My 2.7l EFI project with Carrera intake

[djb25]

Update!

I put a lot of time in over the holidays and tonight I was rewarded - it's alive!

First start attempt and everything. It even took throttle!

Now to finalize some of the wiring and to start tuning...

[scarceller]

The 3.2L intake is resonance tuned for about 5200RPMs and should give you a boost in Torque in that RPM range. Adding SSIs with 2in2out pipe would also be a good combo on this motor.

This intake was designed to bring torque in from 4800-5800 RPMs. And if you wish to lower the peak torque tune slightly down the RPM scale you could build 6 risers that mate at the heads so that in effect you slightly lengthen the runner length maybe 1/2" to 1".

[scarceller]

I have a filter setup a lot like you do now here's a pic:


A nice hi-imp injector that fits in that intake well are any of the Bosch modern day 4 hole injectors. I run Bosch 0280155715 24lb injectors in my 3.2L at 270HP at the crank. The other 19lb injector you can use is the 0280155710 these are very cheap to source and work real well. I've been running 715 injector in my 3.2L for about 2 years now and they help with idle quality as they atomize fuel better than the 1 hole injectors. I also have the injector deadtime specs on the 715s and the 710s should be very much the same.

As far as ignition I'd say save your money and stick with the stock distributor but rip the electronics out of it so it's just a mechanical distributor. Remove the vacuum advance and then install a 60-1 tooth wheel on the crank and feed this signal into your EFI and your all set. Then simply have the EFI control coil events to an old school inductive coil. This is how the factory 3.2L works with the Motronic and it's rock solid. By keeping the distributor you don't need any other sensor other than a simple missing tooth crank wheel that works with your EFI. The trick to making power and throttle response with these motors is in the timing maps, most importantly the Part Throttle ignition map.

I'm not sure what you plan to run for estimating air flow but a modern day MAF can't be beat and will get your AFRs spot on. I use MAFs from the PMAS product line, you see this MAF in the pic above. These devices are so accurate that you have no guess work about how much air is being injested per stroke. It's simple math calc in the EFI and looks like this:
Grams of Air per Stroke = Total Air Flow in GramsPerSec / RPM / 2
It's that simple! Once you know exactly how much air is in the given stroke it's a piece of cake for the EFI to calculate the injection pulse width then compensate the IPW for the target AFR (Lambda). For a street driven car, you can not beat the accuracy of MAF metering, it blows away MAP and/or Speed density systems.

[jpnovak]

Great news. Congrats on getting it started. The tuning part is the fun part.

[djb25]

Sal, thanks for the info. I'm actually running 19 lbs Delphi injectors at the moment. I'm planning on posting a list of components at some point along the way here...

I haven't read much about using a MAF... Maybe I should look into that.

First big setback - my LC1 has failed. It is quite a few years old, but it has hardly been used. Frankly, I'm really annoyed with the thing. It has cycled through essentially every error and simply won't work at all. I guess I'm going to order an LC-2 and hope they are an improvement.

[djb25]

Ok, while I'm waiting on the LC-2 to show up, how about we discuss crank and cam sensors?

The next part of the project (after tuning fuel) will be to do something with ignition control, which means (at least) a crank angle sensor. What's everyone using for a sensor? I want to get something simple and reliable, preferably something that is OEM for some manufacturer and is directly compatible with Megasquirt.

Any suggestions?

My tentative plan is to place the sensor on the right side of the engine. I know most installs place it on the left, but there seems to be a lot more open space on the right of the mounting bracket than on the left. I'm planning on buying a 36-1 wheel and making my own sensor mount.

The Clewett stuff seems really nice, but expensive (since I can make my own bracket, anyway).

[scarceller]

Originally Posted by djb25

Ok, while I'm waiting on the LC-2 to show up, how about we discuss crank and cam sensors?

The next part of the project (after tuning fuel) will be to do something with ignition control, which means (at least) a crank angle sensor. What's everyone using for a sensor? I want to get something simple and reliable, preferably something that is OEM for some manufacturer and is directly compatible with Megasquirt.

Any suggestions?

My tentative plan is to place the sensor on the right side of the engine. I know most installs place it on the left, but there seems to be a lot more open space on the right of the mounting bracket than on the left. I'm planning on buying a 36-1 wheel and making my own sensor mount.

The Clewett stuff seems really nice, but expensive (since I can make my own bracket, anyway).

