Classic turbo problem...... & solution ?
#1
Subaru Tuning Specialist
Thread Starter
Join Date: Jun 2002
Location: 7.74 @179 mph 1/4 mile - road legal
Posts: 6,654
Likes: 0
Received 1 Like
on
1 Post
With a number of 2.0 scoobs now pushing over 400 bhp there are a few of us looking for a suitable turbo for the job.
Ideally we want to have our cake and eat it
Turbo 'wish list'
1- Responsive turbo that starts to spool up around 2500 rpm
2 - Produces at least 1 bar from 3000 rpm
3 - Safely makes maximum boost of at least 1.8 bar from 3500rpm
4 - Supports 450+ bhp.
5 - Affordable ie no titanium or exotic 'one off' parts.
So far there have been a number of turbos fitted by various tuners, including myself, which have achieved a maximum of 4 of the above.
Is this an over ambitious or even an impossible target ?
The problem is mainly due to the compressors being used cannot span a wide enough range of airflow to satisfy both ends of the RPM scale. If the upper end is capable of 450+ bhp then the lower end is running into surge at anything approaching full boost at less that 4000 rpm. (Fails No 3 above)
This applies to EVERY compressor map I have found so far.
A common way to avoid running into surge is to 'clip' the exhaust wheel, this makes the turbine less efficient and spool up of the compressor happens later in the rpm range.(Fails No1 & 2 above) It also allows more flow through the turbine which can slightly offset the reduced efficiency.
I was thinking of another way to avoid surge but without compromising turbine efficiency or spool up time.
By controlling the airflow recirculated through the dump valve at WOT, the position on the compressor map could be moved across to the right. This would allow full boost pressure generation at a lower RPM.
Assuming that the turbine has the torque to spin the compressor to flow more air then more power at lower RPM would be the result.
To control the dump valve, I was discussing with John Banks the possibility of using the JECS boost control output signal to drive a 2 port solenoid valve tee'd into the dump valve signal line. At a preset rpm the signal would be bled away allowing controlled bypass of the charge air. Boost pressure control would be by Dawes or similar as the ECU ouput is 'otherwise engaged'
Initial thoughts are that this may allow 'cake & eat'
Has anyone tried this before Or can you think of any reason why it wouldn't work ?
Andy
Ideally we want to have our cake and eat it
Turbo 'wish list'
1- Responsive turbo that starts to spool up around 2500 rpm
2 - Produces at least 1 bar from 3000 rpm
3 - Safely makes maximum boost of at least 1.8 bar from 3500rpm
4 - Supports 450+ bhp.
5 - Affordable ie no titanium or exotic 'one off' parts.
So far there have been a number of turbos fitted by various tuners, including myself, which have achieved a maximum of 4 of the above.
Is this an over ambitious or even an impossible target ?
The problem is mainly due to the compressors being used cannot span a wide enough range of airflow to satisfy both ends of the RPM scale. If the upper end is capable of 450+ bhp then the lower end is running into surge at anything approaching full boost at less that 4000 rpm. (Fails No 3 above)
This applies to EVERY compressor map I have found so far.
A common way to avoid running into surge is to 'clip' the exhaust wheel, this makes the turbine less efficient and spool up of the compressor happens later in the rpm range.(Fails No1 & 2 above) It also allows more flow through the turbine which can slightly offset the reduced efficiency.
I was thinking of another way to avoid surge but without compromising turbine efficiency or spool up time.
By controlling the airflow recirculated through the dump valve at WOT, the position on the compressor map could be moved across to the right. This would allow full boost pressure generation at a lower RPM.
Assuming that the turbine has the torque to spin the compressor to flow more air then more power at lower RPM would be the result.
To control the dump valve, I was discussing with John Banks the possibility of using the JECS boost control output signal to drive a 2 port solenoid valve tee'd into the dump valve signal line. At a preset rpm the signal would be bled away allowing controlled bypass of the charge air. Boost pressure control would be by Dawes or similar as the ECU ouput is 'otherwise engaged'
Initial thoughts are that this may allow 'cake & eat'
Has anyone tried this before Or can you think of any reason why it wouldn't work ?
Andy
#2
My thoughts on this at the moment are:
1. Implement the gas turbine anti-lag, no lag without the loss of power off boost and exhaust melting of normal anti lag
2. Run a turbo that isn't so efficient, but can run higher boost to get the required flow.
Well, the first is some months away, the second is really a case of trial and error!
Not exactly the answer people are looking for though!
Paul
1. Implement the gas turbine anti-lag, no lag without the loss of power off boost and exhaust melting of normal anti lag
2. Run a turbo that isn't so efficient, but can run higher boost to get the required flow.
Well, the first is some months away, the second is really a case of trial and error!
