Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[nock]

since i ordered all the parts the other day, and i got nothing but time right now, i guess this is my project.

i started screwing around with this the other night. i was able to track and decode all 30 steps per rev. i had it printing position data, but serial.print takes way to long for it to keep track. so i just made it turn on a led at position (Z=0).



the vgt takes about 9 motor revolutions from full closed to full open, the motor can decode 30 steps/rev (10 pole armature, multiplied by three halsensors) so that would be 270 steps. im guessing the whole control word in the original delphi controller is an 8-bit (256) integer.

i've got all the brushless motor driver crap coming from digikey, it should be here this week. now that i've have it positioning, most of the PIDservo code has been written by others. if nothing crazy happens (like me getting a job) it should come together pretty fast.

when its done i'll roll the code into a nice vgt library for the arduino. the library should make it even easier for others to create there own vgt control systems. for right now i plan on making it read PWM from the megasquirt boost controller.

[danz]

you are smart?     .



hm

[Robb]

you are smart?     .



hm

Didnt see that coming. 


However, it was cool until I saw "megasquirt"

[Joseph Davis]

you are smart?     .



hm

Didnt see that coming. 


However, it was cool until I saw "megasquirt"

MS2 is supposed to be alright, but I know what you mean... overexcited MS fags can ruin anything.

[nock]

MS2 is supposed to be alright, but I know what you mean... overexcited MS fags can ruin anything.

megasquirt big cheap-cheap smok-em-lota DB37 plug...

im using megasquirt because its what i know, its just easier. it has a boost controller with a PWM table. so tps vertically, rpm horizontally, and the table itself is filled with pwm%.

all the megasquirt is really doing is giving the turbo a signal that is tps vs engine rpm. this could be done by any ecu with that capability. its not even acting as a boost controller its just a way to tune how the turbo responds to the engine.

boost control itself could be handled by turbo shaft rpm. because compressor pressure ratio is a function of compressor rpm, regulating rpm should regulate the pressure, and keep the turbo from overspeed and exploding. so in the table you'll have tps vs rpm, the cells will represent turbo shaft rpm. so 100% is max rpm (boost) and 0% is no boost (basically whatever rpm that compressor p/p equals one). a nice benefit of doing things this way is that the vgt will seek the largest A/R possible to maintain the desired shaft rpm, this way you will always have the biggest turbine side possible. fuck yea  ;D

or you could just make it control vgt position in % and use an external waste gate. i think i'll try this first because its safer and you could play around with how it responds between shifts.

or you could have it control boost from a map sensor on the arduino board itself. this way is nice to because you can still play with shift points and it simple and cheap, no external waste gate, and you can adjust boost with nothing more than a setscrew.

or you could just hook it up to pwm on an ordinary boost controller. im not sure how this would work, and you would still need tps somewhere to make sure it dosnt try to boost at idle.

also, as if this broke dick ghetto turbo kit wasnt cheap enough already, since the boost controller table has TPS on it, you could tell it to go to max A/R when the TPS is below say 25% and run without a blow-off valve   i have no idea if this would actually work or not, it may end up being what some people consider an acceptable amount of surge, it may not work at all, i dont know. such a system would also require a compressed air tank and solenoid valve hooked to the tps, without this the driver would become disoriented by the lack of blow off sounds and probably crash. 

there is no reason it could not be made to work with any boost controller or ecu, so long as you could put two sensors in a table with PWM output, you could also go stand alone and write an interface that used its own TPS, MAP, and RPM sensors. or like i think will happen, because arduino is so easy to use, people will write there own control structures that work better then mine anyway.

[Gold DA9]

just wow. i wish i never changed my major from engineering

[SloS13]

I've got an arduino sitting right here next to me.  Gotta get off my ass and mess with it some more.  P.S. that video had me glued to my seat.  I was waiting for the exorcist girl to flash on the screen. 

A VGT library would be pimp. 

For non-standalone users, one could get the arduino to take in voltage from the MAP and some type of device that converts the RPM into a voltage (or maybe the arduino itself could read the RPM viaa signal from the ECU?) and you could hook up an LCD and some buttons and put some menus in there to set your values up or read them a .CSV from an SD card....or something.    That'd be cool....so get on that STAT.

