TURBOs!?

[Tapani]

:-)

First I will come up with a setup to measure both inlet and exhaust pressure and temperature and map these against air mass flow. Then try to define the design envelope for the most efficient ball bearing twinscroll candidates.

Do not wait for results very soon, though

Kind regards,

Tapani

[bazza]

Originally Posted by Tapani

:-)

First I will come up with a setup to measure both inlet and exhaust pressure and temperature and map these against air mass flow. Then try to define the design envelope for the most efficient ball bearing twinscroll candidates.

Do not wait for results very soon, though

Kind regards,

Tapani

EG33 have one bank firing cylinders 240' apart. The EJ2X turbo has one bank firing 180' apart (1 4 2 3) and the stock header for the EJ2X is near identical to the EG33 header and is what Perrin copied and what I also did. EJ2X header is good for 100 bhp per cylinder before the pulses create to much of a restriction. Means the EG33 is good for about 600 bhp theoretically before the stock headers become a restriction. I saw 580 odd bhp but the cams were the issue. Now cams fixed I expect it will hit 600 bhp no problems at 6500-7000 rpm.

All EJ2X turbos have always been greater than 1.5:1 exhaust to inlet pressure ratio. VF34P20, one of the best turbos available for the EJ207 is around 2:1 at 20psi when you push the boost up to 18-20 psi like what pretty much every second WRX owner does.

It's only when you stick a big turbo say GT35R on the 2.5L engine when you see 1:1 pressure ratios.

[Tapani]

Yeah,

The two issues must be dealt separately (to begin with). The first concern in my car is the overall restriction the small turbine introduces and potentially the exhaust pressure goes thru the roof and strangles the breathing above 5500 rpm. The stock shiftpoints are around 6300 rpm and the engine does feel struggling...., but lands exactly at the sweet spot after an upshift and pulls strong (for 5 PSI ). This I will try to measure and map across the rev range. Let's see what I can come up with.

The other issue is the basic exhaust arrangement and whether the pressure pulses work against or for us - and how this affects the basics of the cam designs. I do not have a strong opinion for now, but the twin scroll might not be a bad idea to entertain with stock cams and the IRIS system working.

http://www.autozine.org/technical_sc...duction_4.html

http://www.modified.com/tech/modp-09...n/viewall.html

Also:

To summarize, pulse turbocharging is combined with constant pressure turbocharging using a conventional turborcharger in applications that are configured appropriately. The exhaust pulses need to be adequately spaced apart, or else they interfere with each other, diminishing or negating the benefit, the theory being that it is advantageous to recover the high velocity energy available from each individual cylinder's exhaust blowdown event by maintaining the velocity and directing it to the turbine inlet. When these events are spaced too close together within the exhaust manifold, they cancel each other out to the extent that they overlap (one cylinder is in effect exerting backpressure on the other). In practice, a separation of 240 crank degrees is consider optimimum, with 180° being too narrow (having interference between adjacent cylinders in the firing order), and 360° being so wide that the low velocity phase of the impulse allows the turbine to lose momentum between impulses. Applications that utilize pulse turbocharging will have exhaust manifolds that are notably smaller in cross section than those that are designed as purely constant pressure systems. The small cross section maintains the desired high gas velocity and concomitant kinetic energy to the turbine inlet. The ideal engine configuration for pulse turbocharging is the in-line six, with the exhaust manifold split between the front and rear groups of three cylinders each. The turbocharger has a divided turbine inlet with the division maintained throughout the turbine housing all the way to the end of the nozzle section. This keeps the cylinder exhaust events within each half of the divided housing spaced apart 240° (the firing order being either 1,5,3,6,2,4 or its reverse, 1,4,2,6,3,5). This configuration represents practically 100% of today's commercial vehicle diesel engine market. The net benefit of pulse turbocharging is faster spool up and a steeper boost curve. Once the boost is being controlled by the wastegate (e.g.), the benefit is unrealized. Coming back to Bill's question, individual header tubes to the turbine inlet, conforming to the principles of pulse turbocharging described above, is a working concept.

Current internet-wisdom suggests that the evolution of modern turbos may have introduced potential to solve the two issues combined while keeping the low boost levels and the thermal strain on the engine in check. Kind of a "less is more" way of thinking.

The attached clips greatly illustrate the basics of the related time factors which are very important in a road car - also in a race car, I would think.

