Jump to content
SAU Community

Recommended Posts

Hi guys,

I conducted a regression analysis using the output figures and boost ran from the dyno thread relevant to the RB-25 engine with a HKS GT-RS Turbo.

Basically, a regression analysis is a form of statistical modelling that attempts to evaluate the relationship between one variable (Boost - PSI) and another (Power output - RWKW).

This data will help you 'predict' what sort of power you would get if you have an R33 GTS-T (RB-25 engine) with a HKS GT-RS turbo and other supporting mods of course (ie. Front Mount I/C, Full exhaust, Fuel pump etc etc) running the boost of your choice.

Basically, select the amount of boost you wish to run, and look at the corresponding output figure next to it for an estimate of how much power you would be making.

This analysis is ONLY a rough guide, please do no use it as a definite determinant since all cars are individual and differ to one another.

This post should eliminate those repetitive questions such as; 'How much power will I make if I run XX amount of boost with a GT-RS turbo on my R33?'

post-48356-1207094323_thumb.jpg

Cheers!

Edited by Turbz RB-25
GTRS is on my wish list.. when my turbo is dead

So are they bolt on ? as in dump and turbo flange ? or do i need v-clamp dump , and so on...

With stock internals, 14psi is realisitc ?

Its bolt on if you get the T3 flanged one...and if you get a kit it comes with gaskets and a dump pipe too...

Stock internals will take 18psi-20psi easily...as long as support mods and tune are up to scratch...

lol. I did that with Excel to see what min/max/avg gains ppl got with it too. The problem is, a lot of results in the Turbo Upgrade thread have so many different variables (the supporting mods you gloss over), that you can't really draw much more conclusion than you would simply by reading the thread. I think ppl running a fuel or boost controller only (or even in combination if done poorly) are going to get measureably different results to those with full ECUs (and even within the ECUs there would be more variation, possibly less so). Also, it seems a LOT of ppl on that thread post up their results with a little note at the bottom saying 'injectors were maxing out, will go back to the tuner next week' or 'AFM needs upgrading next month'... so that also throws more unknowns into the mix. Having said that, the range between min and max in your results, look about the same as I got. I'll add another result in the upgrade thread in the next week or so :D but I won't have the little addendum at the end about something max out.

lol. I did that with Excel to see what min/max/avg gains ppl got with it too. The problem is, a lot of results in the Turbo Upgrade thread have so many different variables (the supporting mods you gloss over), that you can't really draw much more conclusion than you would simply by reading the thread. I think ppl running a fuel or boost controller only (or even in combination if done poorly) are going to get measureably different results to those with full ECUs (and even within the ECUs there would be more variation, possibly less so). Also, it seems a LOT of ppl on that thread post up their results with a little note at the bottom saying 'injectors were maxing out, will go back to the tuner next week' or 'AFM needs upgrading next month'... so that also throws more unknowns into the mix. Having said that, the range between min and max in your results, look about the same as I got. I'll add another result in the upgrade thread in the next week or so :D but I won't have the little addendum at the end about something max out.

Sl33py,

Regression is some what different from averages, min/max values.

By calculating max/min and average values, you cannot determine anything above or below that, only average boost and average power.

In the simplest form, x (boost - independant variable) vs y (power - dependant variable) axis are plotted. From there, excel estimates a linear (best fit) curve that is best suited to all the inputed data.

Thus, the equation is a simple linear one (y=mx+b), m being the gradient, x being the amount of boost, and b being the y-intercept.

If anyone wants the 'Anova' print out from excel (the regression analysis) send me a PM and I will email it to you.

The JPG picture is merely me filling in the 'x' (boost) value in the equation above and 'y' (power) value will be calculated.

Cheers

Here's the formula if you want to calculate it yourself if you want to run above 20 PSI or below 14 PSI.

