The next incarnation of MegaSquirt, ULTRAMEGASQUIRT is going to have onboard DIS and a driver for the Bosch LS4u wideband 02 sensor. The whole package on 1 board. I can hardly wait. Imagine no more worrying about a dirty dizzy, springs, putting oil on the felt, vacuum advance. Time it once and forget about it (course that one time is gonna be b***h). Run better gas and a hotter chip for the track with a little more advance.

The whole system will use real automotive connectors, no more DB 37 pin connectors. Sounds like the little kid on the block is growing up.

Anyone have a crank trigger on their 4? The only spot I can see for one is replacing the spacing washer behind the fan with a machined trigger wheel. It appears that you could mount the sensor off the studs that attach the oil cooler

This is gonna be one long winter. I got a s**tload of things to do.

Dave

This new son-of-megasquirt sounds great. I am sure I'll be building one (per vehicle).

Regarding the crank trigger: Jake Raby makes just such a mount. It's a billet piece you mount behind the fan-shroud to possition a trigger over a wheel mounted in place of the fan spacer just as you describe.

-Ben M.

FINALLY.

Thanks for the update Dave.

This is GREAT news.


B

Dave,

Ever think of using the flywheel to trigger the sensor(s)?

Any compentent machine shop can index it for you.

Some Motronic systems (like on the 3.6 911 motors) use the flywheel.

All 944's and all 3.2 engines... all Boxster engines/996 engines..

Yeah.. I would say a "few" use the flywheel.


B

A few gotcha's to watch for if you are going to try using the flywheel:
1) Where are you going to mount the sensor?
2) Will your magnet/sensor combo provide the ECU with a sufficient signal width?
What I mean is, that the farther you move the magnet/sensor from the centerline of the crank the less time the magnet will activate the sensor with each pass. My ECU requires each sensor event to be maintained for about 4 degrees of crank rotation. Using the hall sensor and magnets I have on hand (same as unit sold by SDS) this means the sensor can't be more than about 4 inches from the centerline of the crank (ie. an 8" wheel/ring) or the trigger event width falls below 4 degrees of crank rotation and at higher RPM the ECU will start dropping sensor events. You could get around this by using larger magents that held the signal longer on each pass, or by using another type of sensor. What do the late P-cars use? Is it adaptable?

-Ben M.

So.. the resolution is "crappy" on these Elcheapo systems ??


B

No, the resolution is very high - about 0.4 degrees on a 6 inch wheel. The signal just needs to be held for a sufficient length of time. This is so the ECU can clearly distinguish between signals and noise (interference). Square-wave producing sensors don't need a trigger with a huge radius like sine-wave producing sensors. Unlike a coil & reluctor sensor that produces a low resolution sine wave, the hall effect and optical sensors produce a square wave that is either on or off (whereas the old magnetic sensors had no on/off - just increasing or decreasing signal strength - ie crappy resolution).

With the hall sensors the resolution is excellent - however, the accuracy is a function of how precicely the magnets are placed (or the accuracy of the shudder windows if a shudder is used such as in a Vanagan dizzy).

-Ben M.

I gotcha.

Porsche counts teeth.


B

Hey Ben,

I agree with you on the diameter/timing resolution relationship, however, the Motronic uses more teeth (130 and 60 depending on system) than the old school systems like my early Haltech or even your Autronic.

Motronic Math

The Ford EDIS uses a 36/1 count wheel with a diameter of ~4".

Some of the newer Nissans have small encoder wheels with a line count of 360 so that the system will count each degree of rotation (360°'s) !!!!

Dont most new systems have two methods ?? Like counting a cam wheel and flywheel too double check everything ??


B

Actually my plan right now is to use an adaptation of the Ford EDIS with a 36-1 tooth wheel and the Ford VR sensor. Easy and cheap. However the plans for the unit are that it could use a Hall sensor or other tooth wheels so I'm not going to lock into anything specific.

As I get more information I'll try and settle on one way or another. Right now I'm happy to learn that my idea of using the space behind the fan could work. Only problem I see is that it would be a B***h to tinker with.

dave

Yes. And the sensors do fail from time to time.


B

Stardard Trigger: Synchronising pulse for each ignition event (4 cylinder will have 4 pulses)

Multi-Tooth: Counts the teeth and is triggered by the missing teeth or by a "home" trigger


The Subaru is the only one that I know of that measures crank angle and Cam angle, but I wouldn't be surprised if more manufactures used this method.

