Fuel-only Microsquirt

Can one of the experts out there enlighten me on this one - can I run a fuel-only Microsquirt install without a crank-angle sensor? Batch fired obviously, not sequential. This would be for 4 cylinder currently running its original KE-Jetronic. Ultimately, I'd like to convert the entire thing to standalone and maybe even throw a turbo on some day, but if I could work through this in baby steps, that would be awesome. I have never done any standalone stuff, but have been wanting to give it a whirl for a while now. Step 1 would be to use the existing intake manifold, rip out the Jetronic plumbing, build a fuel rail for it, and leave the ignition as is.

This diagram I found for MS2 suggests I can, but it seems like the online documentation for the Megasquirt / Microsquirt stuff is a bit ... scattered. This figure came from documentation with things like "coming in 2012" on it

I'm fairly sure that you're going to need some way to know where the crank is so it knows when and how much to squirt.

What wae said ^^^. You need to be able to figure where the crank is so you can fire the injectors at the correct point in time.

What kind of ignition is currently on there? Megasquirt was originally developed as a fuel-only system. Most guys run it for fuel and spark now, but it's completely capable of running fuel-only.

You don't need a 36-1 wheel or any of that unless you're running ignition. Just a tach signal. The KE-Jet has to supply some kind of tach signal to the current EFI ECU, so you should be able to identify and use that signal somehow into pin 24. Hell, if you're running a standard coil, it'll be turd easy.

Later you can adapt to run full fuel and spark if desired. But not required.

I know that with Megasquirt you can easily grab a "tach" signal from the negative terminal of the coil. Depending on what your car is, you may have a dedicated, noise-free signal to the injection controller that you can use.

I do not know if Microsquirt can deal with a tach input. I remember a lot of people mucking about with capacitors and stuff getting the signal:noise just right when triggering a Megasquirt off of a coil, dunno if there is better circuitry built in to the Microsquirt box to eliminate that need.

hhaase wrote: What kind of ignition is currently on there? Megasquirt was originally developed as a fuel-only system. Most guys run it for fuel and spark now, but it's completely capable of running fuel-only. You don't need a 36-1 wheel or any of that unless you're running ignition. Just a tach signal. The KE-Jet has to supply some kind of tach signal to the current EFI ECU, so you should be able to identify and use that signal somehow into pin 24. Hell, if you're running a standard coil, it'll be turd easy. Later you can adapt to run full fuel and spark if desired. But not required.

I like the sound of that. Running stock ignition, so picking it up off the coil would be stupid easy. Sounds like this may be the way to get started on this project...

Back when Microsquirt was a "due out tomorrow, really" for about 2 years running, I built a V1 based Megasquirt board for the Sportster. It runs fuel only and I pick up the timing off the ignition coil. The stock ignition is used on the bike.

Fuel only is like sticking it in only half way.....

Just go for it don't make it harder than it is. Don't be a fraidy cat...

A MicroSquirt will accept a coil negative trigger signal.

I would also recommend a map sensor to be able to determine load. And lastly a wide band O2 sensor.

You'll want to use the MS2/Extra documentation for the MicroSquirt now.

You don't need to time the injection to anything - it's going to be bank to bank anyway.

Thanks all. Stoked that I can wade into this relatively easily. Formulating a plan of attack as we speak.

bentwrench wrote: Fuel only is like sticking it in only half way.....

That's a rather crude way of directing someone to turn a simple project into a probable hulk in the corner of the garage until it is sold for parts.

Fuel only is a great way to learn tuning if you've never done it before. Tuning fuel is easy and fun, and the hardware setup is simple and easy to work over any new-guy hurdles.

Trying to work out crank trigger setups and sorting out an ignition map can make you rip what is left of your hair out, even if you think you know what you're doing. Doing both at the same time, as a first project, can get you into shrieking-chimpanzee-frustration zone.

Knurled wrote:

bentwrench wrote: Fuel only is like sticking it in only half way.....