[nock]

i think it could be used stand alone with just a low RPM cutoff, say 3000rpm or something you know it wont choke at. a map would give you more control over how it reacts, but having something that just throws the VGT wide open below RPM (X) would work to. all you would really need then is a way to make a single curve for the TPS.

right now, i thinking of this thing like an engine with two throttle bodies, one has regular atmosphere on it, the other has boost on it. the TPS curve just tells the boosted throttle body (VGT) how to move in relation to what the driver wants.

got my digikey order in yesterday. im having some trouble making the interrupts work, i think the arduino interrupt handler itself may be eating up to many clock cycles. i was going to test this out but i thought my o-scope took a dump on me, turns out it was just a bad power cord  . i'll hook up the generator and o-scope and play around with timing more tomorrow.


since you have an arduino, if you want to play around with it here is the code i used it the video. i hope it dosnt show how much i think in assembler all the time.

Code: [Select]

   const int hallAPin = 2;
   const int hallBPin = 3;
   const int hallCPin = 4;
   const int LedPin = 13;
   
   int Z = 0; //posistion
   int N = 1; //hall word
   int lastN = 1; //last hall word
   int HallA = 0;
   int HallB = 0;
   int HallC = 0;

    void setup(){
        pinMode (hallAPin, INPUT);
        pinMode (hallBPin, INPUT);
        pinMode (hallCPin, INPUT);
        pinMode (LedPin, OUTPUT);
        Serial.begin (9600);   
 
    }
    void loop(){
     
      HallA = digitalRead(hallAPin); //read hall sensors A, B, and C
      HallB = digitalRead(hallBPin);
      HallC = digitalRead(hallCPin);
     
      if ((HallA == HIGH) && (HallB == LOW) && (HallC == HIGH)) {N = 1;} else {lastN = N;} //hall word = 1
          if ((lastN == 6) && (N == 1)) {Z = Z + 1;}
          if ((lastN == 2) && (N == 1)) {Z = Z - 1;}
 
      if ((HallA == HIGH) && (HallB == LOW) && (HallC == LOW)) {N = 2;} else {lastN = N;} //hall word = 2
          if ((lastN == 1) && (N == 2)) {Z = Z + 1;}
          if ((lastN == 3) && (N == 2)) {Z = Z - 1;}
 
      if ((HallA == HIGH) && (HallB == HIGH) && (HallC == LOW)) {N = 3;} else {lastN = N;} //hall word = 3
          if ((lastN == 2) && (N == 3)) {Z = Z + 1;}
          if ((lastN == 4) && (N == 3)) {Z = Z - 1;}
 
      if ((HallA == LOW) && (HallB == HIGH) && (HallC == LOW)) {N = 4;} else {lastN = N;} //hall word = 4
          if ((lastN == 3) && (N == 4)) {Z = Z + 1;}
          if ((lastN == 5) && (N == 4)) {Z = Z - 1;}
 
      if ((HallA == LOW) && (HallB == HIGH) && (HallC == HIGH)) {N = 5;} else {lastN = N;} //hall word = 5
          if ((lastN == 4) && (N == 5)) {Z = Z + 1;}
          if ((lastN == 6) && (N == 5)) {Z = Z - 1;}
 
      if ((HallA == LOW) && (HallB == LOW) && (HallC == HIGH)) {N = 6;} else {lastN = N;} //hall word = 6
          if ((lastN == 5) && (N == 6)) {Z = Z + 1;}
          if ((lastN == 1) && (N == 6)) {Z = Z - 1;}
 
     if (Z == 0) {digitalWrite (LedPin, HIGH);} else {digitalWrite (LedPin, LOW);} //turn on LedPin when Z = 0
       
        //Serial.println (Z); //uncomment for debug
        //Serial.println (N); //uncomment for debug
        //Serial.println (lastN); //uncomment for debug
    }

some time soon i should have the interrupt routine figured out and i can start working on something besides tracking position.

[HiProfile]

MS2 is supposed to be alright, but I know what you mean... overexcited MS fags can ruin anything.

boost control itself could be handled by turbo shaft rpm.