You will need a crank sensor and may even need a cam sensor. But if you keep the distributor you only will need a crank sensor. I recommend getting the most teeth possible on the crank wheel as it will provide better timing control. A 34 tooth wheel where each edge on the tooth is counted will give you 68 events per rev so this gives you 360/68 = 5.3deg per event. But a 64 tooth wheel will generate double that at 2.8deg per event. The stock 3.2L has 129 teeth with 258 events per rev 360/129= 1.4deg per event. Some EFIs will then create further breakdown of events in software to simulate more resolution, you should read up on megasquirt to see what they recommend.

If that was my engine I'd start with just a crank wheel, an old school coil like the coil from the 3.2L cars, keep the distributor and see how that works for you. My bet is it will work just fine and save you money to spend else where. IMHO - many aftermarket CDI systems generate a lot of electrical noise and cause real issues in the EFI system, you really need to pick a very good CDI box with decent noise suppression. I have found MSD products to be very noisy and have seen them cause EFI issues. If this setup was turbo-charged with hi-cyl pressures then CDI could be useful but for a normally aspirated motor that does not spin past 7000RPMs and old school inductive coil will work just fine. Set spark plug gap between 0.7 and 1.0mm and you'll be fine.

If you really want individual coils per cyl you could investigate trigger wheels that could be installed into the distributor. If you use this approach you may not need a crank or cam sensor, simply a trigger wheel in the dizzy.

Here's a crank wheel
http://www.ebay.com/itm/7924730471

[Raceboy]

36-1 or 60-2 is more than enough for ANY ignition setup. I would not put any trigger into distributor unless completely original look is wanted because distributor has slight play (not as much as the ones that are cam driven, but still).

Putting a 36-1 trigger wheel onto crank pulley is easy and having more coils per engine does not mean you have to have more teeth on the trigger...

[al lkosmal]

36-1 crankfire works very well.

regards,
al

[jpnovak]

There are several methods for integrating the ignition.

Since you already have the MSD and distributor there is a simple and cost effective method to integrate ignition. Buy a new crank pulley (many sources - Rothsport makes a nice one) and mount a 36-1 or 60-2 tooth wheel to it. Then send a variable reluctor pickup (VR) to the MS and route the timing output signal to the MSD input. Simple. The spark output still goes through the distributor.

The challenges you will have are related to signal conditioning. The VR signal is low voltage and as Sal mentioned the engine bay is noisy. Especially with the high frequency, high power noise that is the MSD multi-spark. You must be exceptionally careful to not have the VR signal wire lay parallel with any wires going to ignition including the main power cables. You also have to make sure the VR signal wire is shielded very well from start to ECU. It would be best to run a single cable without a connector to ensure proper shield ground.

The next problem you will face is that the typical mounting point for the VR sensor is the distributor timing/fixing stud. These two can not occupy the same place at the same time. So, you will have to custom make a new VR sensor mount on the engine console and clock the trigger wheel by the appropriate amount such that the missing tooth has the right offset relative to TDC. There is an offset adjustment in the MS software but you have a limited adjustment. Also, the toothed wheels typically have 4 mounting holes which means you are limited to 90 deg increments of wheel adjustment. Plan carefully.

There are other ways to trigger ignition but this is the most cost effective way and will give you excellent spark and timing control.

If you really want to save money then drill your stock pully and mount some magnets in it. Build a Hall Effect sensor pickup and send this signal to the ECU. The advantage of the Hall effect is that it is a 12V signal and the Signal/Noise ratio is significantly improved at the expense of timing control. Your timing control will only be as accurate as the position of the magnets. This will still be more accurate than a distributor pickup and allow for timing control.

The 2.7 distributor is very difficult to mount a pickup inside. Tried it. In this case, a VR pickup drilled into the side can work with a mounted spur gear. You would run a locked distributor and have issues with rotor phasing.

[Raceboy]

You can use Honeywell 1GT101 HALL sensor (costs ~25 euros NEW) and that sensor does not need any magnets, it can pick up even bolt heads reliably if your ECU is good enough.

But I would do it with VR sensor and proper timing wheel, it gives more precision and is very reliable. To be correct, VR sensor is not low voltage sensor, it only has low voltage during cranking. The voltage is non-linear its amplitude (=strength) rises by quite a bit with revs peaking at 20-30 volts easily near redline. There is circuitry that conditions the analogue VR signal to digital on/off signal that is understandable to the CPU.

HALL sensor is just on/off switch, but its drawbacks are that it cannot read very fine tooth pattern..

[will hung]

Originally Posted by djb25

Update!

I put a lot of time in over the holidays and tonight I was rewarded - it's alive!

First start attempt and everything. It even took throttle!

Now to finalize some of the wiring and to start tuning...

Congrats!

I love the look of your intake, did you paint it or is that the original color? Looks a little darker than factory to me.

