Pretty urgent question tuning a YFZ450

[89Steve]

Hey guys,

 

As some of you may know I'm involved with U of D's FSAE program. Long story short we've built a car from scratch and our motor of choice was a YFZ450. It is fuel injected and managed by an ECU made by "DTA". It's a very high end ECU, but we've been told it isn't the best choice for a one cylinder motor. It has worked for us for a couple years now, but we've now run into a snag. The motor was blown due to incompetence by a member or two, so we've purchased a new one. The problem we're having is that it will not start most times, and when it does start it needs to be held at 20% throttle.

 

Here's what we've narrowed the problem down to:

 

The ECU finds timing using a 12-1 trigger wheel. The sensor, for whatever reason, keeps picking up signals when it isn't supposed to. The screen shots from the oscilloscope will help to visualize the issue:

 

This is our "good readout". Here the engine idles fine and picks up the gap right where it's supposed to.

 

http://i39.photobucket.com/albums/e161/steve7672/Goodreadout.jpg

 

This is the readout we're now getting. As you can see the gap signal is completely sporadic and the ECU is actually seeing 2 gap readings. This then causes a no-start and when it does start it backfires something fierce.

 

http://i39.photobucket.com/albums/e161/steve7672/Badreadout.jpg

 

Does anyone have any clue what could be going on here? The sensor wire is mostly shielded, the tune was not changed, and the engine only ran for about 5 minutes before it went haywire. We're now considering moving on to a different ECU (one that is easier to tune for a YFZ, but less capability). I'd like to avoid this, so if anyone has any quick suggestions they're much appreciated.

[maxboost87]

Sounds like the spark plug could be introducing noise into the system. Are you using a resistor type plug? Is there a resistor inside the spark plug boot? That would soak up a lot of noise coming from that yzf coil"if that is the one you are using" Edited April 5, 2011 by maxboost87

[89Steve]

Sounds like the spark plug could be introducing noise into the system. Are you using a resistor type plug? Is there a resistor inside the spark plug boot? That would soak up a lot of noise coming from that yzf coil"if that is the one you are using"

 

The plug we're using is the OE recommended one, so I would assume that the system would be designed to compensate for any spark noise (from an OE plug at least). Would you have any recommendations on how to rule out the possibility of plug noise interference?

[89Steve]

Here is another readout while cranking the motor. As you can see this one is way messed up and there's no way the engine is going to idle with these readings:

 

http://i39.photobucket.com/albums/e161/steve7672/saeengine.jpg

[pure_insanity]

i dont know jack about the system you are using, but. i do remember reading something in the megasquirt stuff about a hysterisis pot. im not 100% sure just what this does but something to do with cutting out noise. also, i know ms has a couple diodes in it that helps reduce noise also. im guessing your ecu should have something similar.

[button]

has the hall sensor vibrated back away from the wheel?

it kinda looks like what happens when theres too much gap.

pic 1 & 2 are the same other than it thinks it sees a double tooth gap on the 11th tooth every other time.

[Technology]

Yeha, I was going to say too much gap.

 

And if it starts with 20 percent throttle, maybe you should play with cranking fuel. If You can do that.

[Maxzillian]

I sort of had a similar problem with a FSAE car using a KTM 525 EXC engine and a Microsquirt. In our case we had to fabricate a 36-1 ring to slip over the flywheel to provide a more accurate timing signal as the Microsquirt did not like having a single tooth trigger at all.

 

After getting this working I ran into a stumble at over 5000 RPM where the displayed RPM would suddenly go wildly out of control. Ultimately I found that the inductive sensor used on the KTM was producing in excess of 40 volts and flooding the Microsquirt circuitry. I ended up wiring what I think was a 10,000 ohm resistor in parallel with the sensor to reduce the peak voltages.

 

I wonder if you may be encountering something similar, but it is a long short. Regardless, I suggest you scope the sensor directly and take a look at the peak to peak voltages. Just for peace of mind.

[Chad]

Looks like the amplitude of the pickup signal is too high. I deal with RF in a cable Headend all day, and any time you turn up the gain, the noise goes up with it. Is there a way to turn the voltage of the tach signal down? Was something done to make it read higher? What are the scales on that graph.

 

hysterisis is a dealy, the signal stays the same, it just lagging behind the initial source. It's a good way to reduce feedback.