Please don't take away fully from what you've done here and assume a gasoline engine can do that. Even with that large turbo on a D16, you will still get run-a-way boost with an >20cm housing fully open. And since it's a ~65 lb/min compressor designed with high boost in mind, it will make a lot more than you think. If you really want to assume that, at least put a 35mm on whatever you're doing.


Anyways, it would be nice to see if it could take a trigger input (switch to full 12v or ground) and open the VGT at a pre-determined rate. That would really open your market if you decide to sell them. A simple hobbs switch keeping it closed til half your boost would be tits for everyone else.

[nock]

i have thought allot about how this would all work, the only thing i can come up with is, i dont know what the fuck im doing with regard to engine controls. thats why im using arduino so that anyone can change it to do what they need it to do. my main goal here is to make the hardware and software (a C library) that will alow anyone to move the VGT around and read turbo RPM. i would also like to write some control code that would alow anyone to install the thing and run it easily, what form that system would take is the big question here.

if you think about it as the WG opens the VGT will just close to compensate for the drop in turbo RPM. the WG would have to be set to a pressure that is above the max turbo RPM (TmaxRPM) if the VGT is not wide open before the WG opens, the whole system may oscillate.

im sure it can control boost on its own, how well is the question. the frist time its tested we will just watch the VGT, if its opens all the way, we'll know it needs a waste gate.

[onlyflash944]

all this shit is cool, and way over my head


how would you integrate something like what you are doing with crome or ectune?

[92CXyD]

all this shit is cool, and way over my head


how would you integrate something like what you are doing with crome or ectune?

+1 I'm interested in this. :yes:

[nock]

any ecu that you can put tps and rpm in a table, with pwn output, i dont know if chrome or ecutune can do that. you could also just set a low rpm cutoff and tap into the tps. or just about anything else you want. the injector pulse width for example can be an indicator of how much air is going in the engine. but i think TPS and engine RPM are the two best. intent of the driver verses the engine speed, i think, is the best way to select the desired turbo RPM.

the arduino has USB on it  :yes: so hooking a laptop to make changes would be easy, i just have to write a setup program. the arduino itself has six 10-bit A/D inputs, so you can put any sensor you want on it, EGT, MAT, MAP, exhaust pressure, anything.

what im shooting for is to make an arduino add on board (i think they call them "arduino shields") that plugs into the arduino and has the brushless motor driver, turbo RPM interface, engine RPM interface, and a TPS follower circuit. and then make a nice single page setup utility that can run on any old laptop. at that point it could work with any ecu, im just using megasquirt because it happened to all ready have a tps vs rpm table in it.

[ratcityrex]

chrome

didnt know google had a tuning system out.


(Just a heads up, its crome)

[Aero]

I'm sure it can control boost on its own, how well is the question. the frist time its tested we will just watch the VGT, if its opens all the way, we'll know it needs a waste gate.

Cool project with potential.

It won't control boost though.   It will control how fast it spools, and effect backpressure greatly.  With the VGT opening fully by ~10 psi. on a 2.4L 4g63  boost passes 24psi by 4200 rpm at WOT.  If you look at the spool results of some H1C turbos on D16s on this forum in 21cm housings you can see that it won't work either.

[BoostedSchemes]

I can tell you that with a 100% open holset you will be seeing 30psi well before redline so this CANNOT be used as a boost controller, but can help make a more effective powerband, this is very cool stuff.

[Joseph Davis]

It won't control boost though.   It will control how fast it spools, and effect backpressure greatly.

Prepare to monitor drive pressure and correct vs rpm for correct fuel and ignition control, otherwise there are going to be huge swings in primarily desired fuel, and to a lesser extent ignition timing.  Bigger housings are more efficient, and will require a larger amount of fuel for a given intake manifold pressure.  It's really a system that screams MAF, or asks for a MAF correction.

For simplicity's sake I would start off by being able to dictate housing opening vs intake pressure so that the tune stays approximately consistent, and keep those things linked.  You can do the complex shit later down the line if anyone really cares - I frankly do not see a reason to do so.  If the system is less efficient at small housing openings, so what?  You turn up the boost and open the housing to suit, done.

[Aero]

It won't control boost though.   It will control how fast it spools, and effect backpressure greatly.