Not exactly the answer people are looking for though!
Paul
#7
Scooby Regular
Join Date: Nov 2000
Location: 32 cylinders and many cats
Posts: 18,658
Likes: 0
Received 1 Like
on
1 Post
Apart from more expensive methods anyone have any thoughts on downsides of Andy's plan?
A 2 port solenoid is normally closed. By teeing it into the dump valve control line it would have no effect when the duty cycle is zero. By setting the throttle position vs RPM min and max duty cycle maps the same I can control the solenoid. At 3000-4500 RPM on full throttle I could use the 2 port to bleed off the control signal to the dump valve, hence intercooler exit pressure will force the valve open even at WOT and cause the dumped air to recirculate into the intake, reducing the PR across the compressor and improving flow by moving down and to the right on the compressor map - potentially perpendicular to the surge line = ideal If it won't flow well enough I could use a 3 port.
Well that is the theory... wish us luck or are we off our heads LOL?
What gains do you think we would get?
Do you think it would give more airflow through the engine than just keeping the boost down until 4500 RPM?
There should be no reason not to use a Dawes to control if we get rid of the surge issue?
Thoughts on why this might not work even if you think they are stupid are welcome, at least they can be considered and thrown into the melting pot.
A 2 port solenoid is normally closed. By teeing it into the dump valve control line it would have no effect when the duty cycle is zero. By setting the throttle position vs RPM min and max duty cycle maps the same I can control the solenoid. At 3000-4500 RPM on full throttle I could use the 2 port to bleed off the control signal to the dump valve, hence intercooler exit pressure will force the valve open even at WOT and cause the dumped air to recirculate into the intake, reducing the PR across the compressor and improving flow by moving down and to the right on the compressor map - potentially perpendicular to the surge line = ideal If it won't flow well enough I could use a 3 port.
Well that is the theory... wish us luck or are we off our heads LOL?
What gains do you think we would get?
Do you think it would give more airflow through the engine than just keeping the boost down until 4500 RPM?
There should be no reason not to use a Dawes to control if we get rid of the surge issue?
Thoughts on why this might not work even if you think they are stupid are welcome, at least they can be considered and thrown into the melting pot.
Trending Topics
#8
The problem with surge is it can happen because the air physically can't get out of the compressor housing fast enough, not necessarily into the engine.
What I'm talking about wrt to turbo choice is picking something like Bob's turbo. It doesn't quite do everything 100%, but nearly satisfies nearly all the requirements.
It's not possible to meet all the requirements, you can only pick the best turbo for your application. You can get closer by making ultra lightweight parts like titanium turbines and aluminium-berylium compressor wheels. There are other things you could do, but they are all going to cost serious dosh.
and I'm serious about the gas-turbine anti-lag.
paul
What I'm talking about wrt to turbo choice is picking something like Bob's turbo. It doesn't quite do everything 100%, but nearly satisfies nearly all the requirements.
It's not possible to meet all the requirements, you can only pick the best turbo for your application. You can get closer by making ultra lightweight parts like titanium turbines and aluminium-berylium compressor wheels. There are other things you could do, but they are all going to cost serious dosh.
and I'm serious about the gas-turbine anti-lag.
paul
#9
Subaru Tuning Specialist
Thread Starter
Join Date: Jun 2002
Location: 7.74 @179 mph 1/4 mile - road legal
Posts: 6,654
Likes: 0
Received 1 Like
on
1 Post
Paul
The intention is to get the air out of the turbo by recirculating a portion back to the suction. Just exactly as the dump valve does at part throttle.
Bobs compressor is barely any different to the 20g. 200 rpm shift in surge line between them but with corresponding reduction in efficiency ie higher boost threshold.
John
My thoughts were that the PR would not reduce, it would actually increase but the flow would also increase disproportionately (sp) and keep the compressor out of surge.
Andy
The intention is to get the air out of the turbo by recirculating a portion back to the suction. Just exactly as the dump valve does at part throttle.
Bobs compressor is barely any different to the 20g. 200 rpm shift in surge line between them but with corresponding reduction in efficiency ie higher boost threshold.
John
My thoughts were that the PR would not reduce, it would actually increase but the flow would also increase disproportionately (sp) and keep the compressor out of surge.
Andy
#12
You could do worse than to use a turbo which has a flow map similar to the following. Yes, the lack of the surge line is somewhat deliberate, because there isn't one as such. It's more like a soft surge limit, and this particular compressor will operate indefinitely in that area of the map.
Cheers,
Pat.
Cheers,
Pat.
#14
Scooby Regular
Pavlo -
Could you explain more about this "gas turbine" means of anti-lag ? Unless this is the technique of bypassing some intake air into the exhaust manifold, dropping some fuel into it, and igniting close to the turbo, which as far as I understood, would expose the turbo to significant thermal stresses.
