Winter Upgrade - Megasquirt

[bobcowan]

Well, I think I have the answer. I drove the car around the neighborhood this afternoon for about 90 minutes. I was using Analyze Live to tune the VE table. In 90 minutes, I only logged a few sync losses. Not bad, I'm thinking.

This appears to be the answer:





I soldered a 1K resister between the power line and the signal line - a pull up resister. Everything I read in the manual and in the instructions says you don't need it. But a few people on the MS forum recommended it more than once. Out of desperation I tried it. It seems to have worked.

This week end I'll dial in the VE table a little more, but it doesn't need to be dead on yet. Then work on the accel settings. After that.... Wasted Spark!

I'm tellin ya, this has been one frustrating episode. I was about this [ ] close to giving up and going back to my Haltech. Thanx to the pro's on the MS forum, I think I can progress.

I'm not selling the Haltech yet, though.

As for the oil pump, I'm going to leave the distributor in place for now. Eventually I'll get around to cutting down an old distributor to make a pump drive.

[vector1]

so what exactly does the resistor do?

[bobcowan]

Originally Posted by vector1

so what exactly does the resistor do?

To be honest, I have no idea what it does. I don't really understand how computers and complicated circuitry really work, it's all just magic to me.

[Luce]

I don't know about this particular part, but most semiconductors, unless wired up into a complex circuit are "single ended" meaning they conduct in one direction and function by turn on and off. If this sensor is of this type, it can only pull a floating signal to ground, not raise it to supply voltage. It's sort of like a lifter. It opens the valve, but needs the spring (pull up resistor) to close the valve. Or the other much more complex solution would be have a second lobe, lifter, pushrod, and rocker to close the valve. I think some hall sensors have the electronics inside to do all of that, but maybe not all.

[bobcowan]

A VR sensor is a magnetic coil that responds to a metal object (like a tooth) passing through it. As the field changes, it creates a analog signal that looks like a sine wave. The ECU receives that analog signal, converts it to a digital square wave, and then amplifies it for the rest of the system to use. it has only a signal and a ground wire, and does not need power.

A hall sensor is the same as a VR, but it has an additional circuit inside that converts the analog signal to digit. It sends a square wave signal directly out of the sensor. The ECU then gets this signal, cleans it up if needed, and then amplifies it for the rest of the ECU. The additional converter circuit inside the sensor needs power, and uses the same negative ground wire. Because of that, it's not supposed to need a pull up resister - it's part of the 12v circuit.

That's what I understand in a nutshell.

Because it's a Hall sensor, there's no need for a pull up resister anywhere in the circuit. Or so I was told. The people who built the ECU and sold me the Hall sensor says it's not needed. But, more than one person on the MS forum recommended it - even a pro-builder and dealer for MS. I figured, what the heck? Give it a try. It costs <$2, and only took a few minutes to install. I really had nothing to lose. I put it in the pigtail of the sensor because that's easily accessible.

The pictures don't show it, but I covered the leads with liquid electrical tape. Bare wires and leads bother me. (Yes, I'm a little OCD)

[vector1]

interesting these electronics. i was having difficulty with my gen 7 setup with noise and also a powerjection, and i had taken as many precautions as my mind would allow.

i believe the resistor is boosting the signal by allowing 12v into the signal wire, but does not allow it to flow back through the positive wire.

another note, if the magnets (if multiple) in the signal wheel are not oriented correctly or all in the same +/- pole orientation, it can confuse the ecu by a mixed signal i believe.

[vector1]

i was going to add, are you using batch injection or sequential or does it apply to your system? i talked with a well rounded efi guy and he said there was not much to be gained by sequential. that was the problem with my gen 7, it would go into trouble mode and change to batch injection so i ditched it. if i try again it will be batch.

[Luce]

Hang on there. A hall sensor is quite different in the way it senses the magnetic field. It does not use a coil of wire, but rather a transistor that is sensitive to magnet flux. What is in question here is the drive circuit inside the hall sensor. If it's being used to drive the LED inside the opto isolator, then there is no need for any pull up resistor. But if it's driving the VR circuit, there is no mechanism to pull the signal high, so a resistor is used. Also, you may not need a pull up resistor, but may need a pull down resistor. It depends on the polarity of the hall circuit.

And no, of course you never see a pull up resistor on a production car. If it's needed, it's located inside the ECU.

[Luce]

Let me add here, I'm not an EFI expert, but have read a lot on the subject, but I was an EE major in college and have been an electronics hobbyist since the 70's, and do have a good grasp of this aspect of the subject.

[Tony Ripepi]

Hooo Raaaaaah,

Success, I am interested in how the self tuning aspects of the MS work as you go through time and then giving you the performance you are loking for under track conditions.

