The Top DSM Community on the Web

For 1990-1999 Mitsubishi Eclipse, Eagle Talon, Plymouth Laser, and Galant VR-4 Owners. Log in to remove most ads.

Please Support RTM Racing
Please Support STM Tuned

HE351VE Controller Project

This site may earn a commission from merchant
affiliate links, including eBay, Amazon, and others.

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

casuprock

10+ Year Contributor
320
3
Apr 13, 2010
Exeter, New_Hampshire
Starting a writeup about this project to document the work in a new thread dedicated to the controller. The HE351VE controller is designed to position the vanes and keep the compressor in an optimal condition. The HE351VE uses a BLDC motor in the OEM control. I will be using a linear actuator (FA-PO-150-12-2") for the first version of this controller.

The objective is to create something that can be interfaced externally to anything with 5V logic (like an AEM ecu) or to any device capable of communicating through I2C. Digital expansion means I can collect information from any type of sensor necessary to make the system better. The controller also has analog inputs for a variable reluctance (shaft speed) sensor and pressure sensor. This thread will be edited as the controller is developed since I'm sure there will be revisions moving forward. Schematics will be posted after the circuits are tested.

Inputs
- 0-5V Analog potentiometer position input
- 0-5V Differential VR Sensor input ~0 to 300000 RPM
- 0-5V Analog pressure sensor input
- Vin (12-14.5V)

Outputs
- 5V analog supply
- 0-Vin (0-14.5V) PWM motor output

Digital IO
- I2C compatible
- 5V logic

Processor
- ATMEGA48

Basic controller concept

Here is a picture of the controller, before PCB trace routing.
You must be logged in to view this image or video.


The basic control idea is already defined by Holset:
You must be logged in to view this image or video.
 
Subscribed. I have one of these just chilling here, waiting for its day.
 
So would this be a standalone controller that outputs a value that can be logged via a 0-5V signal?

Right now, I have digital IO, with no analog output since that's sufficient for my needs. If there is some interest, I will add an analog 0-5V output as well.
 
Ah, now I see what you mean. Never mind.

I've wanted to build one of these for so long, but the more I looked into it, the more I realized I was in over my head.

Are you making this fit inside where the electronics package originally was?

This version will be built to demonstrate that I can do the control, in the future I might replace the Holset PCB with a custom board of the same shape. For now, the most important goal is to prove the control can work.

I added a 0-5V analog out to the design.
 
Added an extra external 0-5V analog input. (In addition to the potentiometer 0-5V input)
 
Hey man looks good! I am glad to see other people throwing their hat into the ring on this. I think VGT is an amazing thing that the dsm community can really benefit from! I have been working on a controller myself but with a different philosophy. I have gone under the assumption that if you dynamically change the back pressure of the exhaust you are going to change the VE of your engine.

When you change the cross sectional area of the turbine you are changing the back pressure of the exhaust. This is going to change your volumetric efficiency of the cylinders. If you don't have the same VE every time you are at the same load factor/RPM your AFR will be effected. From the design notes that I have seen of yours, I would almost think that you would need your control box to not only control the cross sectional area but also have the ECU account in the variation of the back pressure to get the correct AFR. An example of this is when people change their turbo in their car drastically (ie smaller to larger) you have to re-tune the car. In the original design the Diesel ECU is what determined the vane position.

What we are talking about here is getting the peak efficiency of turbine at all times. Allowing the turbine to extract the most amount of energy, without negatively effecting engine performance. Small area for small amounts of exhaust, large area for large amounts of exhaust. Therefore in my design I am utilizing an approach where I vary the vain position based on RPM. Since RPM is one of my variables on my VE table, and one that I can measure with out my ECU outputting it, I can make it my dependent variable and make sure my cross sectional area is consistent, and therefore my back pressure consistent.

Now with saying that my design has flaws. Like no matter what my load factor is, my cross sectional area of the turbo will always be the same. So I will have to tune each and every RPM point that I want to have. Each point will be unique and every time I make a drastic change to the engine I will have to re-tune each point. Also if I tune the point for fast spool at low load factors, I might choke the engine at high load factors, or vice versa.


Final thing I want to say is that I love your design process. It looks like you are really putting a lot of effort into this and will love to see your results. I am an engineer also myself, and love it when people try things that are different. Even if everyone else has and issue with it (because its not a 16g).
I have taken an assumption that back pressure is going to have a huge impact on the VE. With your design when implemented you may find that that is not the case. So please continue, let me know what you think of my design.

Thanks!
 
Hey man looks good! ...

I have an AEM EMS so I can use a wideband O2 for feedback to control the AFR by injector duty cycle & meth injection. Your idea to control AFR based on back pressure is an interesting thought though and does have merit.

The funny thing about back pressure (caused by a turbo) is this:
Higher back pressure-> lower VE-> higher pressure drop across exhaust turbine->higher turbo shaft speed -> more compressor flow & higher boost-> higher VE.

