ITT: We discuss electronic wastegate ideas
 

Pretty simple idea. PWM controlled linear actuator connected to the wastegate flapper. I can figure out the electronics for it but some discussion on the forces required, how strong/quick of an actuator is needed would be fun. I have some free stuff at work and want to play with some ideas.

Discuss.

Interesting, as I've long pondered a similar concept, using an electrically operated bypass valve in a compount (turbo-supercharger) application, to fully bypass the supercharger once the turbo's boost threshold has been achieved.

Not that I ever plan to build such a thing...


I've heard tell that some WWII aircraft engines had turbochargers with electrically operated wastegates, allowing the pilot to dial in different pressure ratios from the cockpit. A quick search also shows some more recent work in the field.

https://patentimages.storage.googlea...S4656834-2.png


In theory, computing how much force you need is easy. Measure the pressure inside the turbine housing under full load with WG closed, and multiply that against the exposed surface area of the wastegate valve. You just need to drill a hole in your turbine to mount the probe and you'll be good to go. :bigtu:




This press release talks about an electrically operated wastegate for OEM applications: KSPG AG: Electric wastegate actuator ready for volume production


http://cdn.pressebox.de/f/426135bd2f...913.attachment

Wikipedia says electronic control in aircraft was tossed because there was a requirement of separating the electronic systems from the mechanical systems or something.

Wikipedia also says that Gary Oldman is a giraffe, and describes parkour as "the physical act of sucking nuts in an urban setting."

There's long been a superstitious prejudice against electronics on the part of airplane designers and engineers. As late as the 1990s, light, single-engine airplanes were still capable of flying with no functioning electrical system of any kind; the ignitions were magneto-operated and the fuel delivery was a mechanical carburetor with manual mixture control. Once the engine was started, it'd keep running until it ran out of gas even if the alternator and battery were both completely disabled.


That said, I'm a little curious about the ability of your linear actuators to tolerate the heat that will be experienced by a device located an inch or two away from a part of the engine which sometimes glows cherry-red.

Me too :)

Good thing for over boost.

I spent half an hour discussing this with a buddy. I'm a firm believer that it's possible. My solution to the heat problem was a remote mount with a cable linkage, although I'm not convinced can't design a solenoid to take it.

For what it's worth Bosch makes one, but it looked if you have to ask expensive.

A solenoid?

It's not especially easy to PWM-drive a solenoid against a heavy load. A servo would be far better suited to such a task, which I assume is what aidandj meant when he talked about linear actuators. (Many servos, such as those used on remote-controlled cars and airplanes, use a PWM control signal and adjust their position to the duty cycle of the PWM signal. This is a very different thing from just directly PWMing a solenoid.)



Serious question: what would the advantage be of such an arrangement (cable linkage, servo / solenoid, etc) as compared to an air-powered diaphragm controlled by a solenoid-operated air valve?

The latter is pretty much EBC with a standard wg, right? If so, the only real advantage I see is being able to run very low/safety wired open equivalent levels of boost that you couldn't typically buy a spring for (or pre-assembled IWG).

He said PWM control not I. I'd implement the controller on the actuator, and pass it desired boost via the programming, can bus, etc.

The benefit is a one component wastegate control with no reliance on pressure in the system. Done right your have extremely consistent max pressure, with the ability to change whenever you want.

I would add no extra controller or anything. It would be a linear actuator controlled by PWM. So theoretically MS3 boost control would work with no code changes or anything else. I too thought about remote mounting it. But a cable wouldn't offer the same forward and back motion that a linear actuator provides. I'm hoping that I can find a free linear actuator or 2 from my schools robotic lab.

Advantages would be complete control over boost, all the way down to 0 psi. I also think that it would be easier to control.

Right now its all theoretical, and something that I think I could "easily" implement and test.

Fair enough, just different goals. My intention would be a more saleable design that's not stuck with one reasonably obscure control.

