Here is a screen shot of the AEM advanced cam/crank setup, + injector and ignition phasing.

This setup starts and revs in neutral on a VVT motor. I have to verify that advancing/retarding the VVT keeps sync.

This assumes that AEM coil outputs #1 and #2 are being used, and injector outputs #1,2,3, and 4 are being used.

Where coil output #1 goes to coils that fire cyls 1,4
coil output #2 goes to coils that fire cyls 2,3

 

I few quick questions that I hope are related as this stuff is currently over my head.

1. Are you are using the 4 tooth OEM CAS? Or some sort of crank wheel too? It looks to me like your using the oem cas...isn't the CAS on the exhaust cam and not affected by the VVT intake cam?

2. Does the OEM CAS have the resolution to run fully sequential fuel or COP's?

3. Can you clear out your PMs?

 

The VVT motors have the 4 tooth crank angle sensor wheel on the crankshaft. There is a 3 tooth cam angle sensor wheel on the intake camshaft. The factory computer uses the Cam signal verses the crank signal to determine where the camshaft advance is.

As far as running COP's and sequential fuel injectors, the factory sensors are ok for that.

Thanks for posting this information up Jason. I noticed your ign phasing #2 tooth and your crank/cam edges changed from what you posted a while back. I can't wait to give the changes a try this weekend to see if it works.

Thanks,
John

 

This is the result of testing on a friend's 2001.

mistersolo, I completely forgot that on my 2000, I am using coil outputs 1 and 3, instead of 1 and 2, which explains the difference in coil phasing. Can't remember why LOL.

I changed the cam sensing to rising edge because on the VVT motor 1 falling edge "crosses over" one of the crank edges when the VVT is in full advance.

turbotim, all AEMs can run full sequential ignition as long as it is provided a cam signal.

Whenever I see my PM box full I clear some stuff out - you may want to email me instead though - email should be in my info.

 

It worked! I used Jason's settings and after setting the syncing the crankshaft sensor timing to the AEM under CONFIGURE/ECU SETUP/SET IGNITION window, the car ran just like it did when I was using the CAS off of the 94.

The next step will be to try and manipulate the camshaft phasing besides full advance/retard.

Thanks Jason.

John

 

The AEM has VVT conrol built in.
Unfortunately its feedback PI only, not PID.

You'll have to call AEM on how to understand the options setup, and which output to use. Send them your cal file or explain on the phone that the cam signal has a quick double pulse then a single pulse 180 cam degrees later, and that you need to use the rising edge because the crank pulses don't "cross over" any of the cam rising edges as the cam is advanced and retarded. Explain also that while the crank sensor has 4 narrow teeth, they are unevenly spaced and so the crank setup ignores 2 of them and thus it acts like you have a 2 tooth crank wheel. (the '656565653' tooth table pattern)

If you are running an AFM the engine will run better, because if the VVT goes way out from your target, the AFM will sense the increase or decrease in airflow and the ECU can compensate. It won't do that with a speed density (MAP) system.

If you play with AEM's VVT, use a very small I, and take advantage of the error table (which is just a nonlinear P). The small I is for correcting drift in the "hold" value (see below).

One of the things you'll have to determine early are the mathematical signs - is fully retarded zero, and more advance positive or negative? The signs here will also determine the sign of the PI feedback.

If you look at the VVT control screen of the AEM, the 3D table is the VVT advance target, vs RPM and Load.

The 2D duty cycle table is duty vs. advance.

The strategy should be to keep the entire duty table flat, then use the error table and PI parameters to close the loop. This is because the duty you need is not a function of the advance you want; it needs a certain value to "hold" the advance, then more duty advances it from there, and less duty retards it from there.

I suspect you will want I to be some small value, and P to be zero. You then use the error table to effect the 'P'.

First you need to find the "hold" duty cycle value.

Start by keeping everything zero'ed, and set the entire duty table at 35%, with the engine idling. Slowly increase it until it starts to advance (engine will go rough then die). The number you get just before it starts advancing is the "hold" value.

 

Can anyone confirm that this will work using the 90-93 PNP version of the AEM? (p/n 30-1710)

Does anyone have such a set up running? (01+ N/A motor in a 90-93 w/ pnp AEM)

EDIT: Pin out on page 5/6 here: http://www.aempower.com/images/produ...%2030-1710.pdf

 

There is little difference between PnP versions of the AEM other than the pinouts.
Signal polarities of ignition outputs of all miatas from 91-05 are confirmed to be the same.

 

You won't be able to control the VVT because injector outputs 9 and 10 are not routed to the NA Miata connector. You'll either have to do some hacking inside the box, or just get one of the AEM EMS with Honda connectors which have all the available outputs in the connector (they are cheaper than the Miata one too - was it the 1050 box that is suggested? you can find the info at the AEM forums).

 

There are two aux connectors in addition to the two NA connectors, Injectors 9/10 are on one of those.

I am a bit concerned that the VSS input is used for the 2nd cam sensor.

 

Yeah, I know about the extra connectors, I guess I never looked at all the pins description in the table. I just remember they didn't list 9 & 10 in the summary at the top of the document. Injectors up to 8 were listed as available spare outputs, but 9 and 10 were crossed out.
About the VSS part - I guess good that the Miata has just 1 variable cam

 

Hey J_Man,

Do you think this setup will work for both 99/00 and VVT motors using COPs?
Kindly have a look at the spark and injector phasing.

BTW is it good to set injector phasing 360* crank away from spark? (end injection TDC overlap isntead of TDC compression)

Cheers.

 

Update it worked great.

 

Ooops, sorry - I've missed this
Looks like you've already tested it though

 

BTW in the BE "crank trigger" thread, I proposed an 8 tooth wheel for the AEM. In order to work with VVT (wherein the cam signal moves phase wrt the crank), the VVT range must not allow any VVT rising edges to cross over any of the crank rising edge. I am presuming the VVT range is < 45* (which is the distance between teeth on an 8 tooth wheel).

 

What about using a modified NA cam sensor for the cam sync on the exhaust cam?
This way moving the intake cam won't affect the sync signal.

I have no VVT but I am currently running my engine that way - NA cam sensor for cam sync and NB trigger wheel for crank signal.
Initially the NA cam sensor I was running was non-modified, but later I opened it and cut out the wide tooth - now it has just the single narrow tooth.
Very easy on the AEM ECU this way - it is set to look for the single cam edge for sync instead of looking for patterns (no more counting cam teeth between crank teeth and vice versa).
And now because there is no complex pattern recognition I use the "Sync Early" option too - the ECU instantly syncs the moment it sees the single cam significant edge.

And that single tooth cam sensor setup a stepping stone to a trigger wheel with more teeth.

 

This seems ideal to me. Then a 8 or 12 tooth crank trigger option once someone waterjets some.

I'm totally useless with this.

 

the VVT range is 25 degrees IIRC. and if 4 teeth is good enough for mazda, why does the AEM need 8?

 

First the AEM cannot interpret unevenly space teeth properly, so it ignores 2 of the 4 miata teeth. It ends up only timing off of 2 teeth.

Second, with more teeth, the ECU gets updated more often as to where the crank position is. This can be significant if the crank accelerates abruptly or decelerates, such as if the clutch is kicked in at low RPM; or if the timing requirements change abruptly such as tip in from idle. Some numbers I've seen bandied around are that 8-12 teeth are needed for hi perf FI motors.

 

j_man, adding an early CAS seems counterproductive to having a "clean" engine bay
And the slow sync allows oil pressure to build up.