ChrisRX8PR

Hello Guys. I'm looking for opinions on a potential turbo change. This is my build thread:

https://www.rx7club.com/old-school-o...ersion-988372/

I currently have a Turbonetics T-72 HPC Billet Turbo with a divided 1.0 A/R. It has a 72x102mm compressor and a 68mmX76mm turbine wheel (F1).

The turbo hasn't given me any issues but the plans for the car have changed...I would like to keep it at 600whp or less with a better powerband. The turbo I currently have, although it was a great deal for a BB turbo (main reason I got it), it a bit mismatched as far as available airflow from the compressor relative to turbine wheel size. This means that for the power I currently have (about 525-550whp) I am probably loosing 800-1000rpm of spool with this turbo due to the larger mass of the compressor etc which is not being used to the fullest.

I have a few choices to correct this, I can buy a backplate, Housing and wheel from turbonetics and turn it into a 64-66mm front and keep my current turbo. This will probably be around $500-600.

I could also sell my turbo which only has about 2000 miles since new(was replaced under warranty) and buy a Borg Warner better suited to my setup/goals. This brings me to two choices:

S362FMW (older style) with a 1.00A/R Divided - 62x83 Comp and 68x76 Turb(basically same as now) - http://www.ebay.com/itm/141038441780?_trksid=p2060353.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT
S363SXE (New) with a .91A/R divided - 63x88 Comp and 73x80 Turb(bigger than now) - http://www.ebay.com/itm/221845378968?_trksid=p2060353.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT

These are just links for info and does not mean I will buy the turbo there. I'm leaning towards getting a Borg instead of changing my current turbo simply because I could sell the turbonetics and only spend about $100 net to switch to it.

Of the Borgs I am leaning towards the S363SXE but am afraid that although its smaller up front, the larger rear will give me a similar powerband to what I have now and no change will be noticeable. It does however provide a slight edge as far as potential if I can to go up a bit in power.

Im looking for feedback as far as power attainability with either of these turbos and what your experience has been regarding reaction time etc.

Opinions welcome.

I have no interest in any other turbos so please keep the discussion to these options.

Thank you.

Chris

Howard Coleman

iTrader: (6)

Hey Chris,

while i haven't (yet) visited your thread i sure like the idea of putting a better rotary in the RX8..

you have a truly huge turbo for your power target, it has 9.55 sq inches of compressor and 6.6 of turbine area.

given your max power goal (600) you should be looking for no more than 8.1 sq inches from pre 2015 turbos.

given there have been recent tangible increases in efficiency, it is not out of the realm of possibility that a 62 mm (6.54 sq inches) BW SXE might do 550+ (73+ pounds per minute).

Borg Warner suggested that the compressor maps for the SXE line that they had shared w me at the PRI last Dec have been further improved so most of us should be quite happy w the 62 or 63/64 SXE offerings.

the older 62 FMW billet turbo (that you mention) is now two steps behind the new SXE 62 that is currently available at dirt cheap pricing (PN 13009097056).

a key consideration is that there are two SXE "62" turbos currently.

they are the same except for turbine wheels. IMO the larger hotside wheel is too large for the compressor.

13009097056 has the "76" wheel

13009097053 has the "80" wheel

the 7056, priced around $750 might just be one of the single most attractive turbos for much of the rotary community.

stepping up to midsize you have the two SXE offerings at 62.99 (7.18) and 64.4 (7.06). i assume they are similar to the S300 63 midsize of yesteryear but w significant efficiency upgrades. the S300 63 makes 550 happily and can do a bit more when pushed.

my suggestion would be either the 7056 or 7047 depending on your power objective...

500-550 7056

600 7047, 7055

i have a 7056 and will be testing it w the .91 and 1.0 after i finish my 9180.

good luck,

Howard

ChrisRX8PR

Howard, as always your data analisis is much appreciated.

I did find the s362sxe as well and the price is very attractive but the only rear housing I could find for it was 200.00 which made it unatractive considering that I can get an s363sxe for the same price and get a housing for $85.

I am still leaning towards the s363sxe because of its ability to reach 600whp but am still worried about lag in relationship to the current turbo with it having a bigger turbine wheel regardless of its smaller compressor. Do you think it will still spool sooner than my current turbo if I use the .91 divided?

