servo controlled wastegate

[kewp]

Does anyone on here run an electronically controlled servo to operate the wastegate? I know they run one on the FSAE car here in 'gong and it made a world of difference to the boost response. I'm planning to design one for my car using a servo from an RC plane/car and a 4agze map sensor to sense boost, and a simple electronically controller to run the show. I was just wondering if anyone here has done it and what the results were.
I might put some pictures and a short write up when i get it finished.

[wilbo666]

I love the idea.

I saw a factory version on something a while back I think!..

I'm not sure if an RC type servo will have enough pressure to stop getting blown open by the exhaust gas pressure tho...think about how tight a pneumatic wastegate actuator holds the gate closed...



Good luck with it, I'll be interested to see how it goes!

Cheers
Wilbo

[Norbie]

From what I understand the off-the-shelf solutions are very expensive, mostly because of the large and powerful servo required. I agree with Wilbo, a RC servo probably won't be up to the task.

[BMWTurbo]

You can buy 30kg/cm servo's on ebay for $20... ie CHEAP but I don't know how long they'd last in the high temp environment of an engine bay...

[LeeRoy]

It could always be isolated or moved while retaining the actuator rod.

[oldcorollas]

other issue would be response speed of servo? i guess wastegate doesn't need to open that fast... and no real need to close that quickly during gear changes? (as long as it closes by time loud pedal is mashed again?)

http://www.servoshop.co.uk/index.php...050&area=Servo
19kg.cm for 90 pounds monies

so thats still about 4-5kg at 4cm... which might be enough to move wastegate properly...

find out what is area of wastegate actuator diaphragm, multiply by pressure.. that is needed load rating?

[wilbo666]

What about a windscreen wiper motor with position feedback...

The only issue that I see is that the failure method would over boost, as it would lock at the last position... (but you should catch that anyway)

Hobby CNC style stepper motor might be another good thing to look at...

Cheers
Wilbo

[LeeRoy]

Is it possible to set the servo so a failure means it opens?

[kewp]

i figured out I will need about 5kgcm of torque available, and a quick check on ebay shows 11kgcm ball bearing servos are about $20, so this servo operating at half its design rating should last for a while. Tomorow i'm gonna hit up one of the lecturers for help designing the controller for it. The problem with designing the controller is thats its very difficult to measure the dynamics of the entire system, which makes it hard to design an appropriate controller that will give the fastest response without overshoot. At this stage i'm looking into a PID controller but i will see what the lecturer thinks.

[kewp]

as for servo failure, it probably depends what has failed, if it strips a geartooth, then it will go slack and allow the wastegate to flap freely and will probably equal zero boost. If the mechanism jams somehow, then it could lead to serious overboosting depending where the servo jams. The factory management (sr20det) will cut out at about 15psi so this should be okay as i run 13psi, and overboost and it will just cut out.

[MWP]

You could also use a PWM controlled solenoid rather than a servo motor.
Itll be much more reliable (less moving parts, resistant to heat, etc, etc), quicker too.

Itll be a little harder to control (no absolute position feedback info like a servo has), but worth the effort to do i think.

[wilbo666]

You could also use a PWM controlled solenoid rather than a servo motor.
Itll be much more reliable (less moving parts, resistant to heat, etc, etc).

Itll be a little harder to control (no absolute position feedback info like a servo has), but worth the effort to do i think.

Hmm,

The drawback I see of that is that you loose the position feedback and also because the solenoid isn't fixed position it will also be effected by the exhaust pressure...I think these would combine to make it much harder to control!

It does have its good points as you say tho!

i figured out I will need about 5kgcm of torque available,

You take a guess at how much torque you will need...

I think the rule of thumb for turbo manifold back pressure is twice the inlet pressure? If you took a guess at say 3 times (worst case), calculate the area of the wastegate flap, you can work out the force to hold the wastegate flap closed...

Cheers
Wilbo

[kewp]

MWP might be on to something there, because you don't really need position feedback from the servo, the only feedback you really need is the signal from the map sensor. When the controller senses it has reached the correct pressure it will moved the solenoid until the pressure is steady, it doesn't really matter to the controller what phsyical position the wastegate is in, as long as it is holding correct pressure. The disadvantage would be the extra electronics you would need to drive the pwm, although it wouldn't be that difficult.

[MWP]

MWP might be on to something there, because you don't really need position feedback from the servo, the only feedback you really need is the signal from the map sensor.

Yup.
Feed that into a 2D (map vs pwm) or even 3D (map vs rpm/tps vs pwm) tables and you should end up with the required result.

[kewp]

if you had a programmable ecu, you could use mapping (digital), but i'm still running factory ecu and there is a much simpler way of acheiving control using an op-amp configured to perform subtraction of 2 signals. One signal will be from the map sensor and the other from a potentiometer. Say for example the map sensor outputs 4.0V at 13psi, you adjust the pot to 4.0V. When the manifold pressure equals 13psi the output of the subtractor is 4.0 - 4.0 = 0 volts, and therefore the input to the PID and pwm controller is zero and the solenoid will stay where it is. If the manifold pressure is different to 13psi, the map voltage will be different and the output of the subtractor will be a voltage which in turn will cause the PID controller and PWM controller to activate and therefore move the solenoid. I don't know if i explained this very well, but the outcome will be that the solenoid will always attempt to hold the pressure at whatever the input pot is set to.

The PID controller, for those who are wondering, is a network of resistors and capacitors that will increase the response of the system without allowing overshoot. For a system like this it will be impossible to know the correct values of the resistors, so you build it with trimpots so you can adjust the gain so you get no overshoot.

obviously with a programmable ecu, you could do all this in software, but i don't have one and for the sake of actually putting something i've learned at uni into practice, i'm gonna do it the analogue way.