Kind regards,

Tapani

[bazza]

Originally Posted by Tapani

Yeah,

The two issues must be dealt separately (to begin with). The first concern in my car is the overall restriction the small turbine introduces and potentially the exhaust pressure goes thru the roof and strangles the breathing above 5500 rpm. The stock shiftpoints are around 6300 rpm and the engine does feel struggling...., but lands exactly at the sweet spot after an upshift and pulls strong (for 5 PSI ). This I will try to measure and map across the rev range. Let's see what I can come up with.

The other issue is the basic exhaust arrangement and whether the pressure pulses work against or for us - and how this affects the basics of the cam designs. I do not have a strong opinion for now, but the twin scroll might not be a bad idea to entertain with stock cams and the IRIS system working.

http://www.autozine.org/technical_sc...duction_4.html

Current internet-wisdom suggests that the evolution of modern turbos may have introduced potential to solve the two issues combined.

Kind regards,

Tapani

I'm running a 2011 MY11 Forrest S with a very modern VF52 and it's the same as the old turbos - high exhaust manifold pressure at high inlet boost. Nothing has changed here, just has a nicer bend coming off the turbo into the TMIC.

As I said, stick a GT35R on the engine and the pressure is more 1:1 because the turbo is nice and big for the 2.5L - the GT35R has been around for 10 years and the only update by Garret is a new billet compressor wheel which increases compressor efficiency and spool.

[bazza]

Originally Posted by Tapani

Yeah,
The other issue is the basic exhaust arrangement and whether the pressure pulses work against or for us - and how this affects the basics of the cam designs. I do not have a strong opinion for now, but the twin scroll might not be a bad idea to entertain with stock cams and the IRIS system working.
Kind regards,

Tapani

I'd bin the IRIS valve and let the turbo do all the work. Although your issue is most likely the turbo being to small.

Dunno if you saw this, 10 psi pulls nicely. Ignore the jagged boost and power lines - car was wheel spinning on the dyno. If you put that turbo in a daily driven 3.3L it would be the most driveable car about. If you're falling over at 3-5 psi with a smaller turbo then certainly upgrade. Twin scroll is nice but a more expensive solution and usually for people with an EJ2X 2.5L / GT42R type setup who want increased spool.

['E']

bazza what are your thoughts on a GT3576 a/r .65 on the compressor and a/r 1.17 twin scroll turbine? its one of the options im playing with

['E']

the other option im playing with is twin GT2560r with a/r .60 compressor and a/r .64 turbine, i have both options already but your thoughts would be appreciated

[Tapani]

Originally Posted by bazza

I'd bin the IRIS valve and let the turbo do all the work. Although your issue is most likely the turbo being to small.

If you're falling over at 3-5 psi with a smaller turbo then certainly upgrade. Twin scroll is nice but a more expensive solution and usually for people with an EJ2X 2.5L / GT42R type setup who want increased spool.

I would not be looking for maximum horse power - maybe low 300s, but to let it rev freely to the redline and retain the N/A V8 feel it has now. Transient torque chracteristics are very important, I do not wish to loose any of this.

So, you may say that the end result should be like a 21st century OEM set up . Boring, no kick in the pants .

Tapani

[bazza]

Originally Posted by 'E'

bazza what are your thoughts on a GT3576 a/r .65 on the compressor and a/r 1.17 twin scroll turbine? its one of the options im playing with

I think that would certainly work.

Originally Posted by 'E'

the other option im playing with is twin GT2560r with a/r .60 compressor and a/r .64 turbine, i have both options already but your thoughts would be appreciated

That would also work. The only issue with twins is the amount of extra work and cost to get it setup. I considered it for mine with 2 x TD05h's or VF34's but CBF doing the extra work... 2 dump pipes... water pipes... oil feeds and returns. One turbo is annoying enough haha.

Originally Posted by Tapani

I would not be looking for maximum horse power - maybe low 300s, but to let it rev freely to the redline and retain the N/A V8 feel it has now. Transient torque chracteristics are very important, I do not wish to loose any of this.

So, you may say that the end result should be like a 21st century OEM set up . Boring, no kick in the pants .

Tapani

Sounds like my setup. The reason I ditched the EJ257 for the EG33 /GT35R setup is the ability for the motor to give both huge midrange and good top end power. The GT35R turbo I selected is rather small however it's very linear power and instant response across the entire rev range. If you want a lot of power with the EG33, you really need the GT40R or bigger. I could easily get 1000Nm of torque before 4000 rpm with my setup and a bit more boost - is that enough torque for you?