 

Using some diodes will help with unwanted currents/voltages, and they are cheap and easy to work with. If the signal is high impedance like a hall sensor, it may not work (won't have enough current to forward bias the Diode). Hig impedance pickups also are very prone to ingress voltages. Is the shiled grounded at the ECU end OLNY? That is very important.

[yzrider]

If I were building a car using a motocross single cylinder that complain would definitly be running a keihn 41mm carb. Screw efi. Those keihns have great response and make great power. Sounds like a cool tail project though

[89Steve]

Thanks for the replies. I'm going to check the gap tomorrow but I don't see how it could have jumped out of place (its a fully stock motor).

 

Maxzillian -- Microsquirt was actually something I was looking into. Only unanswered question I have on it is did it give you the capability of turning on a fan at a certain coolant temp? Also, what dash did you run along with MS?

[89Steve]

Looks like the amplitude of the pickup signal is too high. I deal with RF in a cable Headend all day, and any time you turn up the gain, the noise goes up with it. Is there a way to turn the voltage of the tach signal down? Was something done to make it read higher? What are the scales on that graph.

 

hysterisis is a dealy, the signal stays the same, it just lagging behind the initial source. It's a good way to reduce feedback.

 

Using some diodes will help with unwanted currents/voltages, and they are cheap and easy to work with. If the signal is high impedance like a hall sensor, it may not work (won't have enough current to forward bias the Diode). Hig impedance pickups also are very prone to ingress voltages. Is the shiled grounded at the ECU end OLNY? That is very important.

 

 

Chad,

When pure_insanity suggested looking into hysterisis I googled it to see what I could come up with. I found another guy suggesting the use of diodes, but my question would be how are the diodes wired in? Do they go in series with the signal wire to the ECU?

 

Also, I am 99% sure it is a VR sensor, not hall effect. What do you mean about the sheilded wire being at the ECU end only? The wire is sheilded pretty much all the way from the VR sensor to the ECU. Is that not the right way?

[Maxzillian]

If I were building a car using a motocross single cylinder that complain would definitly be running a keihn 41mm carb. Screw efi. Those keihns have great response and make great power. Sounds like a cool tail project though

 

Unfortunately the FSAE competition requires cars to use a 20 mm restrictor placed after the throttle body. With a carburetor this would yield terrible results. I should know, we used a Honda 450 fuel injected throttle body and didn't relocate the injector because we were pressed for time. The fuel would pool between the restrictor and throttle body and yielded extremely inconsistent fuel ratios and a pretty terrible running engine. I recall one night we did some testing and the entire exhaust system was visibly glowing. Turned all of the muffler packing into a solid piece of glass.

 

With FSAE cars the ONLY option is port injection or at least an injector immediately after the restriction.

 

You can really see the effect here: http://www.youtube.com/watch?v=7njgYFFfuIs

 

From left to right and top to bottom you have RPM, throttle position, fuel injector pulse width, AFR, manifold pressure, ignition timing, engine temp, and battery voltage. The AFR never really stabilizes until 7000 rpm and by that point we were out of our powerband. The engine sure sounded good pulling up to 9000 though!

 

Maxzillian -- Microsquirt was actually something I was looking into. Only unanswered question I have on it is did it give you the capability of turning on a fan at a certain coolant temp? Also, what dash did you run along with MS?

 

The Microsquirt can easily be configured to control a fan. I believe there is a couple of spare outputs available (or you can use the 2nd injector driver). I can't remember what I used, but we did set the Microsquirt to control the radiator fan.

 

For a dash we used a "cheap" unit meant primarily for dune buggies. It required its own temperature sensor, but also gave us an oil pressure, engine rpm, and wheel speed display. Aside from using the tach output, there was no interfacing with the Microsquirt.

 

http://www.remarkind.com/intellidash.html

 

You can see a video with the dash here:

 

Another alternative may be the Vapor dash, but with a monochrome screen it will not be as visible.

 

http://trailtech.net/vapor.html

Edited April 6, 2011 by Maxzillian

[Chad]

the shield needs to be grounded, but only at one end, the grounded end drains off any unwanted EMI energy. you want it at the reciver end so the ground of the reciever is at the same potential as the shiled. If you ground both ends of the shiled, you create a ground loop, which will place a great deal of noise on your supposed shided signal wire. The shield itself becomes a noise source when it's grounded at both ends.