Power (RWKW) = [6.83 x (Boost in PSI)] + 137.09

So for example, if I want to estimate my power running 12 PSI, I do the following:

Power = (6.83 x 12) + 137.09

Power = 219 rwkw

Remember, this formula is only applicable to an RB-25 Engine (eg R33) with a HKS GT-RS turbo along with your normal supporting mods (exhaust, front mount intercooler, fuel pump, computer etc)

Yes, I studied statistics in uni. It's more the point that the supporting mods are not even remotely close to constant between any of the data you used for the regression, rendering the regression pretty airy-fairy. Simply this: 'Remember, this formula is only applicable...' demonstrates as much because that statement covers a huuuge range of variables. Different ECUs (or fuel/boost controllers), intercoolers (and various setups), AFM used (and inherent limit), injectors and their limit, pump and it's limit, air intake etc etc, not to mention the well known variances on the dyno depending on how the car's loaded up, temperature, bonnet up/down, shoot-out mode or not..... Statistics is all about changing one variable and keeping the rest the same and measuring the correlation, but that's impossible with the (comparatively) small and inconsistent (the supporting mods etc) data set on that thread.

But I can appreciate the analytical thinking put in, more power to ya.

Yes, I studied statistics in uni. It's more the point that the supporting mods are not even remotely close to constant between any of the data you used for the regression, rendering the regression pretty airy-fairy. Simply this: 'Remember, this formula is only applicable...' demonstrates as much because that statement covers a huuuge range of variables. Different ECUs (or fuel/boost controllers), intercoolers (and various setups), AFM used (and inherent limit), injectors and their limit, pump and it's limit, air intake etc etc, not to mention the well known variances on the dyno depending on how the car's loaded up, temperature, bonnet up/down, shoot-out mode or not..... Statistics is all about changing one variable and keeping the rest the same and measuring the correlation, but that's impossible with the (comparatively) small and inconsistent (the supporting mods etc) data set on that thread.

But I can appreciate the analytical thinking put in, more power to ya.

Yeah, that's why I wrote in the first post not to use the table as a definate determinant since indavidual cars are different (i.e different intercooler, different computer, different fuel pump etc etc)

Nonetheless, you are still able to use it as a rough guide as to how much power it'll put out.

Suprisingly, I performed the same analysis on the HKS GT2835 Pro S and the power turned out to be less than the GT-RS at the same PSI.... Strange........ You'd think it would be more??

Anyways, enjoy!

Cheers

Yeah, that's why I wrote in the first post not to use the table as a definate determinant since indavidual cars are different (i.e different intercooler, different computer, different fuel pump etc etc)

Nonetheless, you are still able to use it as a rough guide as to how much power it'll put out.

Suprisingly, I performed the same analysis on the HKS GT2835 Pro S and the power turned out to be less than the GT-RS at the same PSI.... Strange........ You'd think it would be more??

Anyways, enjoy!

Cheers

Yeah but the 2835 is a larger turbo... Larger turbo's usually allow for a cooler charge at a specific PSI/Kpa than a smaller turbo (this of course, would vary from compressors, housings ect ect).. So what one would usually do is ramp up the ignition/cam timing to make up for the shortfall... A larger turbo is made for more upper mid-range/top end grunt, hence although a smaller turbo could possibly make mroe power at lower pressures, obviously, as pressure/rpm climbs, the smaller turbo runs outta puff whilst the larger one just keeps huffin' and chuffin'..

Plz correct me if I'm wrong, but I recall reading somehting in either Zoom or HPI when they tested two garrett turbo's on an LS1 powered VX commie, and the results were suprising...

Edited by Marco-R34GTT
Yeah but the 2835 is a larger turbo... Larger turbo's usually allow for a cooler charge at a specific PSI/Kpa than a smaller turbo (this of course, would vary from compressors, housings ect ect).. So what one would usually do is ramp up the ignition/cam timing to make up for the shortfall... A larger turbo is made for more upper mid-range/top end grunt, hence although a smaller turbo could possibly make mroe power at lower pressures, obviously, as pressure/rpm climbs, the smaller turbo runs outta puff whilst the larger one just keeps huffin' and chuffin'..

Plz correct me if I'm wrong, but I recall reading somehting in either Zoom or HPI when they tested two garrett turbo's on an LS1 powered VX commie, and the results were suprising...

I don't really understand what you're trying to say?

What I meant was, for example at 18 PSI, the car with the GT-RS Turbo put out more power than the car with a GT2835 Pro S Turbo (according to my regression analysis).

Now, I'm pretty cut as I recently purchased a GT2835 Pro S.

So are you trying to say that if you wound the boost up to like 22 PSI that the GT2835 would overtake the GT-RS in terms of power output, but at low boost levels, such as 14 PSI that the GT-RS would be better suited?

