Which ECU's have custom crank trigger "patterns" OR where to buy a signal converter..

[Christopher Chow]

I got an engine, its got a 36-2 crank wheel. I don't want to fit another sensor wheel on it. MoTec has it and the Megasquirt has it, customizable crank/phase triggering patterns. How about the other ECU's?

Ok, How about programming a microcontroller that converted the signal patterns, anyone do this?

-chris

[Christopher Chow]

I got an engine, its got a 36-2 crank wheel. I don't want to fit another sensor wheel on it. MoTec has it and the Megasquirt has it, customizable crank/phase triggering patterns. How about the other ECU's?

Ok, How about programming a microcontroller that converted the signal patterns, anyone do this?

-chris

[B Lewis @ PE Engine Management]

Hi Chris,

Ours is not customizable, but with a 36-2 pattern you can simply remove some of the teeth to get the 12-1 pattern that our system can use.

Making some kind microcontrolled converter would be difficult (but not impossible) because of the timing critical nature of the measurement. Give me a call if you need some help.

[Christopher Chow]

The "teeth" are set in the aluminum flywheel and don't look like they can be removed. I still have to pull it off tho and have a look...

[erny]

Our DTA Fast ECU allows 12-36 teeth and 1-3 missing, or similar. It even allows fancy patterns where missing teeth are not in succession. 36-2 from what iv seen is quite common. I would imagine that most ECU's allow custom patterns.

B Lewis, i am currently doing some research into making an ECU but thought that 12 toothed wheel might not have enough teeth to be accurate enough under hard acceleration, is this not the case?

[B Lewis @ PE Engine Management]

erny,

12 teeth/rev is an excellent compromise between resolution and required processor speed. Using acceleration estimation techniques you can easily obtain 0.1 degree of angular accuracy during even ridiculous accel/decel rates (20,000 RPM/sec). Honda uses only uses 12 teeth on their engines from the factory. Many bikes and cars only use 1 tooth per ignition event as a trigger (4 pulses/rev on an 8 cylinder).

[raska]

Haltech E6X, E8, E11 are all in production and have it.

[DaveC]

SO, going in the other direction, what is the maximum "tooth density" (teeth per inch?) that will produce a clean signal? We were given an ECU that needs a 60 tooth wheel, and we would like to put it where the stock f4i wheel is, which limits its diameter. I believe we tried to make a converter so we could use the Honda wheel, apparently it didnt work quite well enough. http://fsae.com/groupee_common/emoticons/icon_redface.gif

[raska]

What ECU is that? By 60 tooth I guess you mean it requires the standard 60-2 motronic? The biggest part of the tooth density question is choosing an appropriate sensor for the wheel. There needs to be a change in reluctance, so basically the sensor needs to have a small enough area that the fringing fields from the neighbouring teeth won't cause the reluctance to be near constant.

[DaveC]

I'm not familiar with it, but its called an R034 I believe. I've been told its made for Audis, and the 60 tooth wheel gets bolted to the damper (on Audis). This might be one of those thing where we just need to try it out and see if it works...

[Chris Boyden]

You might want to look at the 60-2 wheel from Electromotive, if you don't want to make one yourself. They sell a small diameter wheel that you might be able to use.

[DaveC]

Chris,

Thanks! thats exactly what we need. If someone would have found this last year, the team might not have spent time trying to make a "pulse converter" that didnt work... They also sell a chisel tipped sensor that will work with the small 60t wheel.

Hopefully this might be useful to someone even though it wasnt the original question... http://fsae.com/groupee_common/emoticons/icon_smile.gif

Thanks again,
Dave

[Christopher Chow]

got any tips on building a "box"

[Dan G]

I'm going to be testing a circuit based on this "Divide by N" chip...

http://focus.ti.com/lit/ds/symlink/cd4018b.pdf

in the next few days.

I'm hoping to use it on a VR sensor signal reading off my flywheel teeth (100) and then dividing them by 5 to get a simulated 20 tooth wheel. Then I'll feed it a single reference pulse per rev (couple choices for this) and it will "-1" the 20 tooth signal to give me a 20-1 signal to my engine management.

Hopefully it works! If it does I'll use it on my 951's 132 tooth flywheel to divide by 3 and give me 44-1. The 951 uses a very similar setup from the factory, so I won't even have to fab up the sensors or anything. Stock sensors on the stock bellhousing. Brilliant!

[Chris Boyden]

DaveC,

No problem, hope you guys get it to work.
We had temperature problems with the VR sensors
Electromotive supplied. Be careful mounting them in hot oil. If they aren't temp rated for engine temps, then they will probably fail, and cost you 60 bucks. It seems like the most time we were able to get off a single sensor was about 10 hours, least time a couple hours. That was on an '99 R6. After many failures, I built a crank extension with a rotary seal on the cover and ran a bigger external wheel and the larger VR sensor. That solved the reliability problem, but leaked a little oil. ARghh!