Congrats on your persisitance,

TR

[bobcowan]

Originally Posted by vector1

another note, if the magnets (if multiple) in the signal wheel are not oriented correctly or all in the same +/- pole orientation, it can confuse the ecu by a mixed signal i believe.

I considered using a flying magnet and sensor set up. But they're pretty expensive. And I'd have to modify or replace my balancer to get the proper spacing.

I had a 5" 36-1 tooth trigger wheel made for me. Oddly enough, there does not appear to be an off the shelf trigger wheel for this set up. With so many hobbyists and racers using EFI, that surprised me. Fortunatly, finding some one who could make me one wasn't difficult. I don't remember exactly, but the price to my door step was about $20.

Also, there's no real off the shelf bracket for the hall sensor. I had to make that myself.

Originally Posted by vector1

i was going to add, are you using batch injection or sequential or does it apply to your system?

I'm using batch fired right now. That's what I'v been using for the last 7 years, and it works just fine. Sequential injection is a bit more difficult to install and program. And the only real advantage is a smoother idle with less emisssions. Once you get much above idle, there's no advantage. And some systems will then convert to batch fired, because there just isn't enough time to do it sequential.

CDI's do the same thing. You'll get multiple sparks up to about 2500-3000 rpms' and then it's down to 1 spark just like stock.

The other advantage to full sequential is the ability to use COP ignition. I chose to use wasted spark for a few reasons.

Originally Posted by Luce

And no, of course you never see a pull up resistor on a production car. If it's needed, it's located inside the ECU.

The resister can be anywhere in the circuit. I put it at the pigtail because it was easy. If I was 100% sure it would work, I would probably have put it behind the dash. There is no way I am going anywhere near the ECU main board with a hot iron. That's a recipe for trouble right there.

Like Dirty Harry said, "A man's got to know his limitations".

Originally Posted by Tony Ripepi

Hooo Raaaaaah,

Success, I am interested in how the self tuning aspects of the MS work as you go through time and then giving you the performance you are loking for under track conditions.

Congrats on your persisitance,

TR

I drove it up and down the street today - drove my neighbors nuts I bet. It's just drive and push the button - it was that easy. The computer isn't very bright. So it will sometimes put ridiculous numbers in a cell. So you have to pay attention and know what's happening. You have to drive and tune in a very particular manner to get the results you're after. You can't just drive it around town and expect it do the tuning with no input from you. You need to understand what you're doing and why. It's not rocket science. But it's not as easy as the ads make you think it is.

But just driving around the neighborhood it seems to be driving pretty good. There's a slight bog down low that I have to work on. But it feels pretty smooth.

Tomorrow I'm going to try driving it up and down the highway and do some data logging and tune from there.

It's pretty darned close now. My next step is add the wasted spark ignition system.

[vector1]

another thing to consider

on some of the aftermarket efi setups i believe you can delay how quickly the engine starts through delayed ignition or injector parameters. one or two revolutions of the engine before it fires might be a good idea to give the oil time to circulate if you are routing through lines and some of the oil might have dropped out. i could get my engine to fire pretty quickly and in retrospect don't believe that was a good idea if it had sat for a length of time.

[bobcowan]

One of my concerns with using a coil pack, is what happens if you shred a distributer gear? With a distributer the engine dies immediately, because the rotor stops spinning. The engine quits running before the oil pressure drops. With a crank trigger and a coil pack, the engine will continue to run until the crank completly seizes in the block. That's a bad thing.

So, I routed the power lead of the hall sensor through a summit oil pressure switch. If the oil pressure switch drops below 7psi, it will cut power to the hall sensor. Without a crank signal, the engine will immediatly die.

I don't really know how that will affect starting, especially when hot. a delay when cold is good. A delay when hot, maybe not so good. So, I also installed a bypass pushbutton switch under the dash.

[Caprimaniac]

Good thing you worked it out, Bob.

The MS forum seem to hold a great deal of information.

Why not use a cam phase sensor from a Explorer and LSx coils? Sequential....

[bobcowan]

Originally Posted by Caprimaniac

Why not use a cam phase sensor from a Explorer and LSx coils? Sequential....

Cost, mostly. Not enough bang for the buck. You really don't get much more performance from a COP and sequential injection setup, and the cost is significantly higher.

The cam/crank sensor is driven off the cam gear, just like a regular distributer. Now you're back to the variances of a gear driven component. Plus, they're not cheap. The AEM component is $300 from Summit Racing. Since it's a difficult to locate part, I'd be obliged to carry a spare.

The stock Explorer part can be used, but since that's for a 302/5.0, I'd have to open up the oil pan and change the pump and drive shaft.

When it works, the Hall sensor is relativley inexpensive, easy to change, and more accurate.