I was reading about this back when I was trying to get a GM DIS module to work with my ECU (fruitless attempt). The systems that use two inputs (ie. crank and cam) use the course resolution input (ie cam) just to fire the injectors and ignition at startup before the engine is running, ie when crank speed is erratic (starter spinning crank). When reading the course signal the ECU doesn't try to time any events - it just fires a spark and an injector pulse as soon as it sees a signal (lets say at TDC). Once the engine catches and the rpm's reach a threshhold (say 500rpm) the course signal is ignored, and the ECU opperates off the high-resolution signal exclusively, and the ECU doesn't fire events the instant it sees a signal, but rather calculates the appropriate post-signal delay before triggering the injectors/ignition.

This is why these two-sensor engines seem to fire off with just a bump from the starter. It allows them to quickly index the crank possition and start firing, albeit with rough accuracy. Engines with a single high-resolution crank signal may take a few extra revolutions when cranking the starter before the ECU figures out the rythm and can start firing sparks at the right time.
-Ben M.

Ben'

That's interesting info. i'm still looking to learn a lot. I really hadn't thought of the problems with timing during cranking but all that makes a lot sense. Where are you by the way? Eastern Washington State is a very large area.

Dave

The smart-a$$ answer would be to say It's a big place, but all you have to do is look for THE GUY driving a 914. (So far this year I've seen just one other 914 on the road)

Actually, I'm in the Tri-Cities, Kennewick to be exact. Dry, hot, low altitude. Perfect climate for a rust-prone, low-powered roadster with no heater.

-Ben M.

SDS says that they have only had one reported failure of the crank trigger that was fault of the sensor and they haven’t used that type (brand) since the mid ‘90’s. They have had many a trigger wrecked by improper installation.

I think the SDS trigger would be next to imposable to mount on the flywheel. The sensors are mounted in a 1 ½ X 2 inch aluminum plate. It runs off of three magnets, two triggers and one sync magnet.

The big one is you have to time it with a strobe to get the base line in the computer. I think that I will start and time the engine on the floor then do the install. I just think that it will be easier.

I’m hoping that I can go at the install this weekend. As long as all goes well this week at the ol’ grind.

The following assumes you have the fuel tuning roughed in:

Mark, setting the baseline is easy. Program your ECU for a normal advance curve with 27 degrees advance at >3,00rpm and NO vacuum advance (just to get the baseline set). Also program in your estimated baseline (which you should know pretty closely because you mounted the magnets yourself). Now start the engine and rev it up to 3,500rpm (assuming no fuel issues). Hold a timing ligth on the flywheel in the timing window. You should be able to see the 27-degree mark, but it won't be exactly in the notch. Adjust your baseline in the ECU incrementally until the timing mark is in the notch. You now have the baseline calibrated and 27 degrees in software = 27 degrees on the crank. Now you can program in some vacuum advance and fine-tune your advance curve. This procedure only takes a minute.
-Ben M.

For what its worth, the original design Bruce and Al developed to supply ignition for the Megasquirt crowd involved using a daugter card with the processor on it to provide the programmable map portion of the ignition. This then provided the signal to the Ford EDIS module which provided the power to the coils. Basically a hacked EDIS and messy. Still I went to the wreaking yard and got the EDIS of an old Escort. I now have a trigger wheel and sensor. Interestingly the wheel has an outer ring with the teeth that was pressed onto the serpentine pulley so, it will be really easy to adapt to a machined spacer behind the fan.

Currently there are a couple of options for ignition. One is MegaSpark which is essentially a recoded Megasquirt that drives CDI like MSD. It uses Hall effect and can triggered from a locked down dizzy. It provides a mappable ignition. It's attractive because it is very simple but I'm concerned about accurate triggering and you still have to run the spark back through the rotor and cap.

The new version of the ignition will provide all the EDIS electronics and code so it will not be as messy. The EDIS system is attactive because of capacitive discharge. It is wasted spark, however, I suspect by using the coil pack off of a V8 and tweaking the software you could end up with a coil per plug. Anyway it should be interesting and I'm sure I'm going to learn a lot.

Target date is this fall but everthing these guys have done has taken longer than they expected.

Dave

Speaking of coil on plug... Motec does it.. can this MS be made to fire them ??


B

Brad I don't know the answer to that. There has been some talk about it. We'll have to wait and see the final configuration when they release it. I wouldn't be suprised. All I know is that I'm going to go find an EDIS off an 8 at a pick and pull just in case.

Dave