Cheers

Basically yes..

I gotta dig that magazine and find out the nitty gritty before I start misleading people..

Wouldn't the GT2835 still produce more power than the GT-RS regardless of the PSI run (providing that both turbos run equal PSI) merely because it has a larger compresser wheel??

By that I mean, if both the GT-RS and the GT2835 pro s ran 12 PSI, the GT2835 would produce more power as it has a bigger compresser wheel, thus, pushing more air?

GTRS is on my wish list.. when my turbo is dead

So are they bolt on ? as in dump and turbo flange ? or do i need v-clamp dump , and so on...

With stock internals, 14psi is realisitc ?

Its not a straight bolt-on unless you get the whole kit, i think they go for like 2600 or 2700 now, Stock internal

will handle up to 19, 20psi no problem. You basic need a good set of injectors and other supporting mods.

Now, I'm pretty cut as I recently purchased a GT2835 Pro S

Don't be! This is precisely why the graph doesn't tell you very much. Have you ever looked at those turbo application sheets? They are a lot more complicated than the ol' y = mx + b linear equation.

I chose a GTRS over the 2835 because I PMed about 10 guys I saw that owned them and out of the discussion came my decision that a GTRS would have more power that I would actually use on the street at a sane boost level, and also on the track, whereas the 2835 would have more power up top (where I don't necessarily spend all my time). That, and I hate lag.

Someone correct me if I'm wrong, but using the arbitrary hypothetical of running either of these turbos at 12 PSI is pretty pointless because usually they are run both between 14-18PSI. To talk about numbers outside the 'common' range is... well silly. Trying to analyse turbo performance like you have is just likely to mislead you and possibly cause aggro when it isn't warranted. Who cares if a GTRS would make (some, not a lot) more power than a 2835 lower down? You should have more up top and if you don't, then there's a problem to address.

RE the graph you just pasted. I think it's because you have to wait until more air is being moved by the engine before the larger turbine wheel of the 2835 starts spinning, whereas the GTRS has less to spin up (less lag). Noone will agree that a GTRS would make more power than a 2835 when both are run at or around 18PSI. Everyone knows the 2835 gives more up top (maybe 30rwkw or so? I dunno). You're analysis is too simple in some aspect or another. I would say maybe the data set, or some outriders are screwing up ur numbers...

Edited by sl33py

Hi guys,

Did the analysis for the GCG High-Flow turbo.

Doesn't even compare to the GT-RS and the HKS GT2835 Pro S in terms of output.

Nonetheless, for the money, still produces VERY good results.

post-48356-1207175483_thumb.jpg

Cheers lads!

Don't be! This is precisely why the graph doesn't tell you very much. Have you ever looked at those turbo application sheets? They are a lot more complicated than the ol' y = mx + b linear equation.

I chose a GTRS over the 2835 because I PMed about 10 guys I saw that owned them and out of the discussion came my decision that a GTRS would have more power that I would actually use on the street at a sane boost level, and also on the track, whereas the 2835 would have more power up top (where I don't necessarily spend all my time). That, and I hate lag.

Someone correct me if I'm wrong, but using the arbitrary hypothetical of running either of these turbos at 12 PSI is pretty pointless because usually they are run both between 14-18PSI. To talk about numbers outside the 'common' range is... well silly. Trying to analyse turbo performance like you have is just likely to mislead you and possibly cause aggro when it isn't warranted. Who cares if a GTRS would make (some, not a lot) more power than a 2835 lower down? You should have more up top and if you don't, then there's a problem to address.

RE the graph you just pasted. I think it's because you have to wait until more air is being moved by the engine before the larger turbine wheel of the 2835 starts spinning, whereas the GTRS has less to spin up (less lag). Noone will agree that a GTRS would make more power than a 2835 when both are run at or around 18PSI. Everyone knows the 2835 gives more up top (maybe 30rwkw or so? I dunno). You're analysis is too simple in some aspect or another. I would say maybe the data set, or some outriders are screwing up ur numbers...

Hey mate,

The graph I posted isn't progressive. By that I mean, the graph doesn't represent the build up of boost as the turbo spools up.

The graph represents boost vs power, with the boost remaining constant (i.e being set to a specific PSI)

The graph has nothing to do with how fast the turbo spools.

