Page 4 of 8

Posted: Wed Mar 03, 2004 2:37 pm
by NemesisEJ22t
About the Toyotas, i'm still not too sure how the MAP sensor functions on Subarus, but the Toyota only uses it for ECU boost control and to make the factory boost gauge move around. People do actually unplug the MAP sensor on their cars to remove fuel cut without adverse consequences. The bad thing about that is the way the fuel cut is implemented. Whenever the ECU senses too much boost, it cuts fuel, and then goes into limp mode, giving only 2-3 psi until the engine is shut off and restarted. Not too much fun when your on the track. As for your fuel cut setup, i don't know too much about electronics yet, but it seems that you have really done your homework. I am going to try to build it this coming week to get rid of the "hitting a brick wall" feeling that i get when my MBC decides to spike. Thanks for the info.

Posted: Wed Mar 03, 2004 5:14 pm
by vrg3
On our cars, the MAP sensor is used at least for the same things as on the Toyota. There may be more, though, and we simply don't yet know for sure.

That really sucks that it goes into limp mode on the Toyota.

Posted: Thu May 06, 2004 11:35 pm
by sideforce
Hi All,

Is there any of these avaulble for purchase? could do with one now...

Posted: Thu May 06, 2004 11:53 pm
by vrg3
I can make you one, sideforce. PM me with an offer.

Posted: Tue Jul 06, 2004 1:43 pm
by free5ty1e
Was thinking about building the phase II FCD into a mixed-signal microcontroller, since I have plenty of them and work with them all the time. But the routing needed for this design makes that implementation a bit tough. For now I'm going to build this as is, but will be working on the microcontroller in the background as I may have more features to add.

One question - why not just power the chip with the 5v? I don't see a reason to use the 12v supply if the max. reading needed to deal with is under 5v... I mean, isnt the voltage drop at most .5-.7v? That seems just about perfect...

Posted: Tue Jul 06, 2004 3:19 pm
by vrg3
You certainly could do this with a mixed-signal microcontroller, and you'd get the ability to do other interesting stuff too.

How much does a simple microcontroller with at least one ADC and DAC cost?

On the third page of this thread I briefly mention the power supply thing... the problem is that the LMx24's common mode input voltage range and output voltage swing don't reach high enough. If you used a rail-to-rail op amp instead, a single +5v supply would be fine. The TLV2374 from TI would be a good choice.

Posted: Tue Jul 06, 2004 9:44 pm
by free5ty1e
Well I was toying with the idea, but the virtual routing is difficult to do what your design already does properly. The PSoC chips get down to under $1.00 in large quantities, and aren't too much more in single... the latest family of microcontrollers from Cypress Microsystems have these prefixes and these features (check for prices, they carry all of these):

CY8C22xxx = 4 digital blocks, 1 analog column
CY8C24xxx = 4 digital blocks, 2 analog columns
CY8C27xxx = 8 digital blocks, 4 analog columns

each digital block can be anything from serial communications to a PWM, each analog column consists of 1 continuous time block (PGA, comparator, etc) and 2 switched capacitor blocks (ADC, DAC, filters, etc). The xxx varies depending on chip packaging, flash memory size, and # of pins. Check out the details on - only downside is if you don't already have a burner for these chips, starting out can be a bit expensive. But if you've got an idea you'd like to see on one of these I can probably toss it together for you and send a burned chip for experimentation.

Anyway... your current FCD works like a friggin charm.

One question though... has anyone ever tried just using a 4.3V zener diode and a resistor? Or is there too much current going through that signal line? Because that would be a much simpler way of clamping that voltage...

Posted: Tue Jul 06, 2004 10:15 pm
by vrg3
I don't have much experience with microcontrollers, but those Cypress systems sound pretty cool... I don't understand the terms "continuous time block" and "switched capacitor block" -- is it just that one is a regular ol' combinatorial circuit and the other is a circuit with a clock and clocked devices? How much do burners typically run? Are the chips reprogrammable?

Some people do bypass fuel cuts by simply reverse-biasing a Zener diode between the signal line and ground, the idea being that the Zener wouldn't do anything until the signal voltage exceeded its breakdown voltage, at which point it would pull the signal towards ground. I'm not sure where you were thinking of putting the resistor...

