Megasquirt EFI conversion

[kb67mopar]

For any of you EFI-knowledgeable people following this thread:

Does anyone know if a stepper motor-type IAC valve should completely seal when it is completely closed according to the computer?

I have verified that it is fully extended and the pintle is up against the machined surface in the throttle body, but there is still vacuum there. The pintle and bore are clean. It just looks like they don't entirely match up. It's a very small leak- idle doesn't change at all if you plug it with a finger, but it does suck your finger in. At this point, it doesn't seem to be an issue, but I've Googled myself half to death over this and haven't found an answer.

For anyone else interested, the car has been quite reliable under EFI. Tuning is progressing, but I think the base map was pretty far from ideal for this motor. The initial map was so rich, I think it ruined some plugs. I need to get a new set in there. Cold starts are still an issue, but it runs great at temperature. The logging feature on the computer has been somewhat useful, but the auto-tune feature has really been the best bet. Hopefully I can get the tune sorted a bit better and move on to computer controlled ignition.

I am not sure but I don't think it's supposed to be a positive seal. I would say since you don't need the IAC to be capable of snuffing out the engine and the throttle blades seal off, that there should always be some flow.

 

[dpstark2]

Thanks. That is what I thought.

Now, obviously I'm digging in deeper to this than I've been before, but I'm confused on another issue. I can't seem to find a combined setting of throttle blade position (set with a set screw, like a carb) and IAC position when up to temp that results in both a good idle in and out of gear.

Currently, in gear idle is a nice 800 RPM or so, but neutral bumps way up to 1200 RPM. It's not lean, and there are no vacuum leaks I can find so far. Timing hasn't changed. With a carb, I had in gear idle about there, but out of gear was only 950 or so.

I need to dedicate some time to this, but I just don't have it right now!

 

[kb67mopar]

I don't think you will find a setting that won't change rpm in gear without adjusting IAC location. Ideally when you put it in gear you would need to somehow have the ecm read the shift and change IAC pintle location to allow more air to compensate for the parasitic draw of the transmission. Modern fuel injected vehicles do this by adjusting to keep a set rpm regardless of engine load....drive by wire makes this really easy but was doable with an IAC to a certain load level in the 90's/earlly 2000's.

 

[dpstark2]

MS3 has a closed-loop idle control mode that does exactly that. I tried it, but had too many other issues going on to enter "closed loop" mode. The IAC still operates in open loop until it's up to temperature, but maybe with the map settled down and some of the sensors smoothed out, it will enter closed loop mode now. I will have to try it.

For what I was describing above, I'm just trying to get something more like the carb- a 200 or so RPM drop, not 400.

And I understand there will always be a drop of some sort- the 1200 rpm hot idle is just way too high. It barks the tires going into reverse and lunges hard into drive... that's not good for the trans or me!

 

[dpstark2]

After some tweaking on the IAC table and the throttle plate position, I have a pretty good in-gear/out-of-gear RPM setting when the car is hot. I need to tweak the table more to get it closer how a choke behaves at cold- it throttles up pretty high right now from the start.

I got to drive the car around with Meep-Meep this weekend he was fairly impressed with how it's coming along. I also got a friend who knews EFI better to ride around with me and he confirmed the map is still off. Idle is rich, WOT is lean, but cruise is fine. I let it tune a while and later I'll go back and try to tweak the table by hand to get closer.

The best part is that starts are so easy... no carb bowl boiling, vapor lock, or anything like that. It just turns over a few times and goes!

 

[dpstark2]

Not sure if anyone cares- but if you do-

I smoothed out the map by hand using the tools provided by the associated software and loaded it on the computer. The car started and idled so much better. I let it tune driving around the block a few times again. It's a good drive with some 25 mph cruise, 45 mph cruise, and a nice stretch I can run it hard up to at least 65 mph. Throttle response continued to improve. Watching the map, I saw it fill in the lower rpm cruise positions and adding more fuel to the low vacuum situations.

After tuning, rolling down the road, held in 2nd gear, response was virtually the same as a carburetor. First was better. Way better. Easily breaks the tires loose with a tap from 10-25 mph. The only complication was that the tuning created a weird edge in the graph that ruined (temporarily) hot idled. It was right at the transition between where the computer auto-tunes (off idle) and doesn't tune (idle). It created a lean surge that was easy enough to fix, but I think this is sort of the situation that frustrated me last time. Anyway- I got it back and it was fine.