That's a rather crude way of directing someone to turn a simple project into a probable hulk in the corner of the garage until it is sold for parts. Fuel only is a great way to learn tuning if you've never done it before. Tuning fuel is easy and fun, and the hardware setup is simple and easy to work over any new-guy hurdles. Trying to work out crank trigger setups and sorting out an ignition map can make you rip what is left of your hair out, even if you think you know what you're doing. Doing both at the same time, as a first project, can get you into shrieking-chimpanzee-frustration zone.

QFT. I did fuel and ignition at the same time and the damned car never ran right and took forever to tune well enough to putter around.

The caveat is this: If you have really good documentation of the stock ignition curve, then you can likely get away with doing both fuel and ignition provided you only really tune the fuel. Plug the stock curve in, make sure you have a good MAP signal and blend out any large jumps where possible. This is what I had to do to get my project working reasonably well and translating a distributor ignition curve graph to MegaSquirt wasn't as straightforward as you'd think it would be.

I did ONE car where I had to invent the timing map from scratch.

I... er... did some searching, found some engine controller ROMs from a computer controlled ignition version of the engine on some web forum I've forgotten, downloaded TunerPro, and copied it from there. Engine ran great on that ignition map.

Mind you I think I had eight hours in trying to get the engine to start and run more than just a few hiccups, AFTER all that, because it was a goofy-ass setup in the car.

Moral of the story: The further you go from a plug-n-play setup with an available "base map" that is wrong for your car but still right enough to get you started, to a giant pile of weird, the less hair you will have when you get it sorted. Converting a CIS car to fuel-only will have only a moderate amount of weird, most of that being in getting a TPS set up for it. The bright side is that CIS cars almost always practically stuck the injector down on the valve stem, so the various enrichments shouldn't be too awful to figure out.

Since I currently have zero info about the ignition map, and this needs to remain a (mostly) running project, going fuel only seems like an ideal jumping off point.

Related question - what does MS need for the wideband O2? Just the sensor itself, or one of the Innovate packages? That's one thing I definitely am not familiar with. All the rest I think I'm comfy with. For now...

You need the whole thing for the wideband. The controller, sensor, etc. The controller then has a feed for the MS.

Yep, to run a wideband you'll need a wideband controller in addition to the sensor. In all honesty, I'd suggest start simple with a basic 4-wire narrowband to get you going and keep initial costs down. Hard to make specific recommendations without knowing the engine you're running.

If you're running a mechanical or vacuum advance distributor, keep that for now as well. LOTS of ways to get spark handled when the time comes, almost too many ways, and it's the most complex part. Trying to do conversions to both at once for the first time could be painful.

-Hans

AEM 14.7 Innovate All sell wide bands that use common Bosch sensors. I have been using innovate with good luck. I like the LC2 with a dash gauge because an MS does not have a check engine light.

The new firmware comes with default spark tables that will get you up and running.

Factory ignition Tables are only a starting point as they are 2D and have limitations, a 3D table can do lots more than a wobbly old distributor.

If this is going in a DD you need to line up a loaner till you get it right because unless you have someone experienced to help or do the job for you you are not going to get it done in a weekend. If you are rushed you may well not have a good experience.

If the car is running now, start collecting parts, and survey what you have to see what is usefull. Get the MS built and tested and do all the prep you can get done on the car before you take the plunge.

Are your injectors low or high impedance? If they are low do you have resistors? or you can use a Peak and Hold board to run them.

Idle control, Your old system probably uses a thermal device for fast idle, an MS2 will do closed loop (computer controlled idle). Later style valves will swap in.

I would do a crank trigger and wasted spark or COP.

With an MS2 you can do COP and Sequential injection on a 4 cyl you will need a crank and cam signal (the cam signal is not critical and you can mod an electronic distributor to provide that).

BoxheadTim wrote: What wae said ^^^. You need to be able to figure where the crank is so you can fire the injectors at the correct point in time.