I don't really like that approach for a couple reasons -- one is that it loads the pressure sensor far beyond its specs; it's actually a very high impedance signal. Another is that the breakdown voltage of a zener diode at low currents is not predictable; you may have to try a few different diodes (even if they're all the same part number) before you get one that clamps how you want it to. A more minor issue is that the clamping threshold will be absolute rather than relative to the ECU's 5-volt reference which it uses to interpret sensor signals.

A few extra cheap parts for a high-input-impedance low-output-impedance predictable active clamp with a threshold dependent on the ECU's rail voltage seemed sensibler to me.

But people have successfully used Zeners on lots of different cars... If you do a search I think it's been brought up here too.

Posted: Wed Jul 07, 2004 12:06 am
by free5ty1e
Yes, you have it about right, switched capacitor blocks are clocked switching between variable capacitors within the silicon for functions such as ADC, DAC, and filters... continuous time blocks are analog circuitry that does not depend on clock signals, such as an amplifier or a comparator. The chips are programmable in both intel-based assembly and C (if you spring for the C compiler, which I did... the assembly was very slow going). The burners run between $200-$300, maybe a little less now, its been a while since I've purchased one. The chips are EEPROM with over 10,000 reburns possible. May even be more on the newer family, this is just what I remember from the old CY8C25xxx/CY8C26xxx series. The PSoC Designer IDE is freely available from the website, you can download it and look around at how things are laid out... its very nice. Combinatorial logic such as AND, OR, NOT, XOR, etc is now built into the digital interconnects - that is, any signal going to a pin can pretty much be put through one or more of these functions. Used to be digital blocks. C code can do pretty much anything too. Damn good processors. 93.5khz - 24MHz CPU clocks.

I was a little wary of trying the zener because I figured you'd thought of it already before coming up with your circuit... didnt know how much current was actually in use on that signal line, and didnt know how stable of a clamp that would be. Tested your FCD out on the work bench, that is a damn stable clamp.

Posted: Wed Jul 07, 2004 2:21 am
by vrg3
Wow, those chips seem like they're smarter than I am... Intimidating ;). It does sound cool, although the price of entry is high. I have done a little stuff with PICs and AVRs and it seems like they're much less sophisticated but also much cheaper. Intel-style assembly sounds really cool, since it's what I grew up on. :)

Like I said, people do successfully use Zeners, and I knew about that before I started. In fact, it at least used to be the case that it was impossible to read anything about RS-T projects without also reading about the Zener diode necessary to prevent the ECU from seeing boost. But this circuit is about as simple as they come, which is often a good thing -- you could probably get away with removing the output buffer if you wanted, but that's about it.

Posted: Wed Jul 07, 2004 6:24 am
by evolutionmovement
This is the part where I feel stupid.


Posted: Wed Jul 07, 2004 1:58 pm
by vrg3
:?: I don't get it.

Posted: Wed Jul 07, 2004 4:33 pm
by free5ty1e
I think he's saying he's not a microcontroller-programming kinda guy. If thats the case then I probably just sounded like I was talking in Greek. ...Which I know none of.

Anyway I highly recommend this phase-II FCD to any and all turbo owners with boost control, even if you just want to unplug your wastegate line and throw caution to the wind :D

Posted: Sun Nov 14, 2004 1:51 am
by J-MoNeY
How hard would it be to make this a plug and play?

Posted: Sun Nov 14, 2004 2:18 am
by vrg3
You basically need to cannibalize an ECU and a wiring harness for the connectors... It's kind of annoying to securely connect the wires to the receptacle, but it can be done with some patience. Then you just build a short extension harness for the ECU connector and splice the FCD into it:

Posted: Wed Apr 27, 2005 5:51 am
by Tleg93
I actually just got around to reading this thread. Night shift gives me a lot of time to think about electronics. Pretty intersting stuff. Has anyone tried to use these programmable chips to make a home-brew electronic boost controller? Now that would be something I'd like to try. :)

Earlier in the thread free5style was talking about using C and some cypress chips to do some stuff with with this circuit. V said some other interesting things could be done. Could someone elaborate on this a little bit. The signal from the pressure sensor is analog so are you saying to convert it with these IC's and then do some kind of signal processing in addition to clamping the voltage? I've done some assembly programming and C. I'd like to think about this some more.

