 1)
On the RS2 turbo map that's 73% TE and ~440cfm and a more leisurely 130krpm 
turbo speed.  I'd say there is good reason why a bunch of folks run the 26psi 
software with the RS2 goodies.  

>> If I remember my 
>>calculations correctly, both of them will either be outside the compressor 
>>map or at very critical points in that map. And again, I'm talking about 
>>the 2672 cold side, not the 2470 cold side used on the standard 20vt K24 
>>turbocharger.

See above


SJ >Pipe diameter could be argued as a contributing factor, but the design of 
>the RS2 may offset that somewhat with a revised (less restrictive) wheel 
>design.  Ultimately pipe diameter is a mathmatical reality, but a large 
>factor is the spinning turbine you design into it.

MC>>The RS2 turbine wheel is a very good one, but it's not ideal IMHO. 
Perfect 
>>for anything below 380 real crank HP, but no good for anything above 
>>that... This is based off of first hand experience, not hearsay or other 
>>people's experience BTW.

My first hand experience shows that it can achieve 300rwhp with the stock IC 
and RS2 bits.  My thinking is that the turbo isn't the limiting factor above 
that. 

SJ>All that said, the RS2 is a relatively small turbine (for high HP AND high 
>torque), so heat soak will eventually cause either CAT boost reduction, or 
>knock boost reduction.  I suspect most that see a reduction in boost based 
>on CAT, since knock interventon, IMO, tends to be a bit more dramatic a 
change.

MC>>First off, there's no such thing as knock boost reduction on an I5 20vt 
and 
>>it's associated M2.3 (3B) and M2.3.2 Motronic system. There's only intake 
>>temp boost reduction and timing reduction, but absolutely NO knock boost 
>>correction. Knock boost correction was first introduced on the M3.8 system 
>>used on the very first 1.8Ts, as well as M5.9 used on pre DBW 1.8Ts, then 
>>got to what I consider as a real state of the art perfection on ME7 drive 
>>by wire system used in all current 1.8Ts and 2.7Ts, as well as any other 
>>Audi gasoline engine using drive by wire. M2.3.x has never used the knock 
>>sensors to correct boost, nor is there a way to use them in order to do so. 
>>Even when an engine is pinging like a churchbell (meaning audibly), and the 
>>ECU is trying to do its best to retard the timing, boost is never dropped a 
>>single psi compared to what it is when no pinging occurs. I am 100% sure of 
>>these statements as I've been able to verify them by myself intentionally, 
>>on my own engine.

Interesting Minhea, I'll quote from Audis own Documentation on the 20vt 
motronic: Knock Sensors I & II, G61 and G66...  "The ignition timing of an 
individual cylinder can be retarded if knock is detected.  If knock continues, a riche
r fuel mixture will be used and the charge pressure is reduced.  Knock control 
is now adaptive which means that the system will adapt to fuel quality and 
driving conditions."

Not sure if this calls into question your BTDT, or if boost modifications 
might adversely affect this feature,  but one can easily argue the way the ECU 
was designed, a knock intevention would cause a reduction in boost.  Why?  
Because VE drops dramatically with knock.  Less VE, less exhaust energy, less 
exhaust energy, less boost.

SJ>Exhaust manifold pressures are usually a function of the design of the 
>Nozzles and throat ratios, not so much the backpressure of the 
>turbocharger hot side.  I'd certainly propose that if one is using the RS2 
>exhaust manifold, the RS2 turbocharger unit probably isn't the 
>restriction...  Yet.

MC>>Well, no, the RS2 hot side isn't the biggest restriction, but a slightly 
>>larger hotside will still allow one to get a few more HP from the excellent 
>>2672 RS2 cold side. I know how to monitor knock sensor activity on an 
>>M2.3.x ECU and how to figure out when it's starting to retard timing only a 
>>couple of degrees on one or 2 cylinders at a time, and I've been able to do 
>>extensive tests with a full RS2 spec turbo, as well as a straight K26#6 hot 
>>side-equipped RS2 cold side. Running the same boost levels (except at 
>>pre-3500 RPM where the straight K26#6 hot side doesn't really shine) on the 
>>same car, there's *up to* 5 degrees more timing that can be programmed 
>>before the ECU starts pulling back the advance across the whole rev range, 
>>with the same or even slightly leaner mixtures on the K26 hot side turbo. 
>>This means about 5-10 crank HP above what an RS2 turbo can deliver with no 
>>pinging and any action from the ECU to retard the timing, again, even if 
>>only on one cylinder.

The world of tradeoffs, but I might argue my point better with some actual PD 
measures Minhea.  Let's compare the exducer ("pipe diameters") of 3 turbos, 
the k24, the k26 #6 and the k26RS2.  The k26 RS2 PD sits between the k24 and 
the k26#6.  Using the PD vs HP output, here is a *massive* differential in flow 
and HP between the k24 and the RS2 with a very slight increase in PD.  For the 
*same* slight increase in PD going from the k26 RS2  to the #6 26, there is 
only a slight increase in flow and HP.  I make the claim, it's not the PD in 
the RS2 it's the design of the hot side wheel.

What are the tradeoffs?  Very little.  In fact your "*up to 5" degrees more 
timing" could mean exactly no noticeable HP gain (or dare I say less based on 
mods), and could be just 'none' with more "n" testing of production 
RS2/7200comp turbos.  If we strictly use the PD argument, the RS2 defies logical 
thinking:  It has better lowend than either the k26 #6 or the k24 (at 2250 rpm on the 
RS2 car) That's a good turbo with few compromises, even reading your results.  
I might respond that the 2672 cold side is ok, but there are a couple that 
have better output that can be bolted to the hot side.  My claim, keep the hot 
side of the RS2 until you reach a point where it's no longer capable of 
matching the needs of the motor or the driving style.  

>However, drop an RS2 turbocharger onto a audisport manifold or a tubular 
>manifold, you could easily argue the point.

>>Hehe, yup, totally agreed, but a better manifold will also free up a couple 
>>more HP from the same turbo.

It could give "up to" more HP, it could give less. Define "better" first...

For the 2226 I5, I kinda take a likin' to that RS2 turbo...  

Scott Justusson
QSHIPQ Performance Tuning
Chicago
'91 v8 
'84 URQRS2 Project