[scarceller]

Does megasquirt have the ability to drive a coil directly? Or do you always need some sort of external coil driver?

[cmonref]

HALL sensor is just on/off switch, but its drawbacks are that it cannot read very fine tooth pattern..

ON/OFF yes, but an Allegro 627 is good to 12,000 rpm for a 60-2 disc, per the spec sheet. It has done two years in my 911SC to 6,000 rpm on a 60-2 at DEs and has not missed a beat. Clewett crank and disc and mounting setup for an Allegro potted in a 3/8" aluminum tube.

I'll bet the Honeywell 1GT101 will do the same.

Does megasquirt have the ability to drive a coil directly? Or do you always need some sort of external coil driver?

An MS will drive a "smart" coil directly -- the ones that have the driver setup onboard -- but requires a driver for the others.

[jpnovak]

Originally Posted by cmonref

An MS will drive a "smart" coil directly -- the ones that have the driver setup onboard -- but requires a driver for the others.

The MS ECU can drive a standard coil through the distributor but you have to install a high current FET to make it work well without overheating.

The coil driver pack (ignitor) is about the same as the improved transistor. The difference being physical location of the part.

[scarceller]

Originally Posted by cmonref

An MS will drive a "smart" coil directly -- the ones that have the driver setup onboard -- but requires a driver for the others.

I take this to mean that it has no large driver transistor to ground the coil but can MS control dwell times? I guess I'm asking if MS can send a signal to indicate when to start charging the coil and then when to fire it?

I'm just wanting to compare to what the stock motronic does. The stock motronic has a large darlington transistor that 'grounds' the coil's '-' terminal to start charging the coil for a given dwell time then once the dwell time elapses the transistor shuts off and the coil fires.

So does MS control dwell and firing event or does it only indicate when to fire with no dwell control?

[djb25]

Originally Posted by will hung

Congrats!

I love the look of your intake, did you paint it or is that the original color? Looks a little darker than factory to me.

I actually coated the intake with "Cerakote" - it's a ceramic based finish normally used on firearms. I used the "stainless" color. I'm pretty happy with how it turned out.

It actually looks way better than in the photo. There's actually a better picture earlier in the thread.

[djb25]

Originally Posted by cmonref

An MS will drive a "smart" coil directly -- the ones that have the driver setup onboard -- but requires a driver for the others.

This is why I'm leaning towards using the LS1 coils - well, actually I'm leaning towards using the LS2 "truck" coils, since those seem to be the latest and greatest design.

They're smart coils, and are a "coil-near-plug" design. They have a very good reputation. Mounting may be an issue, but I think I can come up with something.

They're actually the coils used on the 6.2l V8 in my Escalade.

[cmonref]

So does MS control dwell and firing event or does it only indicate when to fire with no dwell control?

MS settings include dwell and adjustments for voltage. It calculates the required dwell, then starts the process the appropriate amount before the firing time. At the firing time, the dwell is finished and the plug fires. I THINK is does this thru a grounding process where the 5v control voltage is shut off causing the discharge.

I'm leaning towards using the LS2 "truck" coils

Good choice! I have them mounted on an aluminum plate that sits on the three 6mm pads on the left intake. 6mm threaded rod and aluminum spacers are used to bolt the two-high stack of coils to the plate, and the plate is then bolted to the intake. This allows the coils to be removed easily as a group, if for instance, you want to replace a coil -- you can do it on the bench. The coils are connected directly to the MS, and 3.2 plug wires are used.


The 6mm threaded rod in the near corner of the plate goes down to the 6mm hole in the support piece between intake runners 1 and 2. That rod holds up the unsupported corner of the plate.

[scarceller]

Is this a wasted spark ignition or is it true separate spark events for each cyl? do you have 6 coils on that bracket?

Originally Posted by cmonref

MS settings include dwell and adjustments for voltage. It calculates the required dwell, then starts the process the appropriate amount before the firing time. At the firing time, the dwell is finished and the plug fires. I THINK is does this thru a grounding process where the 5v control voltage is shut off causing the discharge.


Good choice! I have them mounted on an aluminum plate that sits on the three 6mm pads on the left intake. 6mm threaded rod and aluminum spacers are used to bolt the two-high stack of coils to the plate, and the plate is then bolted to the intake. This allows the coils to be removed easily as a group, if for instance, you want to replace a coil -- you can do it on the bench. The coils are connected directly to the MS, and 3.2 plug wires are used.


The 6mm threaded rod in the near corner of the plate goes down to the 6mm hole in the support piece between intake runners 1 and 2. That rod holds up the unsupported corner of the plate.