219 replies to this topic

[bmhvanbenthum]

This a schematic drawing of the custom manifold i'm thinking of. In low RPM the center butterfly valve is open and the butterfly valve to the secondary turbo is closed. The header pipes to the junction with the up-pipes and the cross-pipe are of equal length. In reality the manifolds of cylinder 1-2 and 3-4 will be directly connected by the butterfly valve without a lot of pipe, so no cross-pipe but only cross valve. So on primary turbo mode all pulses arrive equally at the turbo charger. Switching to twin turbo mode the center butterfly valve will close and the butterfly valve in the secondary up-pipe will open. Then you will have real twin operation. The front bank is driving the secondary turbine while back row of cylinders is driving the primary turbo. The exhaust are all the same length, so equal pulses again.

 

So it will switch from single turbo charging to full parallel turbo charging.

 

According to some Matlab simulations the exhaust gases will reach temperatures up to 700 degree C. So I need to construct some butterfly valves that can hold these temperatures. I want to link them mechanically, one closes the other opens. This way I can still use the ECV actuator. I was thinking to fabricate them out of stainless steel, with normal steel bush bearings to prevent seizing 

 

Or would it be better to keep the cross pipe open, or drill a hole in the valve so there still is a tiny opening after closing??

 

About the position of the secondary turbo butterfly, is it wise to but it before the turbo, or is it better to put it behind the turbo, just like Acres is doing?

 

Another strange idea of mine is to switch from single to single turbo charging, anybody thought about this ?

 

I will try to use the waste gate port of the primary to pre-spool the secondary, but I don't know when it opens normally. Are the boost levels adjustable in the ECU? is the switching point from single to twin adjustable in the ECU? What can be adjusted in the stock ECU ?

[acres]

^ I would think the center butterfly is probably not the best idea unless you replace both turbos with 2 identical models. The stepped turbine sizing will have flow differences so the cross over pipe is necessary to balance the pressures. Wastegate control would also be poor with the center valve closed.

 

 

I might actually finally get to fire this up shortly. Within the next couple of days I should think.

I just have to refit the engine, drain the fuel tank and replace the fuel pump and assembly and re-plumb the black box to suit how I will control the solenoids with the megasuirt.

 

 

 

[duncanm]

 

I will try to use the waste gate port of the primary to pre-spool the secondary, but I don't know when it opens normally. Are the boost levels adjustable in the ECU? is the switching point from single to twin adjustable in the ECU? What can be adjusted in the stock ECU ?

 

Here's a thought I've been having the last few days... why not use the ECV to waste exhaust pressure, rather than the wastegate when on primary - thus pre-spooling the secondary?

 

Only problem I can think of is that it would disturb the exhaust flow direction constantly.

 

That's a pretty impressive set of pipework you've built.

[allpaw4]

Wheres the AFM on that acres? Looks impressive, and I was thinking it might be after the turbos, but that would likely upset the intake air temp setting in the Ecu. Certainly interested to see how it goes. I take it its setup like that so everything bolts up on the car?

[Jimbo]

[acres]

No AFM, I am going to run a MAP based tune. I ditched the pcv valve that is normally threaded into the manifold under the throttle body and drilled and tapped the hole to suit an IAT sensor to enable map tuning.
I just bolted up all the pipework while it's still on the stand as a bit of a final check to make sure everything still lined up properly. It's amazing how easily the pipes warp out of their original location while welding.
The wastegate no longer lines up 100% but I should be able to force it into position when bolting it all back up.

[acres]

So I got it in and running the other week and by the next day I was pulling it back out.

It had a slight coolant leak from both turbos, a slight oil leak from the primary and was pissing out oil from the secondary. Since the package is so tight it is impossible to remove even the turbo's while the engine is in the car, so unfortunately to fix the leaks it all had to come back out. Because the turbos are twisted I had to modify the oil drain lines from the turbos so that they would head in the right direction so my initial assumption for the major oil leak was that I had cocked up the welding of the steel oil drain line on the secondary and the seam had busted open. The other leaks I could see was because the spring clamps I had used on the lines were too light and weren't clamping tight enough. Once out though, I realised that the cause of the major leak on the secondary was not from the welding but simply because some hack hadn't cleaned the old gasket off properly so the flange wasn't sealing once it was bolted to the turbo. After calling myself some filthy names I replaced the gasket, replaced the spring clamps with some heavier units on the primary oil drain and replaced the spring clamps on the coolant lines with some screw clamps and proceeded to put it all back in again.

Got it fired up again not long after and didn't lose a drop of any fluids. Happy days. 