It should be a fairly easy force balance problem, with the load being the wastegate pressure times the flapper area and any spring force you include in the system. My guess is roughly 60lbs is all you'd need.

I suspect you'll need A controller of some kind to interpret the boost own signal into actual positions, although I don't know how MS manages boost controllers.

The MS outputs a PWM signal (intended to operate an air solenoid), starting from a known base duty cycle, and then adjusts it in a closed-loop mechanism to seek a target manifold pressure.

This is done OEM on hyundia turbo cars.
They do it with a simple servo and have some issues with getting the wg to close fully.
I do plan on taking this route, and will likely do a linear servo with a sprung push/pull rod on the actuator arm. Basically you drive the servo to its hard stop and rely on the spring to keep pressure on the wg flapper.

I think I posted this already but I built a circuit that sits between the ECU and the wastegate solenoid that implements a fast local feedback loop to regulate wastegate can pressure. It uses a pressure sensor to monitor and target the can pressure.

So the ECU output instead of going directly to the solenoid requests a can pressure, a proxy for flapper door position. The local loop will adjust solenoid duty cycle to achieve that. The resulting response of the flapper position to a change in ECU output duty cycle is much faster. The overall boost control became faster, more consistent, and with less over and undershoot.

The best way to make this work for a reasonable amount of money is probably to use a 2-way clapper noid if such a thing exists. Or just a monster normal clapper with only a spring holding the gate shut against the exhaust pressure. I dont know what clapper noids are actually called in technical terms come to think of it, we just call them that because they make a clapping sound when they actuate. Its literally just a big dumb coil that moves a steel rod and the clapping noise is the flange on the end of the steel rod hitting the body of the noid at the end of its stroke.

I'm not seeing any sort of worm drive linear actuator being able to move fast enough for a gas engine with an appropriately sized ball bearing turbo.

And one of the stepper linear actuators that work more intelligently than that clapper noid are stupid expensive for the amount of force, speed, stroke we need.

High powered metal geared hobby servos could do this easily. /thread.

Also to the guy who suggested that ms3s boost control via a pwm table is obscure.. Please exit thread.

+1 In the end unless you happen to find a linear actuator that can interpret less than 50% duty cycle pwm as one direction and more as the other you're going to need control logic anyhow.

I like the hobby servo idea in concept. Can you get them making like 60 in-lbs?

I never said pwm based boost control is obscure with megasquirts, I said the entire controller (the megasquirt) is. While the Miata community loves them, if I spent appreciable time and money designing a product I'd want to sell it to the Evo, Sti, 911 etc crowds, none of whom are big MS users.

My browsing history turned up this little guy:
https://www.servocity.com/html/hs-79...l#.VR3Ov-H09sR

Titanium gears, 0.15 sec/60° speed, 333oz-in torque, cheap...
My plan was to buy 2 and see how it held up.

Post here please.

https://www.servocity.com/html/how_d...l#.VR3cJLpViko

Sounds perfect. The PWM signal tells the servo where to be. So just as megasquirt does it 100% is full boost, 0% is closed wastegate. Mount that servo and put an arm on the end of it.

leboeuf what was your timeline for trying it out?

It doesn't matter how obscure MS is, you didn't even say that and I didn't imply you did what you did say was the control method of configure able pwm tables is obscure, infact it is standard and expected on all standalones these days.

Yes you can easily buy servos that powerful that are also super fast.

I see that a couple of other folks have already posted links to ServoCity. I was going to post this link, which has externally gear-reduced servos rated at up to 1,700 in-lbs: https://www.servocity.com/html/top_m...l#.VR3g-PnF98F

They are, of course, relatively slow. Not sure if "slow" in the world of robotics is the same as "slow" in the world of engine management. I've never actually logged the operation of an EBC to see how rapidly it makes large changes in commanded position.

60C max operating temperature is a little bit weaksauce.

For comparison, I found this spec on a commercial wastegate servo
http://www.stoneridge.com/wp-content...eet_Rotary.pdf

2.7Nm torque is 382 oz-in

Ahh, good to know.

