HTA3076 @ 40psi 4th gear on 245 RE01Rs on 17x8.5 wheels


BWEFR @28psi in 4th gear on 295 R888s 18x10 wheels


Do you have a run with the HTA76 that at least runs the same boost pressure as the BW? You said you did a 30psi pull with the EFR. "3012 psi of boost, the car made...........556/499" -Dave B.

https://www.evolutionm.net/forums/9707385-post358.html

The lowest psi pull on the 3076 was this one at 34psi

 

Interesting, the HTA76 has 40whp more at 4100rpm where the EFR is still building boost

 

With all due respect you're just lying to yourself here, you really want to believe that the 7670 is the gamechanger its cracked up to be, just like I did up till a couple of weeks ago until I woke up and ended up cancelling my order, despite being #4 on the waiting list and having placed my order 8 months ago. But lets get back to numbers:

Your calculation above makes it appear that the 7670 turbine is actually THINNER than the GT30, which is clearly absurd... So let us completely remove section thickness from the exponent, and factor it in separately. We end up with
I GT30 / 7670 = (8.54 x 57.44^4t1)/(4.0 x 65.38^4t2)

Take a long objective look at this photo again, with a focus on tip thickness at the inducer and superback sections: http://www.perrinperformance.com/sha...rbinewheel.jpg

Then compare it with this:

http://www.perrinperformance.com/sha...rbinewheel.jpg

Keep in mind that the 2nd photo was done at closer range, making it look much bigger than it really is (compared to the the 7670) , look at the black dot in the top right corner, look at how much bigger it is in the second pic... I can't give comparative thickness estimates of the two turbines above (although I will measure the GT30's section thickness once I get to the shop), I can do the following: If the GT30 and 7670 had exactly the same section thickness, which is CLEARLY not the case, the GT30 would have a 27% greater PMOI than the 7670. And all this STILL doesn't take into account the significant mass of metal on the backdisk that the 7670 has that the GT30 doesn't, right where it does the most harm (PMOI wise), and is easily worth an additional 20-25% fudge factor...

Granted I can't measure the exact PMOIs of both wheels, and I can't measure the section thickness of the 7670 turbine wheel, but then again I don't have to... Jeff Perrin's pictures make it crystal clear to me, beyond any shadow of a doubt, that the 7670 does NOT have a lower PMOI than the GT30 like we thought it would. In the best-case scenario its the same (which I don't believe for one second), but in reality its probably greater... By how much I don't know and don't care. Again, study the photos carefully and objectively, and you should arrive to the same conclusion.

One more time, I'm not passing judgement as to which of the two is "better", I just wanted to debunk the idea that the 7670 has less rotational inertia than the GT30, and with that I won't be commenting any more on this subject.

 

The gtx wheel looks to be taller to me considering its OD compared to the EFR. Were any measurements taken?

 

not too many people on this forum have the experience and knowledge base that you do tim!

DB - i didnt call you a liar, I know you truly believe the posts you write. Of course doing this test on 2 different cars/turbo setups/wg springs is not the most scientific method and a WOT steady state dynopull misses the real benefits this setup delivers. Im not sure what my attending your 2012 shootout event will solve.

the most important data we want from these turbos is to log "snap accel" data during the comparisons on and off the dyno in addition to the dyno vs street change in spool. time TPS RPM boost are the keys here. Basically WOT from 2500, WOT from 3000, WOT from 3500, etc... This can be done either on the road or on some dynos and the measurables are time to torque and time to boost.

It doesn't matter so much at what point you lift... you can run it all the way to redline if desired. torque vs. time and boost vs. time from the dyno would tell the story, the same ECU logs can be collected on the street. Obviously torque is not available in the street dataset (virtual dyno software can help however) but you can get TPS, RPM, and MAP versus time. The easiest way to see whether or not it’s repeatable is to do the same “run” three times and see how good the overlay is.




I expect it would work well, but would be worthwhile to work it out throughthe matchbot. we have had a lot of inquiries lately about that type of setup. based on the EFR 6258 results weve seen on jetskis I expect the bike engine would work well

no, these GTX's are the same GT turbine wheels used for the past 12+ years just like youre old 30R or your hta turbo

 

Hydra you can't get accurate MOI numbers by "looking at things" that aren't simple shapes. Model it, apply an accurate material, and set the axis of rotation. I would say the geometry is complex enough that we cannot use simplified mathematic equations to come up with an even remotely accurate number.

And FYI were not discussing pmoi, since this thing isn't a gyroscope.

Point is you might as well be talking about feelings, because your inputs to your equation are guesses at best.