Thank you for taking the time.

Chris

ChrisRX8PR

Bump! Anyone else with an opinion or experience?

Chris

ChrisRX8PR

So, some new info came to light. AGP turbo has the compressor flow in LBS for each of the compressor wheels used in the SXE series of turbos.

They are like this:

62mm - 76lbs
63mm - 78lbs
64.5mm - 80lbs

Assuming these are accurate, they are not as big a difference as I expected between them. According to info Howard had posted in the past this should equate to roughly 580, 598 & 613whp respectively. Am I correct in these numbers??

If this is the case I may actually choose to go with the 6268 for quicker spool...trading about 30hp for a fatter powerband.

Opinions?

Thanks!

96fd3s

Does not sound right to me.

Full race lists the below sizes available and flow for each;

62mm / aka 8376 = 75lb
64.5mm / aka 8780 = 82lb
66mm / aka 9180 = 94lb
69mm / aka 9180 69mm - 100+ lb

that looks like a more reasonable difference between compressors to me

ChrisRX8PR

I see, but aren't those numbers for the EFR line of turbos and not for SXE? Are the compressor wheels the same? I thought they weren't. SXE comes in a 63mm as well which is the one I was talking about. Either way that is more like what I expected. If your numbers are correct, the 62 compressor should be enough for about 575whp which I'm leaning towards because it will spool way faster than the 63 compressor with the much bigger 73mm turbine wheel. If I could get the 63 compressor with the 68mm turbine I feel that would be ideal for me but that is currently unavailable.

96fd3s

Nope, well sort of. It would appear that up till now they were used with the EFR line.

I guess (now) its just a size designation, not a part number, and does not take into account if its EFR/ Air Werks etc

link to two of them. all of them can be viewed there

http://www.full-race.com/store/turbo...8376-62mm.html

BorgWarner S300SXE 9180 - 66mm - Full-Race.com

arghx

iTrader: (3)

Let me come at it a different way, as I've done low end torque optimization on engine dynos and in cars. Response is really two different but related things: available energy driving the turbine wheel and the inertia of the rotating assembly. When it comes to increasing response, here are two things to consider:

1) the mass of the wheel and its rotating inertia - this affects the spin up, how fast the compressor wheel increases with time. The size of turbine A/R has less of an effect on this transient turbocharger wheel acceleration aspect. Having a heavier compressor wheel matters for this.

2) the max available boost when lugging at low rpm - this is the potential torque curve if you locked speed on a loading dyno, or did a 5th gear pull. The mass of the wheels have much less of an effect on this. A higher mass compressor wheel doesn't matter so much.

So let me ask you to do something on your current setup, if it's roadworthy. You can do this on a loading dyno but that's usually not an option. Take the car out to a safe area. If possible hook up your laptop to your engine management and take a datalog

1) Put the engine in 5th gear at 800-1000rpm. Yes, go that low. Floor it until you get up to about 3500-4000rpm, safety permitting. At least get up to 3000rpm.

Take a look at the rpm vs boost curve. Post the log here.

On a given type of porting and exhaust system this is the potential boost curve according to the turbine A/R and to a lesser extent the turbine wheel design. The mass of the compressor wheel, bearing design (journal vs ball bearing), has very little to do with it.

-- When the turbine A/R gets bigger, this boost vs rpm at 5th gear line gets worse. The max available torque decreases. Boost begins at a higher rpm. Backpressure at high engine speed decreases.

-- When the turbine A/R gets smaller, this boost vs rpm at 5th gear line gets better. The max available torque decreases. Boost begins at a lower rpm. Backpressure at high engine speed increases.


Now, here's another datalogging exercise. This is also better with a loading dyno and a turbocharger wheel speed sensor, but you gotta make do what what you have.

Put the engine into high speed, like say 4500pm, foot off the gas. Just freewheel. Now slam on the gas. Besides the boost control tuning and bearing design, the time it takes to reach full boost when you tip in at high rpm is determined by the mass of the compressor and turbine wheels: rotating inertia. The turbine A/R has a lot less to do with it, because at high rpm you have enough exhaust gas flowing.


