VGT Discussion on turbos

[alansupra94]

Quoting DynastyLCD:
alan, an iac is a stepper motor. basically, as you adjust the applied voltage, its physically moving in or out. same with the Garrett VGT vane adjustment. theres a stepper motor that moves forward or backwards to let the plate that holds the vanes either push them open or closed. its pretty simple.



You think it would be impossible to retrofit a new stepper motor in there?

 

[DynastyLCD]

im sure you could create something to work the vane actuator like a stepper motor, programming it to work off inputs like EBP would be where itd get tricky if you ask me.

 

[desant78]

I did some work with stepper motors, servos ( which i think could work here too) and controllers of each

for those interested in stepper motors, here is a pretty dry (boring) link about the basics of them to understand how to control them intro to stepper motors

I won't lie, I am very lost with this discussion of how to control this turbo, but I am quite versed in c++ and various other types of controllers. It seems like in order to write this, one must clearly define the switches for wanting boost and amount ( like 0% tps instantly opens waste gate) and what actually increases boost (MAP reading?)

so there is my two cents, sorry if it is not very helpful. the coding doesn't seem hard, but understanding what the program needs to accomplish and methods, is what i think is the hardest part. I'll give any input to anyone trying to program this if it is wanted but no promises that anything i know is very useful /ubbthreads/images/graemlins/rofl.gif

 

[alansupra94]

Well upon research regarding VGT turbos and my AEM EMS, I find that I can actually control it quite well.

AEM EMS offers 2D and 3D maps for boost controller solenoid. So I can actually change the map based on many, many factors, including TPS and rpm. Looks like I will be getting a wastegate actuator and a boost solenoid very soon. /ubbthreads/images/graemlins/smile.gif

 

[desant78]

Any progress on anyone end? Anyone have access to any wiring diagram or schematic of the vgt?

 

[curtis]

Ok the Dodge computer and the little ecu board in the controller on the turbo runs off of a can bus. Don't know anything about it but a buddy told me that megasquirt has a new set-up to control a can bus. Haven't researched it since I haven't started on my car but did just find this.


click


I plan on taking pictures of my set-up and downloading and emailing the maps etc to them and see if what I have already have it or if it can be upgraded. I have a profec A as well as an AVCR for the car but hopefully I can just upgrade the megasquirt for the boost controller function as well as VGT control and run 8 injectors. I don't like wiring that much so I have a friend that wire tucked his Honda that will be running all the wires on the car. When we get to that point I'm pulling my harness and getting him to delete everything not needed and I still need to pick up an non eprom ecu to control the stuff megasquirt doesn't. Since I have a tablet PC for the car I'll just build a mount for it and let it be on all the time and use it for everything. Thats the plan anyway just hope everything works as it should.

 

[desant78]

ok lets assume you find a controller, it might be beneficial to all to create a master list of I/O's, or what variables the VGT will be a function of. This way, once you get the controller up and running, you'll know exact how to program this. Also having a master list will allow people (like me) to pursue other solutions. Anyone up for that? I know someone mentioned an AEM EMS solution for this problem.

I understand a lot will have to be tested and proven, but a rough idea what variables we need to measure would be very beneficial. I've read the previous post in this thread...people mention EGT, TPS, Load, etc.

I am starting to pursue a possible standalone PID controller, which I think is the easiest solution IMO. It will tap into the ecu to get this master list of variables we want to use. It makes a lot of sense to me, and in school used a cheap controller (like 60 bucks) to do a few PID projects that controllered stepper motors and servos. Writing the program with the variables and getting the needed inputs does not seem hard.

The challenge lies within communicating with the stepper motor. I am attempting to by-pass the whole CAN-bus route, since it's a pretty propitiatory thing.

oh and a 5 minute research lead to this discovery....using the same controller to make a standalone PID,

PID Standalone controoler

and yes i'm trying to get into contact with this gentlemen now.