['E']

Originally Posted by bazza

That would also work. The only issue with twins is the amount of extra work and cost to get it setup. I considered it for mine with 2 x TD05h's or VF34's but CBF doing the extra work... 2 dump pipes... water pipes... oil feeds and returns. One turbo is annoying enough haha.

I was thinking of using twin vf34's as well (because the compressors are more compact then the gt25's) and using the plumbing, brackets (but obviously deleting the sequential components) to reduce fabrication and keeping it looking oem as possible

[oab_au]

Originally Posted by Tapani

Yeah,

The two issues must be dealt separately (to begin with). The first concern in my car is the overall restriction the small turbine introduces and potentially the exhaust pressure goes thru the roof and strangles the breathing above 5500 rpm. The stock shiftpoints are around 6300 rpm and the engine does feel struggling...., but lands exactly at the sweet spot after an upshift and pulls strong (for 5 PSI ). This I will try to measure and map across the rev range. Let's see what I can come up with.

The other issue is the basic exhaust arrangement and whether the pressure pulses work against or for us - and how this affects the basics of the cam designs. I do not have a strong opinion for now, but the twin scroll might not be a bad idea to entertain with stock cams and the IRIS system working.

The attached clips greatly illustrate the basics of the related time factors which are very important in a road car - also in a race car, I would think.

Kind regards,

Tapani

Bazza. EG33 have one bank firing cylinders 240' apart. The EJ2X turbo has one bank firing 180' apart (1 4 2 3) and the stock header for the EJ2X is near identical to the EG33 header and is what Perrin copied and what I also did. EJ2X header is good for 100 bhp per cylinder before the pulses create to much of a restriction. Means the EG33 is good for about 600 bhp theoretically before the stock headers become a restriction. I saw 580 odd bhp but the cams were the issue. Now cams fixed I expect it will hit 600 bhp no problems at 6500-7000 rpm.

All EJ2X turbos have always been greater than 1.5:1 exhaust to inlet pressure ratio. VF34P20, one of the best turbos available for the EJ207 is around 2:1 at 20psi when you push the boost up to 18-20 psi like what pretty much every second WRX owner does.

It's only when you stick a big turbo say GT35R on the 2.5L engine when you see 1:1 pressure ratios.

This is one of the things that you can't relate back to the 4 cylinder engines, when looking at the 6 cylinder.

The 4 fires each cylinder 180* apart, so that when No.1 has opened its exhaust valve to evacuate the cylinder at about 60* BBDC, its piston continues to the bottom, and is on the way up the bore, 60*from the top, when, 180* later, No.2 opens its exhaust valve to pressurise the system, pushing gas into the No.1 cylinder. So the piston has 60* to push the gas back out.

The 6 cylinder fires each cylinder 120* apart. So when it opens No.1 exhaust at 60* BBDC. its piston is only at 60* ABDC when the No.2 opens to pump up the pressure in No 1. that now has 120* to push the gas back out.

So the interference pressure hits the 6 cyl 60* earlier, to foul the charge. That’s why the pipe connecting the two sides has to be long enough to delay the gas pressure from getting to the other side, till its piston is closer to the top of the bore.


The EZ30R that Perrin used is in another situation as it has variable inlet valve timing, and lift.

Low speed inlet timing.


High speed inlet timing.


When it is at the low speed, fully advanced timing, the inlet valve is open for the 47* that the piston is pushing the gas back, not only into the exhaust, but into the inlet system for that 47* to foul the inlet charge. When the engine gets to 4500??? (don’t know what rpms it switches), the inlet timing goes to full retard and the inlet valve is closed, so that the gas is pushed back down the exhaust only, to let the engine breath clean charge.

So it is no wonder why that engine came to life when the timing changed. When it is on advanced timing all the boost, just blows down the exhaust pipe, till the exhaust closes.

This timing is no good for blowing, and will suffer from the short connecting pipe interferance, more that the EG33 or EZ30.

Harvey.

[bazza]

Originally Posted by 'E'

I was thinking of using twin vf34's as well (because the compressors are more compact then the gt25's) and using the plumbing, brackets (but obviously deleting the sequential components) to reduce fabrication and keeping it looking oem as possible

Yeah they work really nicely. Although wastegate is usually about 7 psi which with 2 x VF34's is a fair bit of power. Also later you could get the P20 rear housing for a bit more power.