If the low speed lugging is not important to you (going uphill without downshifting for example), then making the turbine housing smaller is less beneficial. If transitioning in and out of boost at high rpm (think twisty backroads driving) is less important, then decreasing the inertia of the rotating assembly doesn't matter as much.

ChrisRX8PR

First, I really appreciate your time writing this out.

The car is currently partly disassembled (I just began taking things out to take care of some loose ends I left when originally doing the swap) so taking logs is out of the question. That being said, I took lots of logs...more importantly I completely understand what you are explaining and agree 100%. I will do my best to describe what I like/dont like about the current setup based on my previous turbo rotary experience.

My main problem was that the mass of the compressor was evident between shifts and when going instantly into throttle at higher rpms. I actually did not find the 5th gear "potential" boost curve all that bad, I've felt much worse in other setups to tell you the truth. Since I was not utilizing the full potential of a 72X102mm compressor that is where the idea to downsize the compressor came from. The car actually chugs along at low rpms quite well and builds some boost from as little at 2200rpm. The problem is that the instant explosion when letting out the clutch between shifts that I had felt with other turbos was gone and there was more of a build up. That would be ok if you get a 700whp explosion after the buildup..but for 550-600 I knew I could do better. And this was with a 50mm Tial BOV and a relatively soft spring so it was definitely not surging (it was also not soft enough to cause it to bleed air).

I just really wish I could get the SXE with the 63 or 64.5mm compressor on the 68x76mm wheel but I either get the 62mm w the wheel I want or go bigger on the wheel which may be counter productive for what Im trying to accomplish. I may try out the 62mm compressor...it should hit high 500s before running out of air.

What do you think?

Cheers!

arghx

iTrader: (3)

I know lots of people swear by different brands and such, but really in my experience it comes down to matching more than anything else--the right combination of individual aspects such as wheel trims and housing A/R. The technology isn't drastically different between major manufacturers (Mitsu-based, Garrett-based, BorgWarner-based). If one company doesn't offer the combination you want, you should consider a different company if it's at all feasible.

It sounds like you want medium compressor wheel, medium-large sized turbine housing, and if possible ball bearing for transient improvement (ball bearing does nothing for the low speed potential torque line we've been discussing).

Let me ask this: are there any Garrett-based GTX turbos you are considering? The reason why I ask is that I find there to be a lot of combinations of housings and wheels to choose from for Garrett GT turbos.

Do you think a turbo like this GTX4088 is too big? Garrett GTX4088R Turbocharger, including Turbine Housing (GTX-R Series) : atpturbo.com

it seems to be close to what you're looking for, but maybe cost is the issue. I wonder if the 1.0ish turbine housing A/R would be the way to go though, because of the frame size.

Speaking of frame size:

One thing that's not always clear when comparing turbos with different A/R's is the effect of the frame size or other dimensions of the turbo. So there are two metrics you see for turbine housing size: area, in cm^2 (10 cm^2, 18, whatever), and A/R.

Area is usually used for Mitsubishi (Greddy etc), IHI (Apex'i etc), and Holset based turbos. A/R is usually used for Borg Warner (Bullseye etc) and Garrett/Honeywell (Turbonetics etc) based turbos. So you have to be careful with A/R. The smaller A/R could in some cases not really be smaller in terms of absolute size, as A/R is just a ratio.

KNONFS

iTrader: (14)

Had to get on the highway during lunch hour, and decided to try this test (since I wasn't planning on testing today, I did not get a chance to log it):

5th gear at 1000rpms is about 50mph on my car. Pedal to the metal, and the engine took FOREVER to reach 4000rpms. I am not even sure if the engine got to boost (again, no data logs), however the engine must have been on cell maps that are not tuned properly, since the wideband afr went from 10.X to reading RICH It is safe to say that full boost was not reach (14psi) since the waste gates were not screaming.

This is a fairly safe test, I believe 5th gear at 4000rpms is somewhere around 70mph.

My setup is a stock port REW, TO4S turbo, .91 turbine housing, divided manifold, twin 38mm waste gates.

Will attempt again, and will data log. Curious to see which fuel cells I was hitting on that test.