 

I just thought i'd share a couple of thoughts with everyone. I have personally considered using a linear actuator because it would work well with my current setup, however being able to use the stock electronics would be neat and probably better. The only downfall to using the stock electronics is the extra size associated with it (and finding/building a controller). If I recall correctly from when I first started installing this turbo, with the stock electronic controller the only way to fit in on the modified t3 manifold (and not move the radiator) was if I rotated the turbo 180* so that the exhaust pointed out towards the passenger side. Even if I built a new manifold (still considering) retaining the stock electronics would mean moving or changing the radiator because of its size. With my current setup the exhaust housing is against the engine block and only about 1/2" from touching the radiator. With the turbo being so close the stock radiator tends to get a little warm when I am not moving (even with two electric fans on the other side), but once I am moving temps are normal. Driving with it is fun, my 510 cc injectors are maxed out at about 12 psi so I need some 1000-1200 cc injectors most likely, and boost comes on quick even though I have a few known exhaust leaks that are working against me.

On another note if memory serves right I think the stock gear driven motor had a max torque output of about 7 N*M (about 5 ft-lb)...so I dont think a new motor/lin. actuator would need a high load rating, but having one that can react quickly which be a huge benefit. I had found this information out a while ago b/c delphi sells the exact same controller for other things but I think it is set up to use PWM instead of CAN (same hardware just different wires hooked up).Also with my style setup (mechanic linkage/levers) I can mount a linear actuator almost anywhere, the benefit to this is moving it away from the hot exhaust. but since I am all for building a custom manifold, moving the radiator and using the stock electronic controller if someone can find a way to cheaply manipulate it :-D

Do note that a company has already made and used a VGT controller, the company is Fleece Performance, from a little research it looks like they are going release a standalone unit able to be tuned by usb for $625.....

I used to have a bunch of files on the holset back when I was looking into modifying the stock controller but I did find an alleged compressor map /ubbthreads/images/graemlins/smile.gif

 

I was going through my old files and found some other diamgrams that someone might find useful.

 

[desant78]

Thanks for all the support information! I am working some tweaks in a basic outline for the code..a simple one to start. Currently looking for a turbo. Lets keep this discussion going....no one has really made a controller that was not a 2-D map. The more information the better!

 

[cheekychimp]

Desant, are you simply looking for a turbo for test purposes or are you intending to install one on your car?

 

[desant78]

If you see the video links, he uses a potentiometer to give the position feedback. I am planning on fabricating the housing so I'm looking to purchase one. On the side, if I am looking up possible ways to communicate via the current controller, if that does seem to show some promise, I could use one for testing purposes.

 

[boostx]

I had pick a very low mileage on eBay some time ago for under $200 ship..

 

[desant78]

been looking..but finding pricey ones and I kinda feel odd buying one on ebay unless it's a good deal.

 

[boostx]

That sucks because when i got mine the price range was between 175 to $350 depending on the shape.

 

[desant78]

yea i know....all the post i find are within that price are from 2009-10...and those people already sold them. but I'll find one I just gotta be patient.

 

[boostx]

It must be sick to hit full boost by 3k RPM.

 

[BrandonEchols]

I'll keep a watch out locally.
There's a guy in town who works at the Dodge dealership and gets them every so often and sells them fairly cheap.

 

[cheekychimp]

Quoting Boostx:
It must be sick to hit full boost by 3k RPM.



Yeah but what you aren't actually letting on to the uninitiated here John is that it isn't just the notion of being able to spin a big turbo quickly (people have been doing that with various degrees of success for years by launching at insane rpms, using anti lag and nitrous etc) it is being able to effectively change the A/R of the turbo at will. Even if you can make 700 hp at 3000 rpms it becomes near impossible to hook up the tyres. Drivers end up experimenting with launch after launch after launch doing everything from breaking drivetrain parts to bogging and everything in between (wheel hop and spinning through three gears etc.). The beauty of getting this to work is that you can effectively launch on an A/R of a 16G getting perfect traction and nailing a 1.6 sec 60 ft, but then as soon as you are off the line gradually increase it so that you progress through the equivalent of an HX35 in 2nd, an FP 3065 in 3rd and a GT42R in 4th.

 

[boostx]

You are 100% right Paul but i have all always get rush to see a high HP NA car that can light up the tires of idle.