 

Geoff,
I completely agree on your method, but you're missing a key detail. Before starting the WOT, the turbospeed needs to be at a consistant starting point. If you stabilize your engine speed at the start speed, your turbospeed will be higher and not repeatable. We standardized the process to lift on throttle and let the engine decelerate a few hundred RPMS (remember we were testing 12.9L Diesel engines pulling 80k pounds revving to 2100rpms) maybe for an automobile it would be 2-3k rpms. I know on my Evo and the Supra I tuned we were able to do a 1500rpm drop in 4th gear. To explain it a bit more clearly. If you're going to start your pull at 2500... run the engine up to 4k on the dyno and lift, coast down to 2500, and then go WOT and start your pull.

We would take data from 900rpm to 1700RPM...about every 1000rpms. Like Geoff said we would do 3-4 pulls, throw out any outliers and average the data. I think reasonable RPM breakpoints for an Evo would be 2500, 3000, 3500, 4000, 4500, 5000, 5500, and maybe 6000.

We did the testing based on the time it took to reach 90% of peak torque for that RPM. We used our WOT power/torque curve (same as all of your dyno plots) to calculate what 90% of peak torque was at each RPM. Take the elapsed time from WOT until you reached 90% of peak torque.

And like everything else, your ambient conditions play a large role in these numbers. We did testing at Altitude, cold weather, hot weather, sea level etc.

I know Aaron @ English has used this method and process to quantify turbocharger response. I think it should be understood and implemented if we're going to be evaluating the "performance" of a particular turbocharger or engine combination. We called it T90 times, but whatever...same thing.

 

Can any of you guys with a 7670 measure the section thickness of the blades? I will do the same for the GT30 tomorrow - along with overall height... Meant to do so today but was swamped with work.

 

The engine would definately have to be on a downward movement when the throttle was hit or you wouldn't be testing the transient response. I'm looking to do some testing soon on my 9 engine and am thinking about converting it to drive by wire so that it will not engage the throttle opening until the target rpm is met. It would be easy to allow say only 2-3% throttle opening at full throttle condition so the engine crawls to target rpm then the taps open. If feel that would give a very repeatable series of runs whilst being close to mimicking a fully closed throttle transient condition. It is something i have tested extensively on normally aspirated but never on boosted engines.

Have you tried correcting the time based data for weather conditions? I do a lot of time vs rpm dyno testing but don't have the budget for climate control on either of my dyno cells. I've yet to find anyone else who has admitted to doing it or even thinking about it outside of the very high-end facilities by which i don't mean normal tuning shops.

 

I have a non-EFR S200 here i could measure if i get some time over the weekend. Not sure if that is any use.

 

There's is no point comparing just the wheels. If you're making comparisons you need to compare the entire rotating assembly of each unit.

 

Tim,
We had 1 test cell that was an AVL climate controlled cell with a Horiba AC dyno. All the others were just standard AVL test cells that controlled fuel and air temperature... the inlet pressure and exhaust pressure were adjusted by usign valves in the air plumbing. There were several other variables that were controllable in the test cells, but none of which are pertinent to this discussion. The non controlled cell's ambient air temperature varied somewhat with ambient conditions.

I think your idea that you can sneak up on the target RPM at low throttle may be reasonable and may be repeatable, but it would not be able to be correlated on the road (track) with a test driver while taking data. We were testing with fully instrumented engines, so we could see what the turbospeed was at part throttle and what it was when you were doing a true deceleration. There is quite a bit of a difference, and I think that changes your data quite a bit. FURTHERMORE... doing this at higher engine speeds I think will prove to be very unstable and very hard to repeat. Again, do your testing with both methods and look at your turbospeed... I think you'll find and agree that full decel is much easier to accomplish & repeatable in the test cell and on the test track.

We never took the data and made correction tables for transient response based on altitude, humidity, or ambient temperature, but we did use the data to tune the engine for best response at each condition and fed that back to the test cell to validate emissions and fuel economy impacts. We did testing in Northern Territories, Arizona, Colorado, the South East and everwhere in between. The transient response for these particular vehicles was more dictated by emissions (specifically a low pressure EGT system) & the SFL than the VGT turbochargers.

I learned very quickly that Colorado may be sunny, but the air sucks. Washington is pretty awesome in that respect.

 

AVL . Full AC axle dyno?

I was planning to do the DBW test method on the engine dyno and then later collect data on the road and back engineer and look at trends. Probably the best i can hope for with the time and budget i have available unfortunately. Road data may not be as repeatable as i'd like but there comes a point where budget and accuracy start to part ways at a rapid rate!

 

Yes,
The test cells were the complete AVL package... Kistler sensors...the works. Very nice facility. The Horiba unit was an Dynas3... I don't remember the exact specs, but we could drive the engine with the AC dyno and it would tell us exactly how much power it consumed at each RPM.... I wrote a bunch of dynamic test cycles based on real field test data. Very cool stuff... especially if you were racing on tracks! They never change, so you could put together an amazing engine and validate its improvement before a tire ever touched the track.

 

Per degree pressure analysis, very jealous.