Shainiac

iTrader: (7)

For the sake of comparison, here is a fifth gear spool test on my FC.
13BT 4port, S5 NA 9.7:1 rotors, stock intake ports, slightly earlier opening exhaust port.
GT4088R (not GTX)
EBay FD manifold modified to fit FC, Tial 46mm wg, 10# spring, no boost controller, audibly opening by 5psi
500cc/min water meth, rich as hell street tune
JDM trans with 0.8:1 fifth gear (should hurt spool?)
3" exhaust, large resonator, 7 INCH OD muffler! It's quiet.

arghx

iTrader: (3)

^ With the raw data file you can actually do a whole analysis on the spool by using a tool I have to process boost rate of change curves. There are also some tricks to extrapolate a VE curve to these kinds of cells so that it's not so rich.

I can do a visual analysis of the curve in your logs though.

First, the reason for this 1000-4000rpm test is to very roughly approximate a horsepower rating run on an engine dyno, which is done at very slow engine speed ramps. I'm sure very few of you have worked with engine dynos, but on an engine dyno you can use the brake to hold the engine speed or ramp the speed up and down while sitting there at full load. Loading chassis dynos can do this too. So check this out:



When you see an advertised torque curve like this (GM 2.0 liter turbo engine, successor to the Pontiac Solstice engine), it's a lot different than dyno pulls done on a dynojet. The low end torque looks a lot higher usually, and the "flat" nature of it is because the data is averaged over time. In the rating run tests the dyno brake ramps the engine speed very slowly, so you are seeing a potential torque curve in a quasi steady state.

The size of the wheels doesn't matter so much, because they have plenty of time to accelerate. This is where the turbine A/R is very important--since the time to spin up the wheels is effectively unlimited, the low end torque on a boosted engine is limited by turbine energy. It's one of the reasons why stock turbos have such small hotsides--so the stock torque curve looks more impressive, and journalists who don't know any better can be like "all the torques. right off idle!"

Now let's look at your 5th gear pull running only wastegate spring pressure.



Looking at the yellow MAP line (in kilopascals, where ~100 kpa is atmospheric pressure, 170ish is 10psi, 200 is roughly 14.5psi) we can see several regions of turbocharger operation. First there's a quick jump in manifold pressure as the throttle opens. The rate of that jump has to do with how fast the throttle valve opens and whether there are any misfires due to poor tip-in tuning. On a cable throttle it's straightforward. On electronic throttle there is usually some kind of dampening or delay in throttle opening.

The line between "throttle open" and "WG cracks" is the main thing we are talking about here regarding turbine A/R sizing. Since we are increasing engine speed very slowly due to the gearing, the turbine A/R is the dominant factor controlling the shape of this line rather than the size of the rotating assembly. If you use a smaller A/R, or switch from say an undivided to a divided manifold, this line between "throttle open" and "WG cracks" will shift upwards. Port timing can also affect this but that can be complicated.

The point where "wastegate cracks" occurs depends on boost control tuning initial duty cycle. Typically you start with a high duty cycle on the solenoid, to vent pressure out of the actuator, and then drop to a lower duty to crack the gate open. There is a wastegate opening pressure setting on most external controllers which sets that, and on standalone ECUs it's built into duty cycle tables or some single scalar setting for when to begin control.

The rest of the MAP curve would depend on the solenoid duty cycle setting and feedback system, in addition to the flow capability of the wastegate. In this case it's running no boost control, so it's purely mechanical. If you've got the right setup you can actually force the wastegate full open, run the engine WOT on a loading dyno, then ramp it from ~1000rpm to redline and see what kind of boost it builds. That will tell you whether the mechanical design of the wastegate system is flowing enough for your needs at higher rpm--the lower the boost in that test, the better your wastegate and manifold.

Now I also suggested a transient test where you coast to a specific rpm and do a hard tip in, as a way to test the inertia of the turbocharger rotating assembly. That's just an approximation of something you can do on a loading dyno with locked engine speed:



In the image above, the engine is locked at a fixed speed (1500rpm here, although it will deviate slightly) with an engine dyno brake. Then the rate of torque increase is measured (NMEP is a type of torque curve excluding friction). The rate in the boosted region is highly influenced by the inertia of rotating assembly.