I then concentrated on getting the tune sorted for cranking, cold starts and idle. I was getting really inconsistent results with the initial fueling so after jumping on the megasquirt forum and looking for similar issues came to the conclusion that maybe the power feed I had chosen was the incorrect choice. The oem setup uses 2 different power feeds, with the second feed activated from the ecu itself so when first setting up the patch loom I naturally chose the most direct power feed which is sourced directly from the ignition switch. The second feed also powers the fuel injectors, so in an effort to fix the inconsistencies I switched the power feed to the megasquirt to this second feed. This made a marked improvement and I was now getting consistent and stable starting. From what I can tell the reason for this is because the second feed holds some residual voltage after turning the key off. So with the megasquirt shut down and the injectors still exposed to this residual voltage they become unstable and open uncontrollably till the residual voltage is drained. It is only a very short amount of time but it is enough to inject excess fuel into the manifold when the engine is no longer running, which then causes the inconsistencies next time a start is attempted.

With that sorted I was tuning the idle when a horrible death rattle developed. It sounded terrible but was nothing major though as it was obviously coming from the exhaust. The second hand truck valve that I was using as the ECV had failed. This valve actually had a little bit of play in it to begin with and with just that little bit of use it had seen the shaft had flogged out and was now rattling around as the exhaust pulses hit it. I also used an identical valve as the IC valve for the secondary but this one had zero shaft play and so far seems to be fine.

So as it stands I just need to source another valve and then I can get back into it.

 

It does sound good though when it is running. It is not tinny at all like some headers can sound.

 

 

[Hypnotic]

Sick dude, good to see it's up and at least running!

[CRUISN]

Thats a lot of work. Inspiring!

[duncanm]

Fantastic stuff.

I went back re-read your earlier post - you are using the ECV (your improved butterfly version) for boost control until both turbos online. Good move - opening the wastegate on primary-only always seemed like a waste of good exhaust gas!

I think your definitely on the right track with the headers and moving the ECV well downstream of the secondary.

[acres]

Thanks guys. I was pretty pumped when I finally backed it out of the garage for the first time in years.

Duncan, my thoughts exactly on wasting the exhaust when on primary.

I'll try and explain a little better how I have the ecv and wastegate plumbed up.
I am using a dual port external turbosmart unit as the wastegate. I have the top port of the gate plumbed into the black box to one of the on/off solenoids. This solenoid will switch between manifold pressure and atmo. At lower rpms it will be switched to manifold so that boost from the primary will hold the wastegate shut, above a threshold ~5k rpms it will switch to atmo allowing the wastegate to then do it's thing. The bottom port is teed into the positive pressure port of the ecv, which is plumbed to the black box and directly to the ecv duty solenoid. The on/off solenoid which is normally present before the ecv duty solenoid has been bypassed. The factory setup uses this on/off solenoid to delay the secondary prespool to a narrow rpm band. The negative port of the ecv is plumbed to the black box as normal and is controlled via the one on/off solenoid - switching between vaccum and atmo. While on primary the wastegate will be held shut so the ecv duty solenoid will be able to directly control the duty of the ecv which will both control boost and pre spool the secondary over a wider range. Once the negative port of the ecv is activated and the ecv is held at full open and the wastegate top port is switched to atmo, the ecv duty solenoid will then be able to directly control the duty of the wastegate giving normal boost control at high rpms.
The IC valve is no longer controlled by the black box. It is opened by a wastegate actuator which is plumbed directly to the secondaries compressor outlet nipple, so it should be controlling it's flow independently.

[Hoody]

Great thread. Can't wait to see a video of this

[Shadoxity]

Fantastic stuff.

I went back re-read your earlier post - you are using the ECV (your improved butterfly version) for boost control until both turbos online. Good move - opening the wastegate on primary-only always seemed like a waste of good exhaust gas!

I think your definitely on the right track with the headers and moving the ECV well downstream of the secondary.

FYI, the std tune in the B4 (at least the 01-03 models) is to use the ecv from about 1800rpm and 4psi instead of the primary wastegate. 

 

 

Thanks guys. I was pretty pumped when I finally backed it out of the garage for the first time in years.

Duncan, my thoughts exactly on wasting the exhaust when on primary.