So when you say 'RE the graph you just pasted. I think it's because you have to wait until more air is being moved by the engine before the larger turbine wheel of the 2835 starts spinning, whereas the GTRS has less to spin up (less lag).', I wasn't refering to spool time at all, I was refering to max power output at a defined pressure.

Do you understand what I mean?

I think we are getting mixed up here with to different concepts, one being that the GT-RS will produce more power at earlier rpm than the GT2835 (what you're talking about, I think?) and secondly, max power output for a defined PSI (what I am talking about).

Nonetheless, the analysis I've done is better than nothing at all, and people can still use it as a ROUGH guide as to what power output they would acheive.

As for outliers in the data set, I've already removed them. For example, someone got 234 RWKW at 17 PSI using the GT2835. Way too low in My opinion, something must have been wrong.

And another one where some guys acheived 300 RWKW using the GT-RS. (way too high)

The data may be inaccurate however because people have different mods (ie. computer, intercooler, exhaust etc etc), and also that the data set may be too small.

Nonetheless, I have my fingers crossed when I install my GT2835 on monday....

Cheers

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



  • Similar Content

  • Latest Posts

    • I'm selling my personal race car. 1992 BNR32 Skyline GT-R About eight years ago the engine developed the dreaded low oil pressure problem. I removed it and found the issue. However shortly after I moved half way across the country, moved my work shop, bought a house, got married, then moved my shop again to it's permanent location for the time being. in turn I do not have the time to race this car as it should be. In turn the engine I built for this car will end up in my street car and this car I hope finds a home with someone who can use it to it's full potential. The roll bar was built by Jason Noren Fabricaiton in Pomona California to SCCA and NASA Specs. The car used to compete in redline time attack and global time attack enthusiast class. It also has a fire suppression system.  The car is with me at my workshop in Kyle Texas, USA. Shipping it is no problem, I can personally deliver the car to the port of either Houston or Los Angeles. Any other port depending on where it is may require transport but we can always work that out.   The chassis itself is rust free, rear quarters and under fenders are free of any rust so it's a very solid foundation for anyone who wishes to move forward with it. As posted and at the asking price of $20,000 USD the car includes all of the following;   All OEM Body panels, front fenders, rear bumpers etc.  OEM Nissan Projector headlamps (Pictured N1 lamps were removed) OEM Nismo Side Skirts / Extensions OEM Rear Spats  Front & Rear Subframes are installed car is full roller Rear Subframe has Cusco Camber Arms, Whiteline HICAS Elimination Front Subframe has Whiteline tension rods and whiteline Sway bars w/ endlinks.  Buddy Club N1 Coilovers All chassis wiring and engine wiring is in tact. Attessa System is fully in tact and was functional before engine removal.  BCNR33 GT-R Brembo Brakes F+R with Endless MX72 Plus Pads, DBA Rotors & Stainless Steel Brake lines Billion Racing Radiator Billion Radiator Hoses Fluidyne Engine Oil Cooler Accusump System (trunk Mounted) 5 Speed Transmission w/ Transfer Case F+R Prop-shafts Rear Differential With rear Axles.  Nismo Gauge Cluster Nismo Clutch + Flywheel   Essentially what is needed to make this car running and driving again is as follows: A working Engine, an ECU, and time to put it together. The price posted is or best offer as I am willing to remove some components to accommodate price to a certain extent. If any serious buyer wants full detailed pictures, videos, or any questions I'll be happy to answer them. I believe this car is priced to move quickly considering it's a rust free and very straight chassis but always open to fair negotiation.  I can be contacted via phone or email. (+19517081648) email is juan@bardabe.com   
    • So...to my (and my mechanics doing the swap) surprise...the oil pan from NEO engine does not fit 100% and needs to by modified like this: It is not a huge problem but in the future i want a "proper" oil pan. So question is...what different oil pan can i get to fit without any problems and modification? Thanks! 
    • Some more info I found.  These are the last entries Yoshikiyo Fujii made on his blog before he passed away in 2009: http://blog.livedoor.jp/fujii_dynamics/ And finally this is where the workshop was situated:  https://maps.app.goo.gl/HhTPtHzt3WVcBTiEA  
    • shit boxes, kids, and daily drivers are all welcome
    • Cam seals go hard. It's a very old car.
×
×
  • Create New...