Megasquirt tuning almost nailed

DNHunt · 04-09-2003, 10:35 AM

Sunday, after an hour with my wife driving and me typing away on the laptop I believe I've got the tuning very close. A few missed shifts, wrong gears and its running great. The wife proved its almost impossible to lug to a stall.

Some interesting observations:

1) A fast idle solenoid doesn't give enough control of increased airflow during warmup on our engines because it is either open or closed. I found an idle solenoid off of a Golf that has a a diameter of ~ 5mm and so it idles at about 1000 RPMs when cold. It works but I'd like that a little higher. The solution is either stay with the stock AAR or use IAC stepper control. IAC control will be available with the Megajolt Light ignition (project for next winter?).

2) The injectors on a 2.0L are really large for that engine. It should be possible to fuel an engine up to about 250hp. Can anyone say TURBO? But for me that has presented a problem as pulse widths are so short coasting (1.3 milleseconds) that a slight change in manifold pressure (cracking the throttle) results in a big change in the amount of fuel injected. My solution has been to leave this portion of the fueling map very rich with air fuel ratios of about 11.25 to 11.75. Hate to see that gas go out the exhaust pipe, but if I try to lean it out I get wild fluctuations from 12 to 17 and the car bucks. I know thats not perfect but I imagine the Porsche engineers struggled with this also.

3) Voltage is very important to the DIY-WB O2 controller. It requires over 13V to drive the heater in the sensor. When my car was really warm the old voltage regulator let the voltage drop to around 12.5 V and the controller would try to reset which caused the sensor to read stoich. The solution was a new regulator.

4) I found the DIY-WB controller to be very difficult and I suspect I'm not done with it yet since I don't trust my soldering. Still, it's been great for tuning and Megasquirt uses it in closed loop to correct AFR's.

5) I used the stock cylinder head temperature sensor and a program called Easytherm to calibrate it for Megasquirt. Evidently, some of the readings were too high so the Megasquirt program responded with a sensor failure reading.

This is not as big a deal for Megasquirt as for D-jet. Megasquirt uses the CHT as a coolant sensor to control the fast idle solenoid and warmup enrichment while failure of the CHT with D-jet won't let the engine run. In my case I had the threasholds for both set below the failure level so the car ran OK but the bar graph on the computer bounced around.

My solution was to edit the Thermfactor.inc file to allow higher readings and then reprogram the chip.

5) On the whole I feel the car runs better than with the D-Jet and it didn't run bad before. It definitely is harder to lug down.

6) I'm planning to dyno it when I can get some time. In the meantime I plan on making it to the Seattle 914 breakfast 0n the 19th. Maybe some of the people with more experience than me can play with it and see what they think.

Here are some new links to new enrichments and VE tables. I believe these would be a good place for any 914 2.0L converting to Megasquirt to start.

Enrichments table:http://www.geocities.com/dnhunt2002/Enrichment1.html

VE table:http://www.geocities.com/dnhunt2002/VE1.html

Constants:http://www.geocities.com/dnhunt2002/Fuel.html

Tuning Screen:http://www.geocities.com/dnhunt2002/Tune.html

Dave

pbanders · pbanders's Garage

Quote:

Originally posted by DNHunt
[B]1) A fast idle solenoid doesn't give enough control of increased airflow during warmup on our engines because it is either open or closed. I found an idle solenoid off of a Golf that has a a diameter of ~ 5mm and so it idles at about 1000 RPMs when cold. It works but I'd like that a little higher. The solution is either stay with the stock AAR or use IAC stepper control. IAC control will be available with the Megajolt Light ignition (project for next winter?).

If you've got a working AAR system, I'd just leave it on, as it works pretty well when in good shape.

Quote:

2) The injectors on a 2.0L are really large for that engine. It should be possible to fuel an engine up to about 250hp. Can anyone say TURBO? But for me that has presented a problem as pulse widths are so short coasting (1.3 milleseconds) that a slight change in manifold pressure (cracking the throttle) results in a big change in the amount of fuel injected. My solution has been to leave this portion of the fueling map very rich with air fuel ratios of about 11.25 to 11.75. Hate to see that gas go out the exhaust pipe, but if I try to lean it out I get wild fluctuations from 12 to 17 and the car bucks. I know thats not perfect but I imagine the Porsche engineers struggled with this also.

Some ECU's actually turn off any injection during overrun (early 1.7's, and the '75-'76 2.0L). The decel valve was introduced to bypass air around the throttle plate to reduce emissions due to this fuel flow on motors w/o the shut-off. Shut-off has its own problems, in that it cools the heads and combustion cylinders, and you get high emissions when fuel resumes. In stock D-Jet, on overrun, the pulse widths are about where you say, if you're pulling as much as 25 in. Hg on high-speed overrun, and closer to 2.5 ms if you're on light overrun.

Good luck with it and you're doing a great service by publishing your tables - that's what's been lacking in getting people going on the MS conversion, IMO.

Dave at Pelican Parts · Dave at Pelican Parts's Garage

Dave, is your motor a bone-stock 2.0 one? If not, what kind of modifications does it have? (Exhaust, ignition, valves, compression, etc.)

--DD

Bleyseng · Bleyseng's Garage

Bring the car to the breakfast as I want to see and ride in it.
Geoff

chrisreale · 04-09-2003, 04:10 PM

...or bring it to the AX on Sat. to show it off!

3D914 · 3D914's Garage

Excellent job Dave,

For someone who doesn't feel like he has the necessary experience, you're sure doing it right IMHO! Too bad you're all the way up in Washington.

I too would like to know if you're bone stock on the 2.0 L configuration. I want to pursue the MS for a 2.0L with #73 Webcam, so any variations you have in your engine may be a help as well.

Thanks again,

DNHunt · 04-09-2003, 05:37 PM

I hope this doesn't sound too lame, but I don't know the exact configuration of my engine.

I had it rebuilt in 1993. At the time the I had a 3 year old boy and a daughter on the way and my car was something I hoped I could keep and maybe drive in the future. It was not the center of my universe so, I didn't pay much attention to the details of the rebuild. Since that time, I've put about 6,000 miles on the car, the kids are older and I must be having my midlife crisis because I'm playing with it like teenager.

Anyway, the paperwork is lost and all I have is a vague remembrance of the work done. I'm almost positive the heads, valves and cam are stock. I remember that the pistons and cylinders are Japanese and I vaguely remember the mechanic saying the displacement was around 2.3 liters but I can't say for sure.

I ran a D-Jet FI until I changed to Megasquirt and I do remember having to take the car back after the rebuild because it was running lean. The mechanic made some change and it ran well after that. From what I know now I suspected he changed the MPS and sure enough the epoxy plug was gone.

Sorry I can't be of more help, but if you are considering Megasquirt I'm convinced that my settings could get almost any configuration of engine running (excepting a turbo) and an hour of tuning would have it running well.

Dave