I'll try and explain a little better how I have the ecv and wastegate plumbed up.
I am using a dual port external turbosmart unit as the wastegate. I have the top port of the gate plumbed into the black box to one of the on/off solenoids. This solenoid will switch between manifold pressure and atmo. At lower rpms it will be switched to manifold so that boost from the primary will hold the wastegate shut, above a threshold ~5k rpms it will switch to atmo allowing the wastegate to then do it's thing. The bottom port is teed into the positive pressure port of the ecv, which is plumbed to the black box and directly to the ecv duty solenoid. The on/off solenoid which is normally present before the ecv duty solenoid has been bypassed. The factory setup uses this on/off solenoid to delay the secondary prespool to a narrow rpm band. The negative port of the ecv is plumbed to the black box as normal and is controlled via the one on/off solenoid - switching between vaccum and atmo. While on primary the wastegate will be held shut so the ecv duty solenoid will be able to directly control the duty of the ecv which will both control boost and pre spool the secondary over a wider range. Once the negative port of the ecv is activated and the ecv is held at full open and the wastegate top port is switched to atmo, the ecv duty solenoid will then be able to directly control the duty of the wastegate giving normal boost control at high rpms.
The IC valve is no longer controlled by the black box. It is opened by a wastegate actuator which is plumbed directly to the secondaries compressor outlet nipple, so it should be controlling it's flow independently.

 

Loving your progress. Some solid crazy changes to the system that you have done!

Something that might help you out though, in my many hours I have spent tuning my one, you get much better response and results by leaving the ECV closed until a certain amount of boost and revs. Allows much faster spool on primary, and better boost holding.

 

My theory, which turned out well after enough testing, was that if we can be moving more air using the primary by itself, the ramp up of the secondary turbo will be faster rather than trying to share the low rev/low air between the 2 and regulating the boost by bleeding through the ECV like the stock tune does.

 

I can pull out the maps if you want some info on when I have it changing and what boost etc if that might help?

What ECU are you using?

[acres]

I managed to source a new generic unit of the isuzu butterfly from a bloke in Sydney so it's all back together again now. I thought I should test all the solenoids in the BBOD though before I start working on the tune again and sure enough the ECV negative solenoid had become sticky and wasn't opening when activated. It would click but wouldn't open. I tried to give it a clean but no luck so I switched the hoses around for another solenoid that was no longer being used and that one works fine. 

So with that sorted it is on to tuning.

It now fires up cleanly and idles pretty well. It no longer wants to stall and is happy to run the ac and/or fans at idle and now has zero sync issues with cam and crank signals too. So it is getting somewhere with the tune.

I can't remember where I got my numbers from for the initial spark and fuel maps as it was some time ago when I set them up in the tuning software but they most certainly require a lot of refinement. I have taken it for a few drives up the street to take some logs and have been just adjusting the fuel to target some better afrs. It was too lean in some areas and way too rich in others so I have just been smoothing this out before looking at touching the spark maps which I am guessing are going to require alot more advance since as it stands the car is very slow, makes very little boost and has high egts under this low boost.

 

For tuning the spark, I knocked up an amplifier circuit that I found online for a microphone that I have attached to a clamp and clamped this onto the block next to the knock sensor. I have run the cable through the firewall and can plug in a set of headphones so that I can listen for knock when winding up the timing. It works really well as I can hear all the valvetrain noise, etc but at the same time the headphones mask the sound of the exhaust, wind, etc. I haven't heard it knock yet, which is a good thing, but the timing is still pretty low.

 

Once I get it performing better I will post up some logs and some videos of it running. 

 

FYI, the std tune in the B4 (at least the 01-03 models) is to use the ecv from about 1800rpm and 4psi instead of the primary wastegate. 

 

 

 

Loving your progress. Some solid crazy changes to the system that you have done!

Something that might help you out though, in my many hours I have spent tuning my one, you get much better response and results by leaving the ECV closed until a certain amount of boost and revs. Allows much faster spool on primary, and better boost holding.

 

My theory, which turned out well after enough testing, was that if we can be moving more air using the primary by itself, the ramp up of the secondary turbo will be faster rather than trying to share the low rev/low air between the 2 and regulating the boost by bleeding through the ECV like the stock tune does.

 

I can pull out the maps if you want some info on when I have it changing and what boost etc if that might help?

What ECU are you using?

 

That's some great feedback in regards to the ecv operation, I might have to test out this theory too. 

 

If your happy to share, I would love to see any maps or logs of your's in operation? I would love to see a log of map/timing/afr? Just so that I can get closer to the ballpark. It would also be very interesting to see the primary response and changeover with your ecv setup.

 

I still have those ecu defs you sent me some time ago for the b4 and they have been pretty handy for getting the maps close to where they needed to be in regards to starting, warm up and idling.

 

That's the problem with running a Megasquirt, you are pretty well starting from scratch in regards to the tune. The base settings they come with aren't really all that relevant for Subaru.