I'm the servo torque matter, I'd been considering the waste gate valve being under pressure, but in reality it's just under whatever pressure the exhaust exerts, correct? If so it should be real low force. I'd probably try one of the servos without gear reduction and see if it has the guts. Return spring plays into that force of course.

And be sure to keep them way away from the turbo, those plastic cases will turn to soup if you don't.

Return spring....for what?

Unless you can link to the waste gate with a rigid linkage you'll need a return spring of some kind to close the waste gate. Whether it's built in or not in not sure (i know motion control stuff, not turbos). I've always assumed you'd end up needing a cable to move the actuator away from the turbo.

Servo at 0% duty cycle holds the wastegate shut. Servo at 100% duty cycle holds the wastegate open.

Awesome! So 382 oz per inch is a good baseline. Temp spec would need to be taken into account.

that can be a great idea!

not exactly
it's simple, but not that way
every 20 ms you can send a pulse from 1 to 2 ms
1 ms = 0°
1.5 ms = 90° etc
the servo can rotate from 0 to 180°, but near the limit is a little flappering, even for the expensive units

I don't know how to write ms codes, but I think it can be done.

I've got some 16 kg (35 lbs) servo with metal gears, they are very powerful: if I hold a position I can't move it with my hands!

sadly I have a rev ms2 unit: I don't want to add the hardware for the servo...

you need a 5v power supply for the servo up to 1.5 A (an indipendent one is better) and the logic level pulse for the servo signal (before the mosfet or whatever you are using to drive the ebc solenoid)

please, someone make it!
the heat is not a problem for a pull or two, you are just testing. If it works: heatshield.

The ms already outputs a PWM signal. Couldn't we use that?

Why not a solenoid powered butterfly valve instead of a wastegate?

You guys are of course aware of Hella/Garrett electronic actuators for VNT turbos, yes?



You will notice its mounted directly to the turbo thus safe for underhood mounting. It comes in PWM and CAN flavors for OEM Ford and Mercadies/VAG vehicles in Europe.

More discussion, part numbers, electrical data, etc can be found here: Garrett 712120 VNT Hella 6NW 008 412 actuator.

They can be found used on ebay UK for around $50~$100

deleted

I'll look into that. But diesels run colder, and people have had issues with the VGT actuators on gas motors.

Interesting! I build an electronic wastegate on a BMW a year or so ago.

https://lh5.googleusercontent.com/-v...106_185244.jpg

https://lh3.googleusercontent.com/-n...106_191157.jpg

https://lh3.googleusercontent.com/-8...g%2Bscreen.jpg

https://lh3.googleusercontent.com/-b...106_234841.jpg

It was a 12V DC Servo with a potentiometer for feedback. The ecu I had at the time was able to control it via a 2D table (using MAP) or 3D one (MAP and RPM). I used a 2D table, basically opening the wastegate fully during low loads, then keeping the wastegate closed untill just below the target boost, and then tapering the valve open. This worked very well. Turbo spoolup was improved, Economy was a tiny bit improved, but my favorite part was that I was able to run any boost level I wanted quickly by adjusting a few parameters. I could turn the boost down all the way to zero, or run as much boost as I wanted to.

Eventually one of the butterfly valves I used failed, and the entire thing was binned. I think I'll build something similar again, but using a much more durable wastegate valve. The oem option above looks promising!

I'm not sure what motor it uses or what the feedback signal is, but if it's just a 12V motor with 0-5V feedback maybe this is a suitable and affordable controller:

STPS - controller for servo of exhaust valve. - IgniTech P?elou?

Why not use an electronically controlled 4-port solenoid. I've read about this a lot on the Honda forums. I think connecting a servo to a glowing turbo would be a bad idea unless you could find some reliable way to stop it getting cooked.