Prepare to monitor drive pressure and correct vs rpm for correct fuel and ignition control, otherwise there are going to be huge swings in primarily desired fuel, and to a lesser extent ignition timing.  Bigger housings are more efficient, and will require a larger amount of fuel for a given intake manifold pressure.  It's really a system that screams MAF, or asks for a MAF correction.

For simplicity's sake I would start off by being able to dictate housing opening vs intake pressure so that the tune stays approximately consistent, and keep those things linked.  You can do the complex shit later down the line if anyone really cares - I frankly do not see a reason to do so.  If the system is less efficient at small housing openings, so what?  You turn up the boost and open the housing to suit, done.

That is another reason I have run it with the wastegate actuator control.  Should open approximately the same at a given manifold pressure and kind of work around that issue.  I've had it open early in the boost range to have the tuning error happen when things aren't on the edge also. Beyond ~10psi its fully open so the tune should be the same at all times.  I have asked around about a possible modification to a speed density ECU with some type of a backpressure sensor, somewhat similar to the baro compensation only extended of course.  I would like your input on it.

[nock]

I'm sure it can control boost on its own, how well is the question. the frist time its tested we will just watch the VGT, if its opens all the way, we'll know it needs a waste gate.

Cool project with potential.

It won't control boost though.   It will control how fast it spools, and effect backpressure greatly.  With the VGT opening fully by ~10 psi. on a 2.4L 4g63  boost passes 24psi by 4200 rpm at WOT.  If you look at the spool results of some H1C turbos on D16s on this forum in 21cm housings you can see that it won't work either.

...the master speaks...


i read your thread on the talon, it seems that it will have to have a waste gate (damit) the ultimate go fast cheapness would have been so cool, i guess my cheap  ass can afford a waste gate.

from what i have understood from talking to other people, these things will boost creep while cruising when a mechanical actuator is used. i think it has to be hooked to the TPS somewhere otherwise it will just seek the highest possible manifold pressure weather or not the throttle is open or closed. im sure the system will take off on its own if the turbo is not limited by the TPS.

the change in efficiency is something i hoped to take care of with TPS based fuel enrichment, so a %TPS = %fuel increase, but with the VGT being open all the way that would not work for the whole map. it may have to use something like map switching in order to work properly, like i said, i dont really know what im doing when it comes to controls here. what do you think about map switching?

[98vtec]

its been awhile since something has been completely over my head LOL.

anyone want to dumb this down a little for someone who didnt go to engineering school at the "tender" age of 15   

[Aero]

...the master speaks...


i read your thread on the talon, it seems that it will have to have a waste gate (damit) the ultimate go fast cheapness would have been so cool, i guess my cheap  ass can afford a waste gate.

from what i have understood from talking to other people, these things will boost creep while cruising when a mechanical actuator is used. i think it has to be hooked to the TPS somewhere otherwise it will just seek the highest possible manifold pressure weather or not the throttle is open or closed. im sure the system will take off on its own if the turbo is not limited by the TPS.

the change in efficiency is something i hoped to take care of with TPS based fuel enrichment, so a %TPS = %fuel increase, but with the VGT being open all the way that would not work for the whole map. it may have to use something like map switching in order to work properly, like i said, i dont really know what im doing when it comes to controls here. what do you think about map switching?

Is this TPS based enrichment the TPS enrichment that is already in MS?  Or are you talking about a modification to it?  The TPS based accel enrichment is only for transient fueling on quick changes to throttle position.  Now I think alpha-n is a TPS based fuelling, but IIRC it doesn't work in a boosted application since it doesn't take MAP into consideration.

Map switching would be useful if you had only two operating VGT rack positions, beyond that its sketchy.  Now if you made it where you had your actuator hold a fixed position at a given boost or rpm on one map.  Say, 4000 rpm and 20psi it is always at a 22cm size. 3000rpm and 20psi its at 9cm size.  It needs to be the same each time or you get the VE change that can't be easily tuned around without a mass air system like JD referred to, or some kind of backpressure sensor and compensation if running speed density.
Then switched maps for a fixed mostly open position for economy cruise then I could see that working.  