If there's another form of ALS, I'd be interested.
[Edited by BoxerFlat4 - 3/3/2003 8:54:05 AM]
Could you explain more about this "gas turbine" means of anti-lag ? Unless this is the technique of bypassing some intake air into the exhaust manifold, dropping some fuel into it, and igniting close to the turbo, which as far as I understood, would expose the turbo to significant thermal stresses.
If there's another form of ALS, I'd be interested.
[Edited by BoxerFlat4 - 3/3/2003 8:54:05 AM]
#15
Scooby Senior
Join Date: Oct 2000
Location: Zurich, Switzerland
Posts: 3,105
Likes: 0
Received 0 Likes
on
0 Posts
Only immediate problem I can think of is that you'll be recirc'ing hot air, assuming you're i/c hasn't got temps down to ambient. Whether this would be a problem for the short time you'd be recirc'ing remains to be seen. And the gains would probably make it worth it.
I wonder how easy it is to strap in a 2nd JECS?
Richard
I wonder how easy it is to strap in a 2nd JECS?
Richard
#17
Subaru Tuning Specialist
Thread Starter
Join Date: Jun 2002
Location: 7.74 @179 mph 1/4 mile - road legal
Posts: 6,654
Likes: 0
Received 1 Like
on
1 Post
Interesting map Pat but I would need one that flows 320cfm at 2.8 PR. That seems way out of range for that map.
Richard
As the controlled bypass would only be of short duration ie between 3500 - 4500 rpm (ish) at WOT I don't think heating the charge air should be an issue. The volume of air recirc'd should only ever be a maximum of 27% of the total airflow consumed by the engine.
Consumption at 3500 rpm 2.8pr =320cfm. (approx 340lb-ft torque)
Recirc air = 120 cfm
Total flow = 440 cfm (on map and 73% efficient)
Consumption at 4000 rpm 2.8pr = 390 cfm (approx 350lb-ft)
Recirc air = 60 cfm
Total flow 440 cfm
Richard
As the controlled bypass would only be of short duration ie between 3500 - 4500 rpm (ish) at WOT I don't think heating the charge air should be an issue. The volume of air recirc'd should only ever be a maximum of 27% of the total airflow consumed by the engine.
Consumption at 3500 rpm 2.8pr =320cfm. (approx 340lb-ft torque)
Recirc air = 120 cfm
Total flow = 440 cfm (on map and 73% efficient)
Consumption at 4000 rpm 2.8pr = 390 cfm (approx 350lb-ft)
Recirc air = 60 cfm
Total flow 440 cfm
#18
John,
It's a round metal thing, it sort of tapers toward the bottom. It has chunks cut out of it, presumably to reduce weight, but apparently when you put it inside another round metal thing that looks like a tapered doughnut and turn it very fast, it moves air or something like that
Cheers,
Pat.
It's a round metal thing, it sort of tapers toward the bottom. It has chunks cut out of it, presumably to reduce weight, but apparently when you put it inside another round metal thing that looks like a tapered doughnut and turn it very fast, it moves air or something like that
Cheers,
Pat.
#19
Scooby Regular
Join Date: Nov 2000
Location: 32 cylinders and many cats
Posts: 18,658
Likes: 0
Received 1 Like
on
1 Post
Think a 3 port in line will be better to completely block the signal when activated (normally open) at the high points of the duty cycle. Hopefully it will flow enough when deactivated to get the signal through adequately so the DV doesn't leak otherwise. 2 port might not bleed off enough unless you use restrictors, but that may make the DV leak when you don't want it to.
LOL at enigma Pat Mail me on the exhaust when you have Impreza results BTW
[Edited by john banks - 3/3/2003 10:56:35 AM]
LOL at enigma Pat Mail me on the exhaust when you have Impreza results BTW
[Edited by john banks - 3/3/2003 10:56:35 AM]
#23
Andy,
I'm not suggesting this actual one, just something with a similar design, ie one which can be run at very low flow rates, even at high-ish pressure ratios. If you were to run this one at 2.8:1 PR then you'de probably be looking at 40-50% efficiency which is pretty pants, but if your FMIC is good, then at least you CAN operate there and let your FMIC take the strain of getting temps back down.
John,
my "problem" at the moment with regard to the exhausts is that pretty much everyone who wants one has their car off the road at the moment building projects, or they are running after market headers. The only car that currently has one on with stock headers is my Legacy, and it's fairly loud on that, I don't think it's any louder than a Scoobysport or H&S on the same car. Thing is, it's pretty good at dampening high frequency noise but not so good a low frequency... cars with headers tend to "bark" more than "woof" so that's presumably why it's quiet on them, but not so quiet on stock headers. Shame you ain't more local really, we could have just fitted it and if it was OK you could have kept it, otherwise I would have taken it back, IYSWIM.