DanG,

Are you using National's LM9011 to interface to the VR sensor?

[Dan G]

No, we're using an LM1815N on the "old" megasquirt and the new MS V3.0 layout we'll move to in a month or so goes to a component-level circuit, no ICs.

This is what the "old" conditioner circuit looks like...
http://megasquirt.sourceforge.net/extra/pix/vr-input.jpg (http://megasquirt.sourceforge.net/extra/setup-wheel.html)
From this page...
http://megasquirt.sourceforge.net/extra/setup-wheel.html

And this is the new IC-less circuit...
http://www.msefi.com/msinfo/ms2/vr_trans_opamp_schem.jpg (http://www.msefi.com/msinfo/ms2/pcb.htm)
From the bottom third of this page...
http://www.msefi.com/msinfo/ms2/pcb.htm

The second circuit is really slick. It has trimmer pots that can adjust trigger threshold and hysteresis to the point that you can input a VR signal, Hall sensor, or even just a primary ignition coil trigger.

[Christopher Chow]

??? using a simple divide by N circuit on a crank position signal with a "pause" in it, causes the "pause" to move around. How does this work? Does it matter that the phase setting, long low pulse, doesn't hit on the same time every revolution, ever time the engine is restarted, etc.

[Dan G]

Chris,

I'm definitely not an EE and I didn't come up with the Divide by N idea. So far its still just an idea and is untested.

But the "-1" or pause part of the whole thing shouldn't jump around from tooth to tooth at all, both speed and reference signals are going to be triggered off the crank. 100 flywheel teeth, fixed. 1 absolute position reference tooth (separate from the flywheel teeth), also fixed.

I think you were thinking it was just going to randomly assign the pause after it counted 99 teeth, which it won't. I believe the chip will auto-sync its divisions (by 5 in this case) to line up with the reference pulse. Otherwise I could see it being an issue if it starts counting on the 3rd tooth and when the reference pulse comes around its mid-division. I'll just have to wait and see how it works.

[Chris Boyden]

If you guys run the reference pulse into the reset pin of the div by N counter it will reset the output and start over as necessary.

If you are able to bolt a modified trigger wheel that has an integer multiple of the required pulses (teeth) and that same integer multiple of missing teeth, you wouldn't need two trigger sensors, and would be able to just use the div-N circuit and the missing pulses would take care of themselves with no additional resets, etc. When the clock (trigger wheel) is missing (missing teeth), no pulses are ouput.

[Christopher Chow]

I think the original idea of this post got lost somewhere.

The engine I have dones't have a cam sensor or seperate reference signal on the crank. Only one Hall effect on the crank.

So far, I think its going to have to be a little bit of microcontroller programming to figure out on its own, where the syncing pulse is on the stock crank wheel.

[Chris Boyden]

Chow,

What are you trying to generate? 12-1?

There's a few issues that have to be addressed with a div-N counter if you are trying to generate 12-1 with a 36-2. You'd still have to grind off a tooth resulting in 36-3, and you'd need a circuit that would reset the div-3 counter at the reference. (I understand it may be hard or undesirable to get in there and take it apart, and grind off or remove a tooth). A missing pulse detector with wide limits would have to fire once at starting to reset the divider to adjust for off by one or off by two teeth problems at the start.

A micro controller would essentially have to do the same, grind off a tooth in software. Assuming you are trying to generate a 12-1 pulse pattern from a 36-2, you could:

1: Detect the reference: Unless you have absolute position tracking via encoders then all ecu's have to wait for the reference pulse to come around before they start firing ignition events etc.

2: Ignore the first pulse (grind off a tooth in software).

3: Clear a counter

4: If a pulse is detected and the counter = 0
fire off a trigger pulse. Else do nothing.

5: Increment a counter after detecting the pulse.

6: If the reference is detected, go back to step 2.

7: If the counter = 3, three pulses (div-3) have passed, goto back to step 3. else goto back to step 4.

That should be fairly easy to do in small PIC processor or whatever floats your boat.
It would be able to divide by other numbers, and adjust to other patterns. But what ever scheme is applied it has to be integer multiples of the input pattern the ECU is expecting i.e.36-3 maps to 12-1 etc.
A micro may be easier to implement and deal with the start up conditions. Getting the timing worked out with a missing pulse detector for a div-N circuit may be tricky, but necessary to deal with start up. Plus you still have to mess with modifiying teeth. Either way, it looks like work.