New boost control suggestion
A while back I stumbled across the boost control system used in the Bigstuff engine management system. I found it to be very different compared to other systems as it gave the end user very good control.
The basic setup is a fairly standard PID control loop which operate a valve on the bottom chamber of the wastegate. It has a two dimentional table where you can map boost over the rpm range. Nothing new here.
The alternative boost setup is not new either, a VERY primitive table where you can set a fixed duty for the boost control solenoid for different rpms. This is actually pretty common on race cars as it's simple and precitable.
What is new is that you can use both at the same time! This allow you to use the duty table for the top chamber of the wastegate and the PID controlled boost target table to control the bottom chamber.
The above cover most options but it's badly thought through. A few changes allow a more flexible setup. By using a three dimentional table for the PWM we can make sure that we have full pressure on the top chamber when the boost pressure is below a certain limit.
We could also have a separate limit below which the top chamber solenoid is open but the table offer more control and let us set everything in one place.
The three dimensional table is also beneficial for the boost target table if we use TPS as the Y axis. This allow us to modulate target boost with the last few percent of throttle opening. The 3d table offer significant benefits over a rpm_boost_target table and a multiplicative tps_boost_target table.
I suggest tables with 3 boost/tps sites and 5 rpm sites.
With large external wastegates the physical size of the membrane is often a bit to small compared to the wastegate valve itself. For example an 80mm membrane for a 50mm valve or something like that. That will force a relatively weak spring to be used which will make it hard for the wastegate to stay shut when the exhaust pressure press on the valve. There is a few guidelines when choosing a spring for an high boost car, your spring should give you a minimum pressure that is not more then 0.5-1bar below the target pressure. The higher the backpressure the higher the 'base boost' will have to be. By having a separate boost solenoid controlling the upper chamber you can increase the control range by using a weaker spring and using boost pressure to keep the valve closed.
Jorgen, I think you've got a good idea here but rather than adding these settings I feel that the PID loop should be used to "learn" the proper duty cycle and then have it stored for next time. So we go through a building boost stage where the valve is held, then switch to the assigned duty cycle and use the PID loop to adjust it. With the proper PID settings I've actually found that our existing strategy works fairly well.
Hi Jorgen, I don't understand why it's needed a second solenoid. The way I set up the plumbing lately in the Lancias I tune is by interconnecting the inner side of the membrane in the actuator (originaly breathing to air) with the duct of the solenoid previously breathing to air too. In this configuration the solenoid is able to close the wastegate firmly at any given backpressure even using a soft preload in the actuator (which provides a smoother control). With this configuration wastegate creep is canceled, however the feed forward duty cicle values must be heavily reduced compared to the std configuration since a 50% now will equal to a 100% before. Above 50% the actuator will close with added load on top of the spring preload of the actuator. I haven't played with overboost with vems yet, but would like it to be able to control boost in this configuration, I played with an aftermarket ecu recently no being able to control this.
The way the Marelli ecu works is though a 3D table DC%(Rpm,Tps) of feed forward duty cicle and a boost setpoint vector Boost(Rpm), when pressure is above the setpoint the PID steps the duty cicle back. It is amazing the speed at how the oem ecu controls the situation.
My wishes for changes in current open/closedloop boost control:
Openloop (OL) is good as it is but closed loop (CL) needs changes to TPS-scaling. Problem is that CL uses "boost target" not as target but max boost limit, and TPS-scales that limit down wich causes pwm to hump between 0 and reference value when running lower TPS. My wish is that TPS would scale down boost DC taken from the boost table and left boost target as max limit. In this way boost control would be predictable and stable without much humping.
Simple but useful
Different boost-target depending on
- analog input (multiple b lines) that can be
- or oil temperature if you like
- high EGT can lower boost-target (eg. -100kPa / +100C) above settable threshold (eg. 900C or whatever is configured)
->Can these features be implemented/when ?