Engine modelling software accuracy

timdog · Post by **timdog** » Wed Jan 04, 2006 2:55 pm

I've been trialling the EAP from performance trends with mixed results.
Using data exported from port flow analyzer and comparing with real graphs off my dyno the results are a little inconsistent.
The applications are all jap motorcycles being run on my bike chassis dyno and flow data is from an SF bench.

When i model the bigger engines (1000cc-1300cc) the results come out really good by and large.
When i go smaller (400cc and 600cc), first thing to go out of the window is the power figures.
It is proving really accurate with power curve shapes and also good at showing cam timing change effects.
Where it goes drastically wrong is on port volumes whereby on ports i have reduced in size substantially and seen large power increases, the software predicts power loss! Power seems to be almost totally related to cfm only.
It also took me some time to get my head around the exhaust specs because initially measured lengths were killing the power thousands of rpm before the limiter.

Has anybody else had such issues, know a way around them or found other software to be more accurate on these smaller engines?
Any feedback gratefully received.

Post by **Dave Koehler** » Wed Jan 04, 2006 3:09 pm

Best thing to do is call Kevin at PT. He may know something you are missing. He also likes folks that have some numbers to work/compare with.

Dave Koehler
www.koehlerinjection.com

timdog · Post by **timdog** » Thu Jan 05, 2006 4:22 am

Ok thanks, i'll drop him a line. Most likely i'm entering data wrongly or something else stupid.

ron hamp cycle · Post by **ron hamp cycle** » Thu Jan 05, 2006 3:41 pm

are the smaller singles 4 valve engines that you are trying to simulate its very acurate on 2 valve engines

timdog · Post by **timdog** » Thu Jan 05, 2006 3:53 pm

Yes, 4-valve heads and revving to over 16,000rpm.
I've emailed Kevin and he's looking at my model files.

ron hamp cycle · Post by **ron hamp cycle** » Thu Jan 05, 2006 9:09 pm

one of the things that i had mentioned to Kevin was there needs to be some kind of a head rating for the angle of the port .a down draft port very straight like a yamaha 600 can suport higher port velocitys and keep filling the cylinder producing more power i know that some of the early kawasakis benifited from having a smaller port than they came with kawasaki supplyed Rob Muzzy with a casting and smaller ports for one of the 750s and it made better power .if you had to ports that were identical in cross section and lenghth with the same size valves and identical flow numbers with the only diference 1 being a down draft port and the other having to bend 70-80 degrees the down draft would fill the cylinder better @ higher engine speeds and make more power because of the reduced velocitys @ the short side radius and less flow seperation .if im correct based on what i said the program would need a revision made for straighter ports alowing higher velocitys before the power started falling off or a revision installed for measured velocitys with a peto tube

craig · Post by **craig** » Thu Jan 05, 2006 9:54 pm

I have not used a late version, but the early stuff failed to model my Suzuki motor with the cams that actually worked best. So much for "filling and emptying" formulas.

They seemed to work OK for cars with push rod 2 valve motors, but I stopped using the stuff after finding that it failed to work well with bike motors.

Keep us posted. I'd be very interested if the stuff has improved.

timdog · Post by **timdog** » Fri Jan 06, 2006 5:29 am

I have had some very accurate results with the software as far a generating the curve with the right shape. Power is fairly close and cam adjustments also show up.
Main problems for my application are:-
1) Port velocity changes not showing.
2) Lack of support for complex intake and exhaust systems - ie. multi tapered headers, 4-2-1, unequal length inlets.
Apart from that its very useful.

Yes the 750 kawasaki had way oversized ports. You could fill the majority of the port to almost half its original size before the cfm curve would even move! Yamaha 600 also had slightly oversized ports until this year. The latest ports are extremely downdraft. The ssr will be gone before long!

SWR · Post by **SWR** » Tue Jan 31, 2006 7:28 pm

Hey Timdog,
I've had the exact same issues with the Pro for several years now,and from errors I found in the Pro just a week ago I might have convinced Kevin to do a full library change on the exhaust menu,if we're lucky (and keep pushing him

) we might get tapers on all pipes,good 4-2-1 and 4-1 header simulations,and stepped collectors (that is,a short,very negative taper collector hooked to a full pipe length) in the new EA Pro V3.5 coming this summer.

You've noticed that for a given valve size,let's go for a bigger one like the Porsches I'm doing of 45mm intake,the port can be anywhere from 43mm min CSA (optimal in the Pro) to 38mm (optimal in GT-Power) and the Pro shows no power difference? Stock port for these Porsches are 41mm,in the Pro going to 43mm increases power,although VERY slightly. In the real world,it loses power. And going to a 38mm port in the Pro doesn't lose power (but it shows no gain either),when the dyno shows a 12-hp increase even with a crappy port shape that loses 3 cfm from stock through the range...

If you're still pushing Kevin to look at it,keep up the good work and tell him you have the same velocity/tuning problems as Bjorn from Norway...he knows me well enough.

Then things might start to happen...

By the way,do not use the "find CFM by throttle size"-menu for the Mikuni or Keihin carbs...it gives you much less CFM than it has. I know,I've flowed most any 36,38, and 41mm CV and flatslide carbs in existence...

2T4T · Post by **2T4T** » Wed Feb 01, 2006 4:37 pm

Hello ,
Timdog , I am using a young but efficient wave simulator : Forte-4T is the enthousiast version from http://www.rtz-soft.com. Forte4T is for single cylinder engine or fully separated multi-cylindres (much more rare - who own a 500 MV Agusta ?
Do you want a try ? I can supply you with a YZ426 data and results from a SAE paper for your simulator to compare results.
I have been the tester for one year and we found a length equivalent to valve lift need to be added to head port length for better result (many SAE paper about wave simulators say 0.5D at beginning of manifolds).

Regards
Alain

timdog · Post by **timdog** » Wed Feb 01, 2006 4:59 pm

Thanks for the feedback guys.

Bjorn,
I haven't had anytime recently to look into the simulation. Should have more time by march so will get onto kevin again then.

Alain,
Yes i would be interested to try, but as stated above don't have much time at the moment. In a month i should have much more time.

Great to see i'm not the only one on this.