Sequential EFI is mainly for emissions. Bank injected EFI was pretty common before emission rules tightened up. Back to back testing has shown little if any difference in power or driveability. Without the need for SEFI, there is no need for a cam sensor.

hhaase wrote: Yep, to run a wideband you'll need a wideband controller in addition to the sensor. In all honesty, I'd suggest start simple with a basic 4-wire narrowband to get you going and keep initial costs down. Hard to make specific recommendations without knowing the engine you're running. If you're running a mechanical or vacuum advance distributor, keep that for now as well. LOTS of ways to get spark handled when the time comes, almost too many ways, and it's the most complex part. Trying to do conversions to both at once for the first time could be painful. -Hans

Sorry, engine is Benz M102 2.3L 4-cylinder. KE-Jetronic continuous injection, mechanical (might be vacuum, don't remember) distributor. Actually runs great once warmed up, but keeping it running when cold can be interesting. I want to go MS eventually anyway, so I may as well do it now rather than replace pricey stuff like the cold start valve and then not need it 6 months or a year down the line.

Electronic injection does not need a cold start valve, a proper IAC would be nice though.

CIS injection is not going to provide much useful hardware.

No temp sensor, no throttle position sensor, you might be able to use the pump but will need a pressure regulator,

No crank trigger, Distributor probably has points, the biggest hurdle will be finding injectors that will fit and a fuel rail.

Better to get a manifold with injectors, rail and throttle body off a newer year motor. (that will bolt on?)

hhaase wrote: Yep, to run a wideband you'll need a wideband controller in addition to the sensor. In all honesty, I'd suggest start simple with a basic 4-wire narrowband to get you going and keep initial costs down. Hard to make specific recommendations without knowing the engine you're running. -Hans

I will disagree with this. If one is very unfamiliar with tuning, trying to interpret a normal O2 sensor is not easy- especially when you are trying to tweak and add things. WB systems are complete and can be plug and play, and simplify the calibration process greatly.

Then again, I've not had to calibrate a car with just a basic O2 sensor and nothing else for over 20 years. So my view is biased. Which is to say, even WITH an O2 sensor as the primary feedback, we'd do the calibration with a WB to make sure the system was better in control until it was basically perfect.

If narrow band sensors were any good for tuning no one would buy a wideband.

Narrow band sensors are designed for fast response and have an output that is tailored for switching (@14.7 afr). They are used in imbalanced control systems that constantly hunt back and forth from rich to lean. They do not have a predictable linear output to show for variable O2 levels. The chinese sell dash guages for narrow sensors but they are as useful as blinky lights on the dash, neon under glow, and type R stickers.

I'd spring for a wideband if you can fit it in to your budget. I started with fuel-only on the stock narrowband and my life was changed when I put the wideband in.

When you're shopping for one, stay away from the AEM UEGO series. I've got one of those and it's kind of all over the board. Apparently they start out as having a linear A/F-to-volts output but they don't stay that way. I have no personal experience with any other wideband system, but you'll need the sensor plus a controller. Some have just a controller that will output voltage that MS can pick up, others have a gauge for you to mount on your dash as well. I kind of like having the gauge because when I notice that something isn't quite right in the way the car is driving you've got immediate feedback as to if you're rich, lean, or stoich which can guide you down a troubleshooting path a little quicker.

bentwrench wrote: If narrow band sensors were any good for tuning no one would buy a wideband. Narrow band sensors are designed for fast response and have an output that is tailored for switching (@14.7 afr). They are used in imbalanced control systems that constantly hunt back and forth from rich to lean. They do not have a predictable linear output to show for variable O2 levels. The chinese sell dash guages for narrow sensors but they are as useful as blinky lights on the dash, neon under glow, and type R stickers.

FWIW, the "hunting" is actually a good thing- and is intentionally there. Once someone figured that out back in the 80's, everyone put in their own oscillating pattern- and given the O2 sensor, it's a pretty easy control system to set up. It's why they work so well in the second sensor spot- the gas concentrations at that location are really hard to get right in a WB sensor (as they don't see stoich nearly as well)- so even as a controlling sensor- the catalyst monitor sensor is a NB sensor.