Cheers,
Pat.
I'm not suggesting this actual one, just something with a similar design, ie one which can be run at very low flow rates, even at high-ish pressure ratios. If you were to run this one at 2.8:1 PR then you'de probably be looking at 40-50% efficiency which is pretty pants, but if your FMIC is good, then at least you CAN operate there and let your FMIC take the strain of getting temps back down.
John,
my "problem" at the moment with regard to the exhausts is that pretty much everyone who wants one has their car off the road at the moment building projects, or they are running after market headers. The only car that currently has one on with stock headers is my Legacy, and it's fairly loud on that, I don't think it's any louder than a Scoobysport or H&S on the same car. Thing is, it's pretty good at dampening high frequency noise but not so good a low frequency... cars with headers tend to "bark" more than "woof" so that's presumably why it's quiet on them, but not so quiet on stock headers. Shame you ain't more local really, we could have just fitted it and if it was OK you could have kept it, otherwise I would have taken it back, IYSWIM.
Cheers,
Pat.
#24
Subaru Tuning Specialist
Thread Starter
Join Date: Jun 2002
Location: 7.74 @179 mph 1/4 mile - road legal
Posts: 6,654
Likes: 0
Received 1 Like
on
1 Post
I see what you mean Pat. I'm not sure that with such a low efficiency that it would have a low enough boost threshhold though ?
The benefit of the 20g wheel was that it was so efficient it actually had a lower boost threshold than the 16g for the same turbine clip. Just need to get it stable to operate there though, hence the WOT bypass plan.
The benefit of the 20g wheel was that it was so efficient it actually had a lower boost threshold than the 16g for the same turbine clip. Just need to get it stable to operate there though, hence the WOT bypass plan.
#25
Scooby Regular
Join Date: Nov 2000
Location: 32 cylinders and many cats
Posts: 18,658
Likes: 0
Received 1 Like
on
1 Post
Don't need 120 CFM through the control line/3 port though just the dump valve, the hose is about 2-3cm wide from memory. The piston in the dump valve is probably approaching the size of an inlet valve on the head?
Pat, happy to try it on the same basis and could return it at my expense if it wasn't right? Email me if you think this is a possibility. If it is quieter than a Blitz Nurspec without bung on the back of Magnex downpipe an resonated centre then that will do me fine. Scoobysport sort of volume would be OK, but no louder. Might end up with headers again eventually - would possibly help this surge by reducing low down exhaust gas velocity and possibly increasing VE?
[Edited by john banks - 3/3/2003 11:31:06 AM]
Pat, happy to try it on the same basis and could return it at my expense if it wasn't right? Email me if you think this is a possibility. If it is quieter than a Blitz Nurspec without bung on the back of Magnex downpipe an resonated centre then that will do me fine. Scoobysport sort of volume would be OK, but no louder. Might end up with headers again eventually - would possibly help this surge by reducing low down exhaust gas velocity and possibly increasing VE?
[Edited by john banks - 3/3/2003 11:31:06 AM]
#26
Pat,
get a system shipped up to john while his car is still at 2litres and we will tell you how it does, you and only you will find out the results, if he posts anything without your go-ahead I will knock the **** out him for you .
get a system shipped up to john while his car is still at 2litres and we will tell you how it does, you and only you will find out the results, if he posts anything without your go-ahead I will knock the **** out him for you .
#27
Scooby Senior
Join Date: Oct 2000
Location: Zurich, Switzerland
Posts: 3,105
Likes: 0
Received 0 Likes
on
0 Posts
However, doesn't the DV only open on the signal line vacuum created in the manifold by snapping the throttle shut? How are you going to get sufficient vacuum to open the DV under WOT.
Assuming you could do so, then it's possible to rig a microswitch on throttle, plus a CR timer network, to parallel feed the existing boost solenoid's signal onto the 2nd DV solenoid for the required time. But I guess the required duration in 4th is double that required in 2nd? Maybe a gear-aware EBC is a better option? I'd prefer to keep my JECS for boost control to provide the mid-range/high-end differences I want.
Pat - I've a variety of headers I can use the system with, none of them stock Can I assume things are nearing completion?
Richard
Assuming you could do so, then it's possible to rig a microswitch on throttle, plus a CR timer network, to parallel feed the existing boost solenoid's signal onto the 2nd DV solenoid for the required time. But I guess the required duration in 4th is double that required in 2nd? Maybe a gear-aware EBC is a better option? I'd prefer to keep my JECS for boost control to provide the mid-range/high-end differences I want.
Pat - I've a variety of headers I can use the system with, none of them stock Can I assume things are nearing completion?
Richard