efi Cam position sensor for Megasquirt distributorless COP setup

[///mChris]

I just had a really eye opening conversation with a guy that owns a distributor rebuild shop.

He told me that the backlash between the cam and the distributor gear is typically between 4-8 degrees for a good distributor. So when you get on the gas/off the gas your distributor timing can change drastically. He said this factor is built into the advance curve of a distributor. Therefore any distributor based cam sensor is going to be flawed.

Even though he could sell me a VW/Porsche Hall sensor, it would never work properly.

He said that the cam chain also has slack so the ONLY method is to truely put a sensor on the cam gear.

I've been thinking about doing a sensor like Steve Hauer at http://77e21.info/mscampossensor.htm

Instead of using a bent metal tab, I'm thinking of building an aluminum disc with an embedded magnet and a magnetic hall sensor.

What is everyone else doing that goes without a distributor?

Chris

[nbristow01]

Thats pretty interesting, So he is using both the crank trigger and the cam position sensor? I am not sure I am following since the crank trigger setup is run soley by crank position so there should be no 4 to 8 dgree variable if running through an ECU to coils. I am not well versed in this stuff, still just learning about it so thats why I am asking

[///mChris]

There are 2 ways of going without a distributor. An easy way is the Ford EDIS setup which uses only a crank sensor. EDIS is call wasted spark because it fires the spark plugs during exhaust. It actually only has 2 coils. BMW skipped wasted spark and went straight to coil per cylinder. I figure if their engineers think it's worth doing, it probably is.

True coil per cylinder needs to be able to distinguish when it is appropriate to fire the spark plug for each cylinder. The crankshaft rotates twice for every single cam rotation.

[bent]

Unless you've somehow incorporated variable cam timing on your motor, you shouldn't need that much precision for a cam position sensor. Timing events should be based off of the crank sensor/toothed wheel, the cam sensor's job is only to identify which cylinder is firing to allow sequential injection and non-wasted spark ignition. Anything within 10 or 20 degrees should be more than adequate for that job.

[shauer]

Unless you've somehow incorporated variable cam timing on your motor, you shouldn't need that much precision for a cam position sensor. Timing events should be based off of the crank sensor/toothed wheel, the cam sensor's job is only to identify which cylinder is firing to allow sequential injection and non-wasted spark ignition. Anything within 10 or 20 degrees should be more than adequate for that job.

True, on a megasquirt sequential setup the crank trigger is used for all timing (fuel and spark). All you need from the cam is a single position pulse to qualify the next missing crank tooth as the missing tooth prior to cylinder #1 TDC.

As long as the cam position pulse occurs prior to the crank wheel missing tooth, the exact cam timing will not matter.

- Steve

[jimk]

This has worked satisfactorily for me with an Electromotive sequential setup. Why all the full about a cam sensor? It has a 130deg or more window to signal cyl 1!

[///mChris]

This is all pretty interesting. I keep going back and forth about which direction to go. Distributor or cam sensor.

Jimk,

That looks like exactly what I had imagined. Can you tell me more about what you did?

[rover1]

Im planing on using a cam sensor but for F.I. im just trying to figure what im going to use as the trigger wheel and if i will cut a hole in the upper timing chain cover or the cam cover. as for the comment on distributor/cam gear end play never thought about it 4-6 degrees its a hypiod gear so what happens on deceleration does the timing go back the other way.

[Dave3x72tii]

I like EDIS and COP, but unless I'm missing something, the distributer guy is using flawed understanding to scare you away from a distributor. The distributor has some drag, and so does the cam. That's why a fixed tensioner on a timing belt motor works even if (when) the belt stretches. the drag keeps the slack due to lash on the non tension side. As long as the engine is turning, the lash is absorbed by the drag, unless the engine seizes or you miss a shift, at which point who cares what the actual timing is. This differs from driveline lash, the momentum of the car is greater than the drag, unlike the momentum of the distributor shaft.

[///mChris]

The big difference with the distributor is the lash changes drastically between acceleration/deceleration.

I know looking at my own distributor that the up and down movement in the shaft could account for signification timing change on acceleration.

Going back to steve's comment. If this all doesn't matter with the cam sensor only being secondary to the crank, it won't make any difference.

It certainly isn't accurate, but doesn't need to be accurate.

[nbristow01]

The big difference with the distributor is the lash changes drastically between acceleration/deceleration.

I know looking at my own distributor that the up and down movement in the shaft could account for signification timing change on acceleration.

Going back to steve's comment. If this all doesn't matter with the cam sensor only being secondary to the crank, it won't make any difference.

It certainly isn't accurate, but doesn't need to be accurate.

But if the only reason for the cam sensor is to locate its position prior to the missing tooth then even a 4 to 8 degree variable on the distributor will not matter as long as it signals prior to the missing tooth. I guess I am not seeing the point of the cam sensor mounted on the timing cover when you have a distributor already doing this work. And this is not saying you are wrong I am saying I don't see the point if I have the facts right

[shauer]

I guess I am not seeing the point of the cam sensor mounted on the timing cover when you have a distributor already doing this work. And this is not saying you are wrong I am saying I don't see the point if I have the facts right

Either solution will work equally well. It really comes down to personal preference at this point. A distributor modified to produce a single pulse or a sensor through the timing cover will both work. I personally did not like the distributor idea, I thought the timing cover method was a cleaner solution. Again, just my own personal opinion.

- Steve

[jimk]

I guess I am not seeing the point of the cam sensor mounted on the timing cover when you have a distributor already doing this work. And this is not saying you are wrong I am saying I don't see the point if I have the facts right

Either solution will work equally well. It really comes down to personal preference at this point. A distributor modified to produce a single pulse or a sensor through the timing cover will both work. I personally did not like the distributor idea, I thought the timing cover method was a cleaner solution. Again, just my own personal opinion.

- Steve

Drill a hold thru the timing cover and it may not be the cleanest (oil leaks to deal with)! Plus having the possibility of more junk inside coming loose is a downside.

A pin could also be set in the cam dist gear and mount a sensor in the dist hole, o-rings required again.

[Guest Anonymous]

The way many diesels run a "cam gear" timing sensor that's built into the gear is like this.

you'll see how inside the teeth is little squares. Those have a sensor with a single screw hold down and an O-ring to seal out oil seepage. That's what the motor utilizes for timing.

It would be incredibly easy to utilize longer bolts and a machined piece that stacked with the cam gear for individual timing.

[TobyB]

the cam chain also has slack

Actually, it doesn't so much at full chat- that's what the tensioner damper does.

Until the pockets of the gears wear oval. Then your cam timing goes

south in a real hurry.

But a helical gear HAS to have a bit of backlash, and .008 sounds

like not very much. Then add some vertical slop,

and the relatively small size of the dizzy, and pretty soon

I'd believe you can get a significant amount of error.

Of course, given that an old head design like the 2002's has a 'sweet spot'

of at least 5 degrees at any particular rpm/loading,

and few of us are tuning for dead- minimum emissions

how wound up does one need to get about all this?

Me, I like the crankfire route just fine- cheap, easy, minimalist, far more tunable than I know what to do with...

but I'm sure it's not mission critical.

t