Preturbo 335i Strange Results

[rudypoochris]

Hey Guys,

I have been reading up on WAI for a while and finally decided to take the plunge. I wanted to make more power, but my stock turbos are already far off the compressor map. I decided on pre turbo hoping to get more efficiency out of them. I settled on a mechanical setup as these are supposedly going to have smaller droplet size. I just got this built and running and my results were not really what I expected.

Base setup before meth/water:

1. 2008 335i N54 Coupe
   
2. Compression 10:1 with direct injection running 50-50 E85/91 mix
   
3. Boost set to 17.5 psi (stock is 7.5 or so) so way off the map making 430whp
   
4. After market intercooler and tune, no other modifications (stock DPs, fully catted)
   
5. Currently limiting boost to save turbos. I am NOT octane limited at this time and could probably hit 19-20psi without losing significant timing
   
6. Turbo Tout is probably in the 350F range
   

Adding meth/water:


I went with a Wannaspeed mechanical kit that uses air mix nozzles. I modified it quite a bit. Two nozzles one in each cone filter injecting into 60mm stock plastic intake tubes which have bends and such for about 2 feet before entering the turbos. The 50% meth/water mix is driven from a trunk mount tank with a check valve on the air line (not progressive). Flow is measured using an Aquamist sensor with 100-1500ml/min scale (half graph is 750ml/min or so). Preturbo meth is triggered by boost switch at probably 8psi (just setup last night).


Pictures:







Just to clarify, none of the ports below are for charge pipe injection. That is all boost air to the tank or meth nozzles.










First Results:


I ran the car with and without the meth water running last night. I used the exact same map and ran within 5 minutes of each other. On the no meth run I was going downhill, on the meth run I was going up hill. Ambient was 55F the no meth run started at 59F and the meth run at 69F IAT.The intercooler is heat soaking way quicker with meth on presumably since the specific heat of the meth air mix is much greater than air only. Usually I see IATs no more than 10F above ambient. After one pull with meth I was at 25+ over.


No meth:





Meth:



Note the faster rising IAT, identical PWM (wastegate duty), and surprisingly reduce timing at start of run (fluke?).

The curve to look at specifically (at least I believe) is "PWM". PWM refers to the duty cycle of the wastegate as controlled by the car. More PWM means the wastegate is held more shut. You can see PWM spike during spool up and rise as the RPMs increase and the turbo is working harder to supply the required boost. In both logs, PWM is nearly identical. That is to say despite pouring 600-700 ml/min of meth water into the intake, the amount of work the compressor is doing has not dropped. I thought it would since the WA mix should be sucking up heat during compression leading to more air out of the compressor per turn (higher efficiency).

Am I thinking about this wrong? Am I set up wrong? Should I try biasing more water or more meth? Maybe the mix has completely atomized before the compressor due to my long intake tract? Do I need more flow?

Thanks guys,

Chris

[Richard L]

Interesting project.

Water at ambient is about 1000 times denser than air. I can only assume the compressor is doing more work when methanol was injected on top of air. It is natural to require more energy (PWM) to accelerate a rotating mass.

How is methanol flow regulated (flow control), is it progressive with boost or engine load?

[rudypoochris]

Originally Posted by Richard L

Interesting project.

Water at ambient is about 1000 times denser than air. I can only assume the compressor is doing more work when methanol was injected on top of air. It is natural to require more energy (PWM) to accelerate a rotating mass.

How is methanol flow regulated (flow control), is it progressive with boost or engine load?

So basically the shaft speed may be less, but the work is around the same? Then really the turbo is seeing less "damage" due to overspeed but roughly equal thrust loading wear?

Methanol comes on at a boost switch set too low atm (8psi). It is not progressive at the moment, but I may switch it to be progressive with boost. The check valve makes it not progressive since the tank is pre-boosted to whatever the last pull left it at.

Are other people seeing reduced shaft speed or reduced work with pre-turbo meth?

[rudypoochris]

I am also seeing something else funky. My DME is pulling timing when the meth hits some of the time. Usually this is due to knock, but it doesn't knock/pull timing at these same RPMs with no meth.







Timing loss is only on meth onset and I am not progressive. I wonder if its flooding or something.

[Richard L]

Originally Posted by rudypoochris

So basically the shaft speed may be less, but the work is around the same? Then really the turbo is seeing less "damage" due to overspeed but roughly equal thrust loading wear?

Methanol comes on at a boost switch set too low atm (8psi). It is not progressive at the moment, but I may switch it to be progressive with boost. The check valve makes it not progressive since the tank is pre-boosted to whatever the last pull left it at.

Are other people seeing reduced shaft speed or reduced work with pre-turbo meth?

Although water is heavy but compared to the mass of air going through the turbo, it is much less. There will not be to much variation on the BCV's PWM.