Looking at the map again, it was really neat to see how it's zeroing in on the peak torque areas of the motor- right around 4krpm. I think the throttle body is indeed a bit small and it's choking things a bit since it was built to be at about 4.5krpm. Possibly the air cleaner isn't helping. Still, I'm very happy with how it's progressing now that things are getting settled. I smoothed the map out again, and I hope to give it a go tomorrow.

 

[Mr.SpeedFreak]

Thanks for the updates man, I've been following along and I'm excited to see you're getting the hang of it! My end goal for my car was blown full sequential efi, and for now I'm throwing the engine back together carbureted N/A, but I decided to forego the blow-thru carb step and start piecing together my efi setup starting with a trip to the pick-n-pull this weekend haha. Really excited to save up and start piecing everything together and hopefully get myself up to speed on learning how to tune.

And I'll definitely be grilling Meep-meep with questions about how everything's going on your setup when he comes up here to pick up his Roadrunner lol

 

[dpstark2]

Thanks guys! I recently moved and one great thing is that I can actually get home early enough to work on this thing and drive it. I've been taking it out most days and fiddling with the tune. It's definitely getting there, but there is some weirdness that happens on occasion. The tuner keeps putting peaks in the map here and there that don't make much sense. Then there's the fact I just noticed the RPM bins are not evenly spaced- that's great if you put the narrow bins where it makes sense, but these are not in that range. Anyway- it's super cool to be able to tweak fuel delivery without jets and needles, but it's kind of overwhelming at times. There's a lot of variables to tweak.

I also just noticed that there is a decel fuel cut setting that is on by default, and that's kind of causing weird on/off throttle response at times. That's getting turned off today.

I run my motor at about 195F, and the computer default is to assume 160F is fully warm. I think this is wrong for my situation, but I'll ignore it for now. I think one thing I'll do is to not start tuning until it's well and truly warm. The program won't tune until it thinks the motor is warm, but it seems best to put a couple miles on it before starting. Still- it drove quite well on a 15 miles round trip last night and the map made very few changes. I think I'll get some pics of the program and post them here.

 

[dpstark2]

MS3 setup

Here is a screenshot of what you get when you start up the TunerStudio software to configure the Megasquirt. The gauges are configurable. These are defaults, I think.

Here's a shot of the basic settings. You can set the type of fuel injection and algorithms used here. It's not that complicated- mostly you can use defaults unless you're doing something exotic.

- - - Updated - - -

This is where you setup the number of injectors and the type of motor you are controlling. Most importantly, you can calculate the "required fuel number" here.

The "required fuel" number is how long the injectors would need to be open to supply enough fuel to burn the air in the motor at 100% volumetric efficiency (VE). It is a reference that is constantly modified by the computer- but getting this number close is important in order to get the motor to idle at all. Once it idles, you can modify it and the VE table from there.

- - - Updated - - -

The VE table tells the computer how much air is expected to be in the motor during a cycle based on engine load (manifold absolute pressure, MAP), and at a given RPM. This number is relative to 100%, and yes, you can get numbers well in excess of 100%. This is my current table that does have some weirdness.

It has been auto-tuned and smoothed by hand in the regions I haven't reached or where auto-tune will not tune (idle). You can also get a 3D view you can spin and drag.

- - - Updated - - -

This AFR table allows the computer to target an AFR based on load and RPM, or essentially whether you are asking the motor to go hard (accelerate) or not (cruise). With a narrow band O2, this isn't possible.

I haven't fiddled with this table at all, and it likely can be leaned out a bit at higher cruise settings (55kpa or less, and 3100-3700 rpm).

- - - Updated - - -

This table sets the amount of fuel that is added based on engine temperature during warmup. It's your choke. There is also a setting to increase idle RPM while warming up. This table has been auto-tuned, and I did it wrong. It will need to be redone. The reason I'm posting it is to show how it's handled by the computer.

- - - Updated - - -

Finally, here is the auto-tune page. You can use this tool while you drive to allow the computer to see what changes are being made to adjust for AFR, then it will update the VE table for you get it to be a closer match to your AFR table.

On the right, there is a pane titled "Cell Weighting". As you drive, this fills in showing the cells that have been adequately visited. Green is good, yellow means visit more, and white is unvisited. Ideally you want to get this all green. Below it, the "Cell Change" pane shows what changes have been made, either richer or leaner. I'll try to get a picture of it in action later- but it's pretty neat.

 

[kb67mopar]

thats pretty cool. I really want to try to do that one day!