[nock]

in the firmware you just delete the sec in volt/sec and thats it. v/s is really only used for one thing so i dont think it would be hard, you would have to fatten things up to compensate for not having any accel enrichment. that was what i came up with for the change in bsfc. but, it wont work if the vgt is wide open by 4000RPM.

i cant believe i never considered table switching. its so easy, just pick a big turbo size, say 20cm2, tune it up, then pick a small turbo, say 8cm, tune it up, then switch them when the small turbo starts to run out of hot side. its perfect, you could even use the vtec solenoid to trigger it.

as for the maf i agree, in fact i have read some patents that talk about using turbo RPM for a mass air meter. they are pretty vague, i cant see how it works because turbo RPM has more to do with pressure then flow, but apparently somebody figured it out, mabye turboRPM divided by pressure or something. another crazy idea i had early on was to use turboRPM to run the engine in a maf/map blend arrangement, fuckin weird.

i think just having two VGT positions and using table switching is good for now. i think you could make it progressive with engine RPM, having the A/R change constantly as you go up in RPM. that would at least make sure that each vertical row had the same A/R. i like the idea of having TPS hooked to it as the driver could just make the call on how much boost and when, but as you said, without some kind of maf it dosent look like it could work.

i think table switching would be easier to use, and easy to tune as well.

also, i guess i ordered the last mc33035 from digikey, fagots. this really pissed me off because the only reason i went with that chip was because they had over 2000 on hand. either somebody ordered the other 1998 right after me, or they liquidated them as soon as they sold one. fuck im pissed, the whole fucking mess just went back to the fucking drawing board. FUCK!!!

[Joseph Davis]

It won't control boost though.   It will control how fast it spools, and effect backpressure greatly.

Prepare to monitor drive pressure and correct vs rpm for correct fuel and ignition control, otherwise there are going to be huge swings in primarily desired fuel, and to a lesser extent ignition timing.  Bigger housings are more efficient, and will require a larger amount of fuel for a given intake manifold pressure.  It's really a system that screams MAF, or asks for a MAF correction.

For simplicity's sake I would start off by being able to dictate housing opening vs intake pressure so that the tune stays approximately consistent, and keep those things linked.  You can do the complex shit later down the line if anyone really cares - I frankly do not see a reason to do so.  If the system is less efficient at small housing openings, so what?  You turn up the boost and open the housing to suit, done.

That is another reason I have run it with the wastegate actuator control.  Should open approximately the same at a given manifold pressure and kind of work around that issue.  I've had it open early in the boost range to have the tuning error happen when things aren't on the edge also. Beyond ~10psi its fully open so the tune should be the same at all times.  I have asked around about a possible modification to a speed density ECU with some type of a backpressure sensor, somewhat similar to the baro compensation only extended of course.  I would like your input on it.

I think someone has to build/live with the system in order to really shake any more functional insight out of it. 

nock's concept of table switching is not bad as long as interpolation takes place, look towards the K-series ECU structure with respect to cam angle for a good method of implementing this.

[bigdaddyvtec]

its been awhile since something has been completely over my head LOL.

anyone want to dumb this down a little for someone who didnt go to engineering school at the "tender" age of 15   

NO NO NO.. The rebuttal was the NOT so tender age.... hahahah Seems your comprehension is lacking as well negro

[98vtec]

its been awhile since something has been completely over my head LOL.

anyone want to dumb this down a little for someone who didnt go to engineering school at the "tender" age of 15   

NO NO NO.. The rebuttal was the NOT so tender age.... hahahah Seems your comprehension is lacking as well negro

i wasnt talking about JD's age

[Joseph Davis]

I'm from the trailerpark, nothing tender about me.

[nock]

I'm from the trailerpark, nothing tender about me.

Trailerpark University of Applied Engineering (TPUAE)

tpuae alumni get 10% off at walmart

[92CXyD]

I'm from the trailerpark, nothing tender about me.

Trailerpark University of Applied Engineering (TPUAE)

tpuae alumni get 10% off at walmart

Where can I enroll?   ;D

[Joseph Davis]

I'm from the trailerpark, nothing tender about me.

Trailerpark University of Applied Engineering (TPUAE)

tpuae alumni get 10% off at walmart

Where can I enroll?   ;D

Kick me your address.  As soon as Obama cuts me my first dirty white trash rebate check I'll mail off your Doctorate based on life achievement.

[bigdaddyvtec]

What kind of life achievement award do you get for a blue dash???