So an initially lower VE results in a higher VE once the turbo gets up to speed. It's a form of positive feedback, the energy from the exhaust pressure drop is fed back into engine by increased intake pressure. What this means is the VE is lower while the turbo isn't spooled (i.e. lag) and MUCH higher after the shaft speed is up; the reason a turbo works.

The job of a VTG is to get the turbo up to speed faster by making it behave like a smaller turbo (i.e. higher exhaust pressure drop) then transitioning the exhaust pressure drop to a lower point as the engine flow rate increases. It all comes down to shaft speed; the VTG has done it's job once the shaft speed is up as quickly as possible.
 
I have an AEM EMS so I can use a wideband O2 for feedback to control the AFR by injector duty cycle & meth injection. Your idea to control AFR based on back pressure is an interesting thought though and does have merit.

The funny thing about back pressure (caused by a turbo) is this:
Higher back pressure-> lower VE-> higher pressure drop across exhaust turbine->higher turbo shaft speed -> more compressor flow & higher boost-> higher VE.

So an initially lower VE results in a higher VE once the turbo gets up to speed. It's a form of positive feedback, the energy from the exhaust pressure drop is fed back into engine by increased intake pressure. What this means is the VE is lower while the turbo isn't spooled (i.e. lag) and MUCH higher after the shaft speed is up; the reason a turbo works.

The job of a VTG is to get the turbo up to speed faster by making it behave like a smaller turbo (i.e. higher exhaust pressure drop) then transitioning the exhaust pressure drop to a lower point as the engine flow rate increases. It all comes down to shaft speed; the VTG has done it's job once the shaft speed is up as quickly as possible.


I don't have any experience in AEM EMS, so forgive me if I make any assumptions. I have only a limited experience with tuning and only with ECMLink and megaquirt (project of my friend).

I should have explained more clearly the situation I was thinking of. When your engine is operating in closed loop you should be fine. Your ECU is reading your Wide band and adjusting pulse width of the injectors and in your case, meth injection. So any variance in VE would be taken care of by the ECU.

But when you are not in steady state/closed loop you have to make an open loop look up table, where the ECU has set values of VE that estimate the amount of oxygen that is in your cylinders. If you change the back pressure when you are in open loop, your burn will not be optimal.
Also in your example above where high back pressure low VE works its way through the turbo to high intake pressure/high VE, there is a lag in that process. So even when you do get to the correct shaft speed, if your engine is not in closed loop you will get the same effect of having a mis-tuned open loop map.


Really what I am saying is that there is nothing wrong with your controller, its the way that the ECU's available to us operate. Now if you were able to make a Open loop look up table with 3 variables (RPM, load factor, and backpressure) ... :hmm:
 
I don't have any experience in AEM EMS, so forgive me if I make any assumptions. I have only a limited experience with tuning and only with ECMLink and megaquirt (project of my friend)...

You're absolutely right. If I were planning on running the stock ECU with this VTG and no wide band feedback, I would have to build a look up table. The table could also be based on throttle position, intake manifold press, and RPM if back pressure were not easily measured.

I chose to "suffer" through the pains of setting up this AEM instead so that I can use a big MAF and not have to worry about changes in weather or altitude affecting AFR like the speed density folk (I bet some SD guys are going to hijack this thread because of that comment :p ).

But... the controller I'm making for the VTG is a stand-alone system. It can interface with 5V digital or analog which will hopefully make it useful for people who run any type of engine management, either open loop or closed loop AFR.
 
I've been wondering how long it would take for someone to step up to this challenge ever since aero tried using a mechanical only system with a WGA positioning the vanes on his. As long as the adjustment is quick enough, this should prove exponentially better.

Kudos to you sir.
 
Added a surface mount ambient pressure sensor (Freescale Semiconductor MPL115A2T1), for use in reading barometric pressure when calculating pressure ratio. This makes two pressure sensors; external mani pressure and the barometer located on the controller.

Working on calculating shaft speed vs pressure ratio from HX40 7 blade compressor wheel (same compressor map as HE351VE as far as I can tell, but there are misinformation and rumors about Holset maps). Shaft speed has good sensitivity as a function of PR. The exhaust vanes control shaft speed by changing pressure over the exhaust turbine. Plotting out the compressor PR vs shaft speed parameter gives a nice polynomial fit which is well behaved. Tells the most efficient path through the map, so I can adjust the vanes to seek this path. At least it's a good base to tune from.

You must be logged in to view this image or video.
 
Hey man looks good! I am glad to see other people throwing their hat into the ring on this. I think VGT is an amazing thing that the dsm community can really benefit from! I have been working on a controller myself but with a different philosophy. I have gone under the assumption that if you dynamically change the back pressure of the exhaust you are going to change the VE of your engine.

