View Single Post
Old 05-26-2007, 04:43 PM  
Scooby Specialist
Member#: 88322
Join Date: Jun 2005
Chapter/Region: MWSOC
Location: Indianapolis, IN
2009 BMW 135i


Just a quick write up for running a 4 port with your 2-port EWG. I'm going to be brief and assume you've read post #1 thoroughly.

Grainger PN 6JJ40
ARO Premair 4 port pneumatic solenoid with speed control
There's another one that is otherwise identical but without speed control. I don't think it is needed, but I got it to play with.

Basic principles of using a 4 port with EWG are the same as running a 3 port with a single port IWG.

The point of the 3 port is to be able to get a wider pressure range to the wastegate. From vacuum from the intake up to full boost out of the compressor.

(Quick aside: You can certainly run a 3 port with your EWG just like an IWG. Just use the bottom wastegate port. Leave the top port open to air. It works just fine.)

Running a 4 port on an EWG allows you to get the same range on both the upper and lower control port. You get full intake vacuum to full boost, and vice versa, on both ports. When the solenoid clicks on, it swaps which port gets the vacuum and which gets the boost.

Here's how the lines run. The blue lines represent how the ports are connected when the solenoid is at rest (0% WGDC, or low side of the PWM cycle). The red lines show how the ports are connected when 12VDC is applied to the solenoid (100% WGDC, or high side of PWM).

Here's a basic 3 port "bleed" setup on EWG vs. a 4 port "dual interrupt" (what I'm talking about):

Since this setup not only releases pressure from the push-open side, but also moves that pressure to the push-closed side, it is very reactive. In addition, most EWG come with much stiffer springs (~1bar instead of 0.5-0.6bar like a stock TD04 or VF--). This means it takes less duty cycle to reach boost, and small changes in duty cycle have a large effect.

When replacing a 2 port stock to a 3 port, I generally recommend multiplying your Initial WGD, Max WGD, Turbo Proportional, and Turbo Integral by about 0.7 (i.e. reduce by about 30%) as a starting point. Your actual WGD might drop from say 80 to about 55.

With an EWG, you might only need 20-30% total WGD to reach 20 psi Remember, you'd hit 15 just on the WG spring by itself, and the 4 port. Thus, you need to reduce your initial, max, and turbo closed loop controls by even more than a 3 port on a 8psi-sprung stock type turbo.

Here are some example maps to give you an idea of what kind of changes to make. (FWIW, a 3 port only running bottom EWG port is not too far from this, at least it is way closer than stock mapping).

The initial maps here are just alittle lower than what it takes to hit target boost. Since I know I can't hit 18psi at 3200rpm, I don't ask the car to try. I still have the WGD really high. Basically following the boost threshold of this car's setup (Perrin GT30 STI).

Notice how low the values are in the proportional and integral maps. This is important. The stock map has values about 2-4 times as large and will cause spiking and oscillation.

You may not recognize some of these maps, but they're there and are useful. Merchgod has put them in the Enguinity definitions. I'll talk about each of the odd ball maps briefly.

With a large rotated turbo and 1 bar spring in the WG, the WGD is going to matter much at low rpm or when desired boost is way less than the spring pressure. I don't want the car trying to think too hard trying to hit boost at a rpm so low it's not going to matter. Or, more technically, I do not want the car to evaluate the integral term, because it can wind up or down (look up integral windup on Google, wikipedia or buy a control theory book). Basically, if you know the control won't affect the output, you don't want your loop evaluating!

I changed the RPM to enable turbo closed loop control up from about 1600rpm to 2200rpm.

I also increase the desired boost threshold 1D map (on this car it is like 0.3 bar default). It just doesn't matter what the solenoid is doing if you're only trying to hit +0psiG, so I set it to not even worry about it until 1.1 bar (about +1.5psi). The stock mapping on most of the DBW cars seems to set this map to about 1/2 the spring pressure in absolute. I.e. if the stock IWG spring is ~8psiG, that's about 22.5psiA, and the stock map turns on the PID control loop at about 10psiA desired boost. So with a 15psiG spring, that's about 30psiA, or 2 bar, so I set the ECU to not even bother trying to effect output when its effect is so minor. You could probably go as far as to set this to about 5psi less than the spring. Might be messing with this on your setup.

There is also a revolving integral min/max that defaults to -20 and +15 on this car. To keep it from running away on me, and since small changes have a big effect, I cap it at only +4.5. This means integral can only be able to pull up the initial WGDC map by 4.5%. Again, this is to avoid integral windup issues. It's just about as effective to change this cap as it is to use the Max WG duty map to limit WGD. Perhaps even better in some respects.
* Registered users of the site do not see these ads.

Last edited by Freon; 12-31-2014 at 04:36 AM.
Freon is offline   Reply With Quote