Hi,
Have found this thread through the TDI board, very interesting reading!!
I am also carrying out a VNT conversion at the moment and have gone through surprisingly similar processes and problems!
Libbypapa re. your poste above:
"The control system I imagine for the vanes would be as follows.
I would first have the vanes be normally open (spring pressure).
I would have a throttle arm controlled spring to exert pressure to close the vanes. I would run another cable off the trottle arm, to a spring at the vane control. Whenever throttle was increased it would increase pressure to close the vanes.
I would add a control that would be rpm dependant. With increase in rpm, there would be an increase in pressure to open the vanes. I am toying with the idea of using a vacuum diaphragm from a cruise control unit. I imagine that engine vacuum could be used to control the diaphragm. Hopefully, a happy relationship to the three controls could be found. I am wondering if vacuum from the vacuum pump is linear with rpm.
I imagine that I would still want to use a wastegate system from the EGR port similar to the one you used in order to control boost spikes.
I imagine that the control system outlined would have the effect of keeping the vanes open when not demanding more boost (in order to lower EGT's, and increase engine efficiency), eliminate the hyper boost response at cruising, and be far more likely to fail to a safe(er) mode (vanes open). "
There is no boost dependancy here so the system has no feedback for such things as air temp, altitude or any other factor that a pressure referenced actuator accounts for. Also the 3 springs would have to be so accurately balanced to get the correct boosting characteristics that it would be almost impossible! Remember you would have to include the forces exerted on the turbo vanes by the flowing air in your calculations, which measuring would be no mean feat in itself!! You are talking carburettor style accuracy!
The rpm dependant method may yeild VERY strange boost characteristics, as although airflow into the engine may be linear with respect to rpm, air flow accross the compressor is not proportional to these for the boost curve. An engine load referenced spring may work better.
How can a cruise control work? There is no engine vacuum on a diesel?
Also, the problem of the turbo overspeeding if a boost hose blows off is alot less catastrophic than if the vanes stick closed when the boost hoses are still on! Zero p2 pressure means that the compressor wheel has effectively no forces accross it and as such is in no danger. The problem is that the shaft speed may become high enough to cook the bearings or wobble onto the housings at either end. This will NOT happen after 5 secs (at most) before you notice you have popped a hose! In fact last time a hose popped at full boost myself and passenger actually noticed quite quickly, it was apparrent by the huge !!POP!! (and passenger cursing) that nearly made me swerve into oncoming traffic!
FSPgtd:
Have to say that your setup is looking pretty sweet, and such nice fabrication on the exhaust!
I had designed a very similar setup to yourself, although the only extra trick i am using is a ballvalve connected to the throttle arm that allows vaccum from the servo to pass to the -port on a 2 port actuator when the throttle is closed, meaning that when the throttle is closed beyond a certain point, the VNT vanes will be opened, allowing less backpressure at idle. The ball valve has to be completely closed at cruise speeds however as this loses too much driveability.
It also means that when you lift off suddenly or brake the vanes open up. This gives a slight lag, but nothing like a normal turbo, and avoids using the compressor past its surge point. Have you compared your boost at certain rpm’s to the flow map for the GT15? It may be producing boost at the wrong point (but probably not by the sounds of how well it is running!!). There is a map here:
http://not2fast.wryday.com/turbo/maps (very good site) or try
http://www.turbobygarrett.com/turbobygarrett/products/turbochargers.html if you have the part number (vnt has a slightly different comp spec to wastegate GT15).
There are so many adjustment points now though that testing is becoming a heavy task!
There is:
1) Vnt start point adustment (on threaded rod)
2) Vnt finishing point (small allen stop on turbine body) (gt20)
3) Bleed valve to boost + side of actuator to determine vane movement (boost) curve
4) Ball valve active point
5) Thinking of installing another bleed in the vacuum line to see if can use more proportionally!
A better solution to the economy/overboost situation may be to use a tps module (or similar potentiometer) on the throttle arm that will vary voltage output for different throttle openings. This can then be used as an input to a small circuit to control an amal solenoid bleed valve (through PWM) connected to the vacuum side of the 2 port actuator. Maybe a very simple digital boost controller uses this same circuit (in reverse)?
Anyway, very interesting reading guys, will keep following this one!!
Cheers
Greg.
Citroen ZXVNTD (GT20)