Here is a diagram that Jaydee made for a twin supercharged K24.
https://www.miataturbo.net/attachmen...ine=1429259403.
He's using a MAC 4-port solenoid to control a 38mm Tial wastegate which looks to be very successful. He's running boost by gear and can dial in max boost via his steering wheel. I think you need to run a lighter spring in the wastegate so that you get a good range of adjustable PSI.

A setup like this is what I want to run when I upgrade to a Rotrex.

How am I just finding out now that this is a thing...just 5 years late btw.
I need to hang out on this forum more often.

Bump...

Before I go buy a 4 port MAC and a 2 port actuator, has anyone gotten anywhere on this?

I've got some of the 1/4 scale super cheap hobby servos and would be willing to give it a go. I'd probably just start with a separate uC with a pressure sensor to save the trouble of trying to get the MS to do it to prove if works well or not.

If I were to try it I would use some stupid long linkage. And then a uC. I wouldn't use a pressure sensor though. Just translate DC from the MS to output position. Then boost control should still work fine.

Did I post this already? https://www.amazon.com/dp/B00CBKBQBE/

https://cimg9.ibsrv.net/gimg/www.mia...384f8e842b.jpg
As far as I can tell that's a BorgWarner EFR/KKK turbocharger recirculating diverter/BOV 3 bolt flange. If I buy one of these things and a pigtail will one of you guys figure out how to make it work? ;)

So dumping boost instead of bypassing exhaust? Sounds like a good way to overspin the turbo. Send me a turbo speed sensor with it and I'll try.

Actually, that's a good point. Probably need to disregard ;)

someone on our MT FB page (that doesnt understand electronics whatsoever) is trying this...

But why... I mean I guess he demonstrated that the actuator isn't stuck? /Sigh

I should message him and ask him to send to to me, and in return I'll tell him how it works, but not return it...

You haven't seen much from tech salvager have you. There's a reason he's banned here.

We'll that's a little better... but I mean, wow.

wait is that tech? all his posts make so much sense now!

is it?

See username

Seriously, that was the first thing I noticed.

I attempted this on my 1.6 turbo build. Used the mahle electronic wastegate shown above.

Mahle 03F145725G
Found on any of the 1.2 TSI VAAG engines, engine codes CBZA and CBZB, since about 2012

If you pull one from the junkyard, dont make the same mistake as i did and make sure you get one with the connector. The wiring colours seem to differ between engines but the connector is a 6 pin connector with only 5 pins populated.

https://cimg8.ibsrv.net/gimg/www.mia...94779ba41d.png
2 of the pins connect internally directly to the motor. The other 3 are 5v, 0v, signal out. Its a hall sensor inside so nothing too difficult to deal with there.

It needs an H-bridge driver capable of "some" amps to drive the motor. I never got my setup making boost with the actuator attached but i cant imagine it takes more than about 5 amps at 12v to hold down the wastegate flapper under boost

Its worth noting that the linkage arm coming out of the actuator is not constrained linearly so it can be bolted directly to the wastegate arm if mounted in the right plane. The total travel of the linkage is (from memory) about 10mm.

https://cimg5.ibsrv.net/gimg/www.mia...69652eec75.jpg

Dont make the intermediate arm between the linkage and wastegate like i did in that pic ^^^. The arm is already designed to follow the arc of the wastegate arm so making extra linkages makes it just tweak about rather than actually operate the wastegate.

Unfortunately this car met its end on its first outing on track running a normal wategate actuator and i haven't turned my attention to it since. So i have zero data on how well it works, in practice, let alone any data on spool ect.

Not that that data would have been any use, because its a 1.6 anyway. This build was in progress way before i had done any reading here, but im now dropping my rods-only VVT lump in this weekend so id like to get this ball rolling again soon.

Back from the Dead.
 

Hey Forum,

I know its an old thread but i though it would be worth re-visiting. Has anybody implemented a working electric actuator-based wastegate control as discussed in this thread? A lot of European cars (BMW,MB,VW) are now using electric-servo driven wastegates. Some of those can be had for $70. Does anybody have a working system in their car and could chime in on MS3 control implementation using the PWM table?