 

[Gpuller]

^^What he said^^
Lots of learning to do.

 

[dpstark2]

Speaking of learning- I noticed as I reviewed the pictures I posted above that the "Incorporate AFR Target" setting is "don't." That means it's not even using the AFR table... Something else to change tonight. I guess that explains a few things.

 

[747mopar]

:moparts ghost:

Here is a screenshot of what you get when you start up the TunerStudio software to configure the Megasquirt. The gauges are configurable. These are defaults, I think.

Here's a shot of the basic settings. You can set the type of fuel injection and algorithms used here. It's not that complicated- mostly you can use defaults unless you're doing something exotic.

- - - Updated - - -

This is where you setup the number of injectors and the type of motor you are controlling. Most importantly, you can calculate the "required fuel number" here.

The "required fuel" number is how long the injectors would need to be open to supply enough fuel to burn the air in the motor at 100% volumetric efficiency (VE). It is a reference that is constantly modified by the computer- but getting this number close is important in order to get the motor to idle at all. Once it idles, you can modify it and the VE table from there.

- - - Updated - - -

The VE table tells the computer how much air is expected to be in the motor during a cycle based on engine load (manifold absolute pressure, MAP), and at a given RPM. This number is relative to 100%, and yes, you can get numbers well in excess of 100%. This is my current table that does have some weirdness.

It has been auto-tuned and smoothed by hand in the regions I haven't reached or where auto-tune will not tune (idle). You can also get a 3D view you can spin and drag.

- - - Updated - - -

This AFR table allows the computer to target an AFR based on load and RPM, or essentially whether you are asking the motor to go hard (accelerate) or not (cruise). With a narrow band O2, this isn't possible.

I haven't fiddled with this table at all, and it likely can be leaned out a bit at higher cruise settings (55kpa or less, and 3100-3700 rpm).

- - - Updated - - -

This table sets the amount of fuel that is added based on engine temperature during warmup. It's your choke. There is also a setting to increase idle RPM while warming up. This table has been auto-tuned, and I did it wrong. It will need to be redone. The reason I'm posting it is to show how it's handled by the computer.

- - - Updated - - -

Finally, here is the auto-tune page. You can use this tool while you drive to allow the computer to see what changes are being made to adjust for AFR, then it will update the VE table for you get it to be a closer match to your AFR table.

On the right, there is a pane titled "Cell Weighting". As you drive, this fills in showing the cells that have been adequately visited. Green is good, yellow means visit more, and white is unvisited. Ideally you want to get this all green. Below it, the "Cell Change" pane shows what changes have been made, either richer or leaner. I'll try to get a picture of it in action later- but it's pretty neat.

Aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaah....................:moparts ghost: TMI, looking at all that makes my brain hurt! My hats off to you for putting yourself threw this haha, should be pretty cool once you get it all dialed in

 

[dpstark2]

It's pretty close already. I'm looking forward to another good day of tuning and turning on AFR correction. What I can relate from this experience is that EFI is not for everyone. I'm a competent electrical engineer, and I've been stumped a few times. I've also been too smart for my own good and thought I knew better than the internet. I do not.

I'll bet people have better luck with the big-box systems- MSD, FAST, Holley- but this is considerably cheaper. Even those big name systems have their quirks. I'm happy to help if anyone wants to start their own system- as much as I can, at least.

 

[Mr.SpeedFreak]

Thank you so much for that post! I've been doing a lot of research mainly on the hardware end of things, but I've just started getting into the tuning/software side and that was a great look into the tunerstudio. I asked meep-meep Alex about your setup while he was up here and got some good conversation going about setting it up and tuning, wish you would have been able to come up with him!

I've started piecing everything together (don't mind the dizzy and orange box) and hope to have my ecu picked up in the next month or so.



I have no doubt it will be a steep learning curve and require a lot of time and work to set up, but I've got plenty of time and I'm definitely willing to learn haha

 

[Mr.SpeedFreak]

Also here's something I've found as far as researching the tuning, this guy has a series of videos I've found really helpful even without having my car set up yet, figured it'd be good to get an idea of what I'll be looking at ahead of time. Here's the video where he talks a little bit about the include/don't include afr target

[video=youtube_share;ThK5tD1oOy4]http://youtu.be/ThK5tD1oOy4[/video]

 

[dpstark2]

I highly recommend instrumenting a running, carbureted engine before going EFI. It really does help to sort out any potential electrical and sensor issues. Plus, it leaves the car running and driveable should you end up with some sort of failure out on the road. I drove mine instrumented for at least a year before swapping to an electric fuel pump. At that point, I had the ECM controlling the fuel pump, so I would know if the ECM ever lost sync. If it did, it would shut down fuel, the carb would empty, and I'd be stalled. All it would take is a jumper to force the fuel pump on.