When you change the cross sectional area of the turbine you are changing the back pressure of the exhaust. This is going to change your volumetric efficiency of the cylinders. If you don't have the same VE every time you are at the same load factor/RPM your AFR will be effected. From the design notes that I have seen of yours, I would almost think that you would need your control box to not only control the cross sectional area but also have the ECU account in the variation of the back pressure to get the correct AFR. An example of this is when people change their turbo in their car drastically (ie smaller to larger) you have to re-tune the car. In the original design the Diesel ECU is what determined the vane position.

What we are talking about here is getting the peak efficiency of turbine at all times. Allowing the turbine to extract the most amount of energy, without negatively effecting engine performance. Small area for small amounts of exhaust, large area for large amounts of exhaust. Therefore in my design I am utilizing an approach where I vary the vain position based on RPM. Since RPM is one of my variables on my VE table, and one that I can measure with out my ECU outputting it, I can make it my dependent variable and make sure my cross sectional area is consistent, and therefore my back pressure consistent.

Now with saying that my design has flaws. Like no matter what my load factor is, my cross sectional area of the turbo will always be the same. So I will have to tune each and every RPM point that I want to have. Each point will be unique and every time I make a drastic change to the engine I will have to re-tune each point. Also if I tune the point for fast spool at low load factors, I might choke the engine at high load factors, or vice versa.


Final thing I want to say is that I love your design process. It looks like you are really putting a lot of effort into this and will love to see your results. I am an engineer also myself, and love it when people try things that are different. Even if everyone else has and issue with it (because its not a 16g).
I have taken an assumption that back pressure is going to have a huge impact on the VE. With your design when implemented you may find that that is not the case. So please continue, let me know what you think of my design.

Thanks!

If you use a maf based air metering method, changes in ve will have no effect on afr's. Speed density uses a lookup table based on ve, ait, and map. With a maf no calculations involving ve are needed to hit the target afr's in the fuel table. The air is actually being measured, no calculations are needed involving ve.
 
A company already has a controller to put HE351 turbo on a 98-05 Cummins Dodge. Not sure how much would be similar between the two applications...

HE351 Turbo Controller

Not much similarity unfortunately... I was hoping to use this controller but turns out it's designed for the specific application. Not too useful for installation on our cars, even with an AEM other standalone. You'd have to do all the work to communicate with it, plus buy it in the first place.

I am making a standalone controller that is designed to work on a generic VTG application with direct sensor input (not relying on an ECU or specific communication protocol). What I want to end up with is a device useful for any tuner with a VTG turbo and an electric linear actuator... not tied to a specific ECU or make/model.

Hey, I was thinking... it would be neat to put a Bluetooth transmitter on this design so I can transmit data directly to a laptop for logging and tuning. Roving Networks makes some good Bluetooth serial modules I've worked with in the past. What do you guys think about that?
 
Work is coming along on this controller!

Recent update:
Moved to stepper motor from brushed motor with position feedback. This eliminates the need for analog feedback from a potentiometer and also makes the unit more 'robust'. There are some high torque steppers in NEMA packages that will work great. Since I switched my design from a brushed motor to a stepper, I deleted the H bridge on my design and replaced it with a Texas Instruments DRV8412DDWR which is smaller and better suited to drive these bipolar steppers.


Things slowing completion: My work is too busy :rolleyes: and I need to get a top mount tubular manifold $$$ fabricated for the HE351VE. Maybe there's someone out there who can fabricate a top mount manifold in exchange for a VTG controller...

I'm considering making this into a kit once the design has been tested. Thinking about what will be included in a kit-
Motor, machined aluminum mounting bracket, modified VTG actuator shaft (either splined or set screw mounting to drive motor), controller board, rebuilt HE351VE.
What price range will the market bare for a Holset turbo with VTG?
Is this something a lot of people want, or just a few people?

I talked with Thomas Dorris from ECM Tuning. We discussed the lofty goal of connecting my VTG controller to Link. He was interested in the prospect of a joint project, but both of us are very busy right now. Just out of curiosity:

If you don't already have Link, would you buy Link because it supported an electronically controlled VTG turbo (i.e. could tune VTG maps, bidirectionally communicate with ECU, etc.)?
 
seeing how good used vgt holset can be had for around $250, I doubt many people would like to pay more for a controller than the actual turbo.

I really hope this will be available by next summer

as far as buying link, personally that would be a no. for some the 98 99 flashable ecu is a cheaper option
 
Support Vendors who Support the DSM Community
Boosted Fabrication ECM Tuning ExtremePSI Fuel Injector Clinic Innovation Products Jacks Transmissions JNZ Tuning Kiggly Racing Morrison Fabrications MyMitsubishiStore.com RixRacing RockAuto RTM Racing STM Tuned

Latest posts

Build Thread Updates

Vendor Updates

Latest Classifieds

Back
Top