[bazza]

Originally Posted by oab_au

This is one of the things that you can't relate back to the 4 cylinder engines, when looking at the 6 cylinder.

The 4 fires each cylinder 180* apart, so that when No.1 has opened its exhaust valve to evacuate the cylinder at about 60* BBDC, its piston continues to the bottom, and is on the way up the bore, 60*from the top, when, 180* later, No.2 opens its exhaust valve to pressurise the system, pushing gas into the No.1 cylinder. So the piston has 60* to push the gas back out.

The 6 cylinder fires each cylinder 120* apart. So when it opens No.1 exhaust at 60* BBDC. its piston is only at 60* ABDC when the No.2 opens to pump up the pressure in No 1. that now has 120* to push the gas back out.

Okay, EJ engines 101. The EJ20 has the firing order of 1-3-2-4 not 1-2 etc as you've suggested and the EG33 has 1-6-3-2-5-4.

This means 1 and 3 on the EJ fire 180' apart and 1 and 2 are obviously 360' apart. On the EG33 1 and 3 fire 240' apart which means the EG33 cast iron header is actually better off than the EJ2X cast header as it gets a touch more time between firing to clear.

Also the merge is a long way from the exhaust valves and tends not to be a problem as shown with GT Spec headers on the EJ's. The merge is more about ensuring the pipe design is big enough to ensure good flow and no restriction - for example my first header design used a 3" merge into the up-pipe and was crazy responsive.

So the interference pressure hits the 6 cyl 60* earlier, to foul the charge. That’s why the pipe connecting the two sides has to be long enough to delay the gas pressure from getting to the other side, till its piston is closer to the top of the bore.


The EZ30R that Perrin used is in another situation as it has variable inlet valve timing, and lift.

Low speed inlet timing.


High speed inlet timing.


When it is at the low speed, fully advanced timing, the inlet valve is open for the 47* that the piston is pushing the gas back, not only into the exhaust, but into the inlet system for that 47* to foul the inlet charge. When the engine gets to 4500??? (don’t know what rpms it switches), the inlet timing goes to full retard and the inlet valve is closed, so that the gas is pushed back down the exhaust only, to let the engine breath clean charge.

So it is no wonder why that engine came to life when the timing changed. When it is on advanced timing all the boost, just blows down the exhaust pipe, till the exhaust closes.

This timing is no good for blowing, and will suffer from the short connecting pipe interferance, more that the EG33 or EZ30.

Harvey.

You you've never tested this or tuned any sort of Subaru boxer turbo system alone without bring AVCS or AVLS into the mix and then understanding how the theory relates to the real world. The one thing with theory is it never addresses all the variables. I did rather well in fluid dynamics and thermodynamics at one of the best Uni's in Australia and many years ago I used to spend hours doing the theory - however one of the things I quickly learnt with engines after sitting on dynoes and modifying the shizzle out of these things for a long time is there are to many variables for any theory to deal with. The theory is really only best applied when you have a working setup and you want to take the next step however of course this then requires testing to determine if it works or not. For example, we all know the best engineers in the world can spend 12 months doing the most perfect theory and get it absolutely wrong as seen in the F1 over the years.

So my advice to readers is start with a proven design and then innovate from there. Going on theory without proven results is a very risky business especially when it's your coin being spent at the end of the day.

[bazza]

Further:

An old set of WRX headers (EJ2X):


Here you can see the merge - the diameter is nasty yet this setup supports 100 bhp with the 1-3-2-4 firing order and high rpm. You can see how close the runners merge and they fire 180' apart.


The "GT Spec" header - long runners. An upgrade for EJ's above 100 bhp (per cylinder) due to the very small diameter in the headers and nasty merge above. Note the merge location doesn't change, it's about the same length from the exhaust valves.

[bazza]

My header version 1.0. 3" up-pipe ensure massive flow at the merge. Nice fat GT35R turbine and EWG ensure low header pressure.


My header version 3.0. Planning to run > 600 bhp eventually so will need to seperate the pipes at the port level:


Cast iron stock header:


What not being conveyed and my point is that for low boost and low power applications such as Tapani and Bobswood, the stock header is absolutely fine. If you've got the coin to go nuts - but don't think you HAVE to build a fancy header system from the get go - that's coin you can spend wisely elsewhere - like a fully tunable ecu etc.