But anyway- doing things incrementally seems like a really good idea to me. For example, since I had a rock solid O2 reading before adding the fuel injectors, I was fairly confident that the "new" noise on the O2 sensor had something to do with the injectors. It did.

There's a lot of info out there and sorting through it is part of the problem. I still have a few issues, but it's coming along. I also have a multi-port manifold and injectors, but there's no way I'm touching that stuff until I have this stuff sorted out. No need to introduce more problems yet.

Good luck- and let us know what parts/ECM you're planning to use!

 

[Mr.SpeedFreak]

Good to hear! That lets me know I'm on the right track, as that's what the dizzy and orange box are there for. Going to get my car running on electronic ignition and carburetor well before I start implementing any of the efi stuff. However, I'm planning on going straight to a sequential setup (no batch-fire test run) so hopefully that won't be too big of a jump.. but anyway here's what I have planned:

- Megasquirt ms3x ecu
- Edelbrock pro-flo intake and fuel rails
- Ebay 92 mm throttle body
- oem 55 lb/hr Shelby GT500 injectors with injector dynamics extension hats (adding 8 psi of boost after sorting everything out N/A first)
- IAC, TPS, IAT, MAT sensors all standard chevy LS equipment
- Corvette zr1 3 bar map sensor along with the ms3's 2.5 map for altitude correction
- '01-04 Jeep straight 6 Cam Synchronizer sitting in a bushing I'll machine myself
- Ford EDIS-4 style 36-1 trigger wheel and crank position sensor, trigger wheel modified and welded to the back of the factory crank pulley, fabricated bracket for the sensor
- Chevy D585 "truck" coils
- Obviously o2 sensors on the header collectors
- Custom fabricating my own carbon fiber instrument cluster with Speedhut gauges, to make room for adding boost and A/F
- Aeromotive 13301 regulator
- Aeromotive Phantom 340 in-tank pump in modified factory charger tank
- 3/8 sending and return lines
- various AN fittings and everything I'll need to plumb and wire this whole shebang haha

 

[dpstark2]

I had no end of trouble trying to use the EDIS-type sensor without the EDIS box. I ran it to the ECM, shielded, grounded properly, etc., and had frequent loss of sync issues. I even put this setup (pulley, trigger wheel, etc.) on a lathe and verified that it worked. However, I could not get it to work in the car. I believe the issue to be noise on the power lines when sparks are actually flying. I can't explain it, but I have encountered a few noise issues with the MS3 that really should have been addressed in the PCB design. My suggestion is to use a Hall sensor if you aren't using the EDIS box itself.

Since it isn't much trouble to do a batch fire to sequential conversion, I would still suggest doing batch fire first. Once it's verified and working, going to sequential will be easy. However, if you have problems with sequential, I feel like it would be difficult to sort out without knowing that batch fire worked first.

Otherwise- sounds like you have a good plan and list! I hear good things about your car and skills, so keep at it!

- - - Updated - - -

Here's what tuning with auto-tune looks like after another decent run.

You can see the yellow and green where the computer feels that cells have been adequately or not quite adequately visited. The red and blue show adjustments that have been made to the table. Red means fuel removed, blue fuel added.

At this point, the map seems pretty stable from 1400-4500 rpm. I haven't spent enough time above 4500 rpm to know what's going on there. There is a block of red on the left that is where I manually removed fuel from the idle range. I need to smooth this out and further investigate how much leaner I can make it. Once it's good there, I'll re-run the cold start tuning and see if I get a curve that makes more sense.

The beauty of this stuff to me is that I can so easily make changes and then undo those changes. It's a little less rewarding than the mechanical work, but it's still pretty fun.

 

[dpstark2]

Last weekend I turned on closed loop idle control. I was having a lot of issues with in gear vs. neutral RPM. In talking to a friend, he suggested trying closed loop idle. In this mode, the computer uses the idle air valve to set the RPM of idle whenever it believes the car is just sitting at idle. Using this approach, I was able to back off the idle set screw a full turn and let the computer use the IAC to set RPM. It works pretty well! Now I have a consistent 750-800 RPM idle in gear and in neutral. It took some tweaking to get a good response, but it's way better now!