Flowtesting/porting I5 turbo exhaust manifolds...

[OorQue@aol.com]

On Saturday, I took the friend-of-a-friend up on his offer to try out his 
home-built flow bench.  He built it based upon the instructions given in John 
Dalton's book, Practical Gas Flow (which I had pointed out to him, hence the 
reason for his offer), and claims to have had pretty good results with it so 
far.

Unfortunately, as is often the case with most low-cost alternatives to 
high-cost devices, it has a few shortcomings, the most important of which is 
that it's *not* calibrated in the absolute sense.  As such, it's only useful 
for making relative flow comparisons and not determining absolute CFM.  In my 
case, I'm not sure this is all that important as I'm not planning on porting 
heads professionally and don't intend to model the I5 engine on a computer.  
It's also a lot cheaper to rent than the Superflow 110 that is available to 
me -- a six-pack of MGD and lunch at Burger King v. $35 per hour -- and since 
I'm on a fairly limited budget, this is a not-insignificant point in its 
favor.

Because this was just a trial run, I brought only two manifolds with me to 
test: 1) Audi's two-piece upgrade and 2) the Extrude-Honed one-piece manifold 
Ben Howell used on his Ur-Q.  For simplicity's sake, we didn't attach them to 
an actual cylinder head but ran them open to the atmosphere ... once I've 
finished porting and prepping a head, we'll test them again to see what 
differences, if any, there may be.

Anyway, the bottom line is that the Extrude-Honed one-piece manifold 
consistently outflowed the stock two-piece manifold by 4-5 percent.  I was 
surprised by this since I'd heard such great things about how much better the 
two-piece manifold flowed compared to the one-piece -- unfortunately, I 
didn't bring an unmodified one along with me -- and I'm guessing this speaks 
well for the Extrude-Hone process.  I tried measuring flow velocity at the 
turbo flange -- Yes, it was a pain to reach in there! -- and so far as I 
could tell, it was about the same for both manifolds.

I also tried to see how equally each of the three separate passages flowed by 
blocking off the intakes with a plate ... needless to say, they weren't very 
equal with the runner for the cylinder No. 1 flowing the most and the runners 
for cylinders Nos. 2 and 5 flowing the least.  On the two-piece manifold, I 
was able to even things up a bit with my Dremel and after about two hours of 
grinding here and there, I got things to the point that it was outflowing the 
one-piece Extrude-Honed manifold by 7-8 percent ... clearly, there *are* flow 
gains to be had but they're fairly small as these things go -- based upon 
knowledge from other sources, we're probably talking about an improvement of 
less than 10 cfm in total -- and whether it's enough to keep up with a 
freer-flowing exhaust port in the cylinder head has yet to be determined.  
It'll be interesting to see how well the entire system flows with everything 
bolted together.

All in all, it was an interesting experience.  I'm debating whether it's 
worth the money to rent the Superflow and put proper numbers to everything or 
to save a couple hundred bucks and use the home-built bench instead ... 
perhaps the best thing would be to compare the results from each flowbench 
and see how similar they are before deciding.  

BTW, driving home, I started wondering whether a using a different exhaust 
system before the turbo -- I'm thinking of something along the lines that 
Porsche used on its 944 turbo but starting with a non-turbo Audi manifold -- 
may be the best way to go ... certainly, putting the turbo on the other side 
of the motor will help cool the intake air and cause it to run cooler because 
of the air being blown across it from the radiator fan.  I don't how how easy 
it would be to run the downpipe underneath the motor/tranny, though ... I 
guess I'll need to investigate this further.

JG   (NOTE: I'm losing access to this account tonight so it'll be a week or 
two before you hear from me again.)