Micro controlled VNT project

[regcheeseman]

Do you have a bird? you have way too much time on your hands...

   - No bird......... no time what with having a wife, two young kids, house and full time job. Most of the PCB prototype work I do in my lunchtimes.

These two PCBs are around the same size as an EDC15 ECU. Maybe a bit over engineered for the simple job it is going to do?

The image is not actual size, I'm dealing with large packages (not SMD devices) on a single sided board. The lower board is the display/boost gauge

The circuit drives the display / shift light / rev counter AND controls the VNT servo.

[tomnik]

am I right considering the controller as 2 parts:
PID and pwm generator?
If so a LED dimmer could be the the pwm part.
I ordered such a thing that outputs a pwm 140 Hz, 12 V, adjustable with a poti.
With that I should be able to move either the stepper motor or via solenoid the vac.
Now the PID only has to "operate" the poti.
Is there any kind of PID available that could do this?
Or does it make no sense to split the complete unit like that?
I am not an electronic, just want to get my Benz running after the mechanical installation of a VTG.

Tom

[regcheeseman]

The 28X2 micro processor has two PWM outputs built in it also has several analogue IN lines and configurable digital inputs/outputs

a quad op amp looks after the pressure input from boost sensor
throttle position is measured via pot

vnt is controlled via servo motor driven by motor driver chip - positional feedback is via pot

An engine speed signal is taken from pump tranducer and via a freq/voltage convertor LM2917 to give an analogue value for revs.

[tomnik]

so what is needed to complete the controller?
The code only?
Are the mentioned hard ware components proved and tested and available?
I don't want to sound silly, just try to learn where you are and how I can give input.
(not personally me but buddies who know more about electronics than i do)

Tom

[regcheeseman]

I appreciate the offer of help, but electronics is my profession,  The hardware is all prototyped and bench tested, the code written and working. There were a few problems with input signal levels and op-amps but these are now all resolved.

I've improved the design to incorporate new features and increase processing speed.

The only thing that's holding up the completion of the project is life in general - if you could send your buddies round to re-plumb my central heating, finish the house extension, cook the meals, wash-up etc it would be much appreciated

[regcheeseman]

I discovered that I needed a buffer stage on an output to drive revcounter, a spare buffer was available on the board but to use it required re-working half of the design, 6 hours later I have yet another print and it's a bit smaller. Main board is 110mm x 50mm



Hopefully it will be produced fairly quick and I'll have a working version

[regcheeseman]

My project is just about complete, all sensors are returning required values, the servo driver is responding as expected = I just need to implement a new map and would like to bounce my ideas about on here....

I have a number of variables sampled

Tp = Throttle position  0=accelerator not pressed 255=pedal mashed into carpet
Ba = Boost actual 0=no boost 255=max boost  (circuit currently built such that max boost is 30 psi)
Revs =Engine speed 0 = 0rpm 255 = 7300rpm (again this max depends on circuit configuration)

I need to manipulate these variable to give a desired vane position Vp

If I simplify the setup considerably and make each variable a binary value I get this


Tp   Ba   Revs   Vp
0   0   0   0
0   0   1   0
0   1   0   0
0   1   1   0
1   0   0   1
1   0   1   1   
1   1   0   0
1   1   1   0

Therefore Vp = Tp AND (not)Ba

It would appear that the revs value is redundant, however I think the over-simplification has caused this and I see the actual value of VP being a function of Revs value - providing the basic Tp AND (not)Ba is adhered to.

If the variables are taken back to 0-255 analogue variable, (not)Ba would be (255-Ba) then the basic equation would be

Vp = Tp - (255-Ba)   All values over 255 = 255 and any negatives would be 0

If I bring revs back in, we get

Vp = (Tp - (255-Ba)) - revs


And tuning factor Tf could be introduced to shift the performance up or down

Vp = Tf x ((Tp - (255-Ba)) - revs)

Does this sound feasible? any ideas?

[gnavs]

I love how I don't understand 90% of the jargon being used in this thread, but nonetheless I approve!  Good luck bringing this to fruition.

[R.O.R-2.0]

I love how I don't understand 90% of the jargon being used in this thread, but nonetheless I approve!  Good luck bringing this to fruition.

x2 on not understanding half this thread..

but i know i want one when they are done..

[regcheeseman]

At the best of times I've only got 80% comprehension of the thread.

Been working with the MAP, it seems that the oversimplification of the map may well be right - convertuing it from 2D to 3D when you add in a REVs variable, actually has very little influence on the map profile.

The replacement of the REV variable with a fixed Tuning factor, serves pretty much the same purpose - and is much quicker through the processor.

So the only thing driven by rev signal is the shift light, I'm tempted to rip it out of the final circuit board.   

I'd build a seperate revcounter driver with shift light - that would hook straight up to any four cylinder petrol revcounter.

[regcheeseman]

FINISHED! 

Finally got it sorted and fully operational.

For those that dont know it's a circuit devised to control my VNT turbo using a servo motor, basically the servo motor opens and shuts the vanes on the VNT turbo, affecting it's gas flow geometry

The linear pot in the forground is in effect the accelerator position and the syringe is replicating boost pressure  - peaking at 28 psi

The revcounter has the boost gauge incorporated as a bank of LEDS, and also a status light (partially obscured by white wire)


So what happens in the video? I reach over and switch the power on, the servo unit cycles shut then open to check for stuck vanes, when this is complete, the status light changes to green and the unit is in it's run mode.

Any input from the accelerator is monitored and responded with respect to existing boost pressure, the nature of this response is tailored by a tuning factor entered in the program code.

[rumbling_caddy]

   Congratutation!    Must be so exciting to see the fruit of your labor working as envisioned.  Will you be needing guinea pigs?

[Thezorn]

Wow thats awsome. I want to see a vidoe of that thing in action. You must be pretty excited, its been a long time in the running.

[745 turbogreasel]

That is  a thing of beauty.

I have a devils advocate moment-
Don't the OEMs limit boost off idle/low vehicle speed to protect the motor from crushing force at low oil pressure, and save the teeth on 1st gear?

And some of my related interests-
Braking mode, high boost off throttle
Maybe this can be accomplished by interrupting the TPS circuit.
Launch mode, spool up before you hit the gas
Maybe  I'm over the edge...

[regcheeseman]

OH thanks for that, just when I thought I'd got it finished

Not sure what you mean by braking mode off throttle - what response do you need in what instance?

Launch mode would be easy enough, the circuit had a pushbutton circuit, you could set a launch mode that cancels when the TPS or REVS input gives a certain value or you could just interrupt the TPS signal as you suggest to force the turbo to spool.

Don't the OEMs limit boost off idle/low vehicle speed to protect the motor from crushing force at low oil pressure, and save the teeth on 1st gear?

I'm hoping the map will allow for this to some extent, and my inbuilt mechanical sensibility will stop me going stupid in first gear....hopefully.