 

For any one interested;

Megasquirt - http://www.msextra.com/index.html

and can be bought here - http://www.diyautotune.com/

[Shadoxity]

I attached some old logs (July 2012! lol) There should be a lot more newer logs but i cant seem to find them right now. WIll have another look.

 

Thats on a map that is about 8 revisions behind what I have been running for the last 2 years lol.

The ones in 123007 dont have the wideband attached cause I was testing out some other things and fueling was all good by then.

That map was running 17psi on primary and secondary. 

 

It will give you a good idea about boost and timing etc, but those logs are all starting from close to 3000rpm, so I am not sure what I was testing back on that map.

 

 

120710 has some more full logs from low revs. These are even older though, but that was at a fairly advanced stage in the rom. After these ones it was testing on things to affect the changeover and holding boost. I will do some fresh logs later next month as im going to give it a bit of a touch up tune after doing the solenoid box with new hose.

 

 

Couple of pointers to look at to get an idea of some things.

 

Ex. Gas Positive Pressure (On/Off)

Goes to a '1' when the EGC valve starts as the boost control and start spooling the secondary.

Ex. Gas Negative Pressure (On/Off)

Goes to 1 when it switches to twin mode

Intake Air Solenoid (On/Off)

goes to 1 as above, which opens the secondary to the intercooler.

 

 

 

The std ones are logs from a car Im about to tune which has the std tune. It will give you a better idea (hopefully) about how the std one works (take note of the ex gas positive pressure) and how slower the boost builds.

 

Ill dig up some other stuff when I get a chance. The timing and fueling maps would probably be helpful to you.

 

Attached Files

•  123007.zip   18.7KB   13 downloads
•  std.zip   8.64KB   5 downloads
•  120710.zip   29.84KB   5 downloads

[acres]

Ta mate, appreciate it.

 

Have downloaded and had a quick squizz. Do you know if there is a program that can interpret those logs into a graph? Be a bit easier to read that way and I can't see any way for romraider to translate them.

 

Thanks

Aaron

[Barbbachello]

This thread hurts my brain in a good way

[Shadoxity]

Ta mate, appreciate it.

 

Have downloaded and had a quick squizz. Do you know if there is a program that can interpret those logs into a graph? Be a bit easier to read that way and I can't see any way for romraider to translate them.

 

Thanks

Aaron

 

 

Dont know about a program, but Excel has some half decent graph ability if you know how to use it (Which i dont know to well lol)

Oh actually, Airboy's spreadsheet might give you graphs from that data? I dont know, I have only used it for timing and maf scaling before - http://www.romraider...m/topic957.html

 

This might help.

http://www.romraider...php?f=33&t=6156

 

I normally go into excel and go view > Freeze panes  > Freeze top row.

That way the header row gets locked in place and everything else scrolls and you can see what the column headers are.

And cause most of the logs are just a WOT run, from top to bottom is just going from low revs to high and you can get an idea of what is happening.

Just found this tool which makes things a bit easier to read.

http://www.romraider...php?f=32&t=5446

 

EDIT: Oh and found this as well. Could do the trick

http://www.romraider...php?f=32&t=6003

 

Or this

http://www.datplot.com/

[acres]

Bit of a late reply, got swamped with the build up to christmas and just couldn't find the time to work on it. Airboy's spreadsheet (it's got me buggered how somebody can make excel do that) did the trick though. After transferring the numbers from the logs you provided it was able to translate the numbers into a graph. So using that info as a bit of a guide and making a few runs I have so far ended up with some maps that look like this:

 

 

 

Throttle response is excellent and acceleration is smooth however power is still poor and primary boost is shocking. Boost builds better at around 4200rpm when the secondary kicks in but will still only reach a max of 1 bar, generally less than this though.

 

This is an image from a log with this tune. You can see how MAP drops off and the injectors are also hitting 100% duty at 6000rpm. With boost at only 11psi though this shouldn't be happening, so obviously it's running way too rich - which the spark plugs can confirm with their thick layer of soot. At the point where MAP flattens the AFR is registering as 12.8, which is making me doubt the accuracy of this wideband.

 

 

This image is from the next log where I added 3 degrees of timing to all cells above 100kpa and pulled 3% from the fuel map in all cells above 150kpa and after 5100rpm. This created some light knock which I could pick up through the headphones but interestingly the afr went richer with the extra timing even though i pulled fuel too. You can see the AFR drops to 12.4 in the image.

 

 

I have noticed 2 small cracks in the headers which may be contributing to the poor boost build up so I will get that sorted before pushing forward again. I also obviously need to replace the injectors as the 440's aren't going to flow enough.

Once it's boosting properly and I am positive there is no more electrical issues it will be taken to the dyno for the pros to wind in the power.