Now in the image above they've got a piston engine at 1500rpm to simulate an engine cruising around, locked in top gear. In an Rx-7 that would reflect a feeling of delay when you climb up a hill around town in 5th gear and you want the turbo to spool up to its potential torque curve so you don't have to downshift. If you do such a transient test at 4500rpm, it will reflect any sense of delay you would feel in a backroads driving or autocross application.


So when we talk about boost curves and low end torque, we've got all these effects working together:

1) mechanical design of the wastegate and manifold
2) inertia of the turbocharger rotating assembly (size, bearing design)
3) tip-in fuel tuning
4) turbine energy, especially as it relates to turbine A/R
5) boost control tuning (wastegate opening point, feedback control)

Shainiac

iTrader: (7)

Thanks for the analysis.
I forgot to mention, I have a .95 AR twin scroll exhaust housing and a divided twin scroll manifold.
I'm also running water-to-air IC, so very short plumbing from the turbo to the throttle body.
Also, my VE maps is wack because of some incorrect injector dead times and overall poor tune. I'd be interested to see how a leaner mixture would effect spool. Less mass through the turbine housing vs higher velocities and more heat?

arghx

iTrader: (3)

Raising the exhaust temperature should give you a little more boost in that area. That can be accomplished by leaning it out--the temperature effect normally outweighs the reduction in total mass flow out of the exhaust ports.

Leaning out may raise boost a bit in this kind of lugging test but still not matter enough to be noticeable when driving normally though. Leaning it out will also give more torque because it improves the efficiency of combustion. Retarding timing will increase boost in this kind of test by raising exhaust temperature but reduce the efficiency of combustion. You pretty much have to do a bunch of tests to understand the sensitivity of these factors--it could be a big difference or hardly any.

Either way, if something as fundamental as injector settings needs to be redone, you probably have a bunch of tuning to do.

Shainiac

iTrader: (7)

I was experiencing boost creep from 10psi spring pressure to 14 psi by the end of 4th gear. I added 2 degrees of timing at 6000rpm and the boost creep actually dropped 1psi the instant the timing advances.

BLUE TII

iTrader: (9)

First, I really appreciate your time writing this out.

The car is currently partly disassembled (I just began taking things out to take care of some loose ends I left when originally doing the swap) so taking logs is out of the question. That being said, I took lots of logs...more importantly I completely understand what you are explaining and agree 100%. I will do my best to describe what I like/dont like about the current setup based on my previous turbo rotary experience.

My main problem was that the mass of the compressor was evident between shifts and when going instantly into throttle at higher rpms. I actually did not find the 5th gear "potential" boost curve all that bad, I've felt much worse in other setups to tell you the truth.

I am going to jump in here with something beside choosing between the two turbos you mentioned because you stated you currently have your set up apart.

Exhaust velocity!

This is not only related only to the AR you have on the exhaust housing.

I have experience with using the same turbo and same AR exhaust and simply enlarging the exhaust manifold and turbo exhaust runners.

The turbo response (punch) is much lower with the larger exhaust manifold volume despite full boost being reached at exactly the same rpm.

Take a look at your exhaust manifold up to the turbo and see if you can improve velocity with smaller diameter runners. The manifold runners should be no larger area than the turbo exhaust housing runners' entry at the turbo flange.

Instead of 2" Sch 10 on a long runner manifold use 1.75" Turblown RX-8 REW swap manifold or make your own 1.75" one.

If you have excellent exhaust velocity you wouldn't have a problem with soggy response from the large compressor, but rather compressor surge.

Just decreasing the exhaust housing AR is not the same thing especially on the long runner RX-8 manifold.

In this case you still drop velocity in the manifold and then stack up enough exhaust at the turbo entry to accelerate the exhaust gas again, but at huge cost to pumping losses.

It is more beneficial to size your exhaust manifold runner diameter for good velocity.

arghx

iTrader: (3)

^ That's another good point.

This becomes a system wide optimization problem and it's hard to say what the trade off is for change in wheel dimensions, change in turbine housing, and change in manifold. Each change has an associated cost in money and time.