Posted: Wed Apr 27, 2005 2:56 pm
by vrg3
Some of the simplest forms of electronics boost control are implemented with certain aftermarket ECUs -- they just pulse a solenoid with a duty cycle looked up in a 1-dimensional table indexed by engine speed. People have good success with this.

A little more sophisticated is the use of a 2-D table that's also indexed by throttle position. In fact, aside from the slight complexity of tuning this table and the fact that it doesn't incorporate feedback, this is actually a very powerful form of boost control. Subaru has used it for many years, actually.

The former would be pretty easy to implement with analog circuitry, if you wanted; you'd be better off implementing the latter with a little microcontroller. I think Chris was talking of integrating it into the LegaCU... right, Chris?

You could also try to design a closed loop PID controller that uses the pressure sensor for feedback. Without other intelligence you do run the risk of overspeeding the turbo at part throttle, though. And you need a pressure sensor that reads up to the range you're dealing with, which may mean the stock one is insufficient. The advantage is that you just tune the three gain constants, and specify the desired pressure. It would in principle be able to adapt around changes in the system's behavior and stuff.

Posted: Wed Apr 27, 2005 3:04 pm
by mikec
Hey Vikash,

I came across this the other day. Its an old thread, but the FCD they talk about at the beginning of the thread is one that ramps the signal above a threshold, effectively raising the fuel cut point, instead of eliminating it completely. Very similar to the one HKS sells, except that theirs has 2 LEDs to let you know when you've reached the 2 points set.

Granted, what you've said about our stock pressure sensors still holds. Does our stock sensor reach 5V around 14psi? Because if not, then I'm wondering if a sensor off a WRX would work?

I really need to read this thread from start to finish....

Posted: Wed Apr 27, 2005 3:12 pm
by vrg3
Yeah, our sensor tops out somewhere around 15 psi of boost...

If you want higher range, a 3-bar GM MAP sensor is probably the way to go, since its transfer function is well known and it's easy to find.

Then you'd have to first condition the signal to match that of our stock sensor, and then apply the ramping thing.

Note that I disagree with the way HKS ramps the signal; it'd be better to pass the signal through unchanged until you start getting up to higher boost.

Posted: Wed Apr 27, 2005 4:08 pm
by free5ty1e
V - yes, the electronic boost control using our stock controllers will be a part of the LegaCU. I finally got it set back up on the new workbench (still in Orlando, at a friend's house) and should make some more progress on it this week.

Maybe one of the first tasks I'll tackle is boost control...

Posted: Thu Oct 06, 2005 12:44 am
by andys2
Why not just unplug the MAP line from the exchange solenoid?

Posted: Thu Oct 06, 2005 12:50 am
by andys2
vrg3 wrote:
You could also try to design a closed loop PID controller that uses the pressure sensor for feedback. Without other intelligence you do run the risk of overspeeding the turbo at part throttle, though. And you need a pressure sensor that reads up to the range you're dealing with, which may mean the stock one is insufficient. The advantage is that you just tune the three gain constants, and specify the desired pressure. It would in principle be able to adapt around changes in the system's behavior and stuff.
I think maybr you don't need the D in this setpoint regulator type application. you only need a PI controller. the good news is that the entire PI controller can be implemented with a resistor(s) and a cap and an op amp.

Posted: Thu Oct 06, 2005 12:58 am
by andys2
and the easier way to do it electronically is to use a zerner diode clamp and a series resistor. It's non adjustable but it's the bomb reliability wise - and WAY easier to construct.

R (10000 ohm or so)
to ecu ---------/\/\/\/\-------- to sensor
4.5V zener (or whatever)

But i still don't know why we just don't disconnect the line entirely.

Posted: Fri Nov 25, 2005 7:54 am
vrg3 i think my fcd or should i say your fcd broke, its got to be -20 upper where i am in canada and the snow is comeing down hard. its time to play in the snow with my subaru, today my car was hiting boost cut, i used the lm324 becuse i cant find the lm224, is that why it broke becuse of the drop in tempature and becuse i used the lm324 chip right?

i must make a new one right away i cant have this, where can i get this lm224?

Posted: Fri Nov 25, 2005 5:10 pm
by magicmike
I dont think its "broken". Resistors change their values in extream temp conditions. maybe Vikash could think of a fix for this.