You cannot guarantee the distribution of water drop when it travels toward the compressor wheel. Some will pool on the flange wall and migrate/trickle towards the utmost tip of the turbo. This will exert most load on the turbo. Long term, it will unbalance and pit the compressor wheel, followed by excessive shaft wear.

I am not sure why you are using methanol instead of water. Water has more cooling capacity.

Until you have balanced the air against the meth flow fed to the nozzle, you cannot
guarantee how well the mixture is atomised. I suggest running run water for one test, clip the jet on the windshield and see how well the spray is atomised. (driving test)

If you have a air compressor handy, you can do it without driving the car. You can also measure exactly how much water/methanol is injected for 1 minute.

[Richard L]

Originally Posted by rudypoochris

I am also seeing something else funky. My DME is pulling timing when the meth hits some of the time. Usually this is due to knock, but it doesn't knock/pull timing at these same RPMs with no meth.

Are you sure it is knock? I would check the STFT and see you are having methanol inbalance distribution to the cylinders. You can do this by comparing the two banks.

Can you log individual cylinder timing? This is a real useful tool for your kind of project.

[rudypoochris]

Originally Posted by Richard L

Although water is heavy but compared to the mass of air going through the turbo, it is much less. There will not be to much variation on the BCV's PWM.

You cannot guarantee the distribution of water drop when it travels toward the compressor wheel. Some will pool on the flange wall and migrate/trickle towards the utmost tip of the turbo. This will exert most load on the turbo. Long term, it will unbalance and pit the compressor wheel, followed by excessive shaft wear.

I am not sure why you are using methanol instead of water. Water has more cooling capacity.

Until you have balanced the air against the meth flow fed to the nozzle, you cannot
guarantee how well the mixture is atomised. I suggest running run water for one test, clip the jet on the windshield and see how well the spray is atomised. (driving test)

If you have a air compressor handy, you can do it without driving the car. You can also measure exactly how much water/methanol is injected for 1 minute.

If PWM/work doesn't drop, what is the benefit of injecting pre-turbo? I thought the idea was to get more air through the turbo before going out of the efficiency range (increased efficiency). It seems if work in is the same and so is boost air out, that I haven't changed the efficiency at all. Or is it that the turbo is equally efficient, but can handle more flow before overspeed?

Are people actually documenting erosion? I haven't been able to find a solid source of it yet? It makes sense to me for large droplets or saturated walls, but it would seem the walls of the tubing in the engine bay are quite hot and the speed of the air high enough that pooling probably doesn't occur. I of course have no way of testing for this right now. I do know though that a lot of PCV is diverted through the turbo inlets and that does not erode the compressor despite being likely significantly larger unatomized drops.

Interesting. I might do the atomization test. I am not sure how knowing exact water/meth quantity will help. I do have the aquamist flow sensor feeding into the DME though and you can see it in the logs above. Half scale should be 600ml/min or so.

What do you mean by balancing the air against the meth flow? You mean the flow into the 2-port nozzles? The boost air will always be higher than the pressure of the fluid by a couple points. Say 17.5 psi boost with 15ps or so water/meth.

I am using a combination of water/meth since I speculated meth evaporates quicker (reducing chance of erosion) and because the meth helps out with the fueling a little bit. I could go full water or full meth if I needed I guess.

[rudypoochris]

Originally Posted by Richard L

Are you sure it is knock? I would check the STFT and see you are having methanol inbalance distribution to the cylinders. You can do this by comparing the two banks.

Can you log individual cylinder timing? This is a real useful tool for your kind of project.

I need to flash the DME to log individual cylinders. I could do this if I got a cobb... maybe I should do it. How could there be an imbalance if I am injecting pre-turbo into both turbos? Even if there was imbalance, I have more than enough E85 in the fuel itself to cover the boost targets I am testing at. It could be the DME reducing timing for some kind of torque reduction (according to the tuner), but I don't think that it. Weird.

The STFT are on the graphs as "Trims". They are almost maxed out since I run a lot of E85 with no flash.

[Richard L]

Dismiss the STFT for the time being. In the absence individual cylinder log, it is not easy to go further on finding the cause of knock.

I cannot think of any way to pin point how the timing was trimmed, other than it was caused by meth injection.

[Richard L]

Originally Posted by rudypoochris

What do you mean by balancing the air against the meth flow? You mean the flow into the 2-port nozzles? The boost air will always be higher than the pressure of the fluid by a couple points. Say 17.5 psi boost with 15ps or so water/meth.

I am using a combination of water/meth since I speculated meth evaporates quicker (reducing chance of erosion) and because the meth helps out with the fueling a little bit. I could go full water or full meth if I needed I guess.

It is quite tricky to set up an air assisted nozzle. You really need minimum amount of pressure on the fluid and ample air pressure. If you have too much fluid pressure, the atomisation is not that good.

Look at the following pictures, the difference is very visible.