MS3 - Starts then dies
 

First, the configuration:
1988 GTU, streetported 13B
MS3 with MS3x, with ms3-pre1.1alpha18 firmware
Denso yellow RX8 fuel injectors - primaries
Bosch EV14 650cc - secondaries
adjustable FPR set to 44 PSI (injector flow rated at 44.5 PSI)
emissions removed (straight test pipe in place of catalytic converter)
Air pump driven 5th/6th port valves controlled via emissions solenoid
A/C using A/C control.
MS3 with MS3x sequential injection
stock CAS with new shielded cable.
GM temperature sensors
Ford F-150 TPS
NGK Powerdex AFX wideband

Here's the problem. When I start the engine, no matter how the throttle is positioned, the engine will catch, run for maybe a second or two and then die; sometimes it'll shoot up to 2000RPM, or barely catches at 800-1000. The exhaust smells very rich. I've tried disabling ASE and reducing fuel on the warmup enrichment, but still the same result.

I have a very short log file and the current tune. I haven't touched the VE table yet. Vaccum hoses are in place, there are no obstructions in the intake or exhaust. All sensors are calibrated and responding.

All my tables came from aaroncake's MS2 writeup.

Does it bog and die with lots of blackish smoke from the exhaust?

Usually this sort of problem means that you don't have enough fuel.

The way I usually deal with this is by increasing the req_fuel value until the engine will catch and stay running, then I reset the req_fuel to the original value, and scale the VE table up accordingly. The tuning software should have a way to do that.

Make sure you flood clear between each startup attempt too.


Ken

It does seem to bog and die, but I'm not sure about the color of the smoke.

There are two other possibilities unrelated to the ECU that I only thought of this morning.

In addition to the mods, I also installed a new fuel filter and fuel hoses. I also have a Crane HI-6 ignition module on the leading coil that was in working order (hooked up directly to the battery).

So, here are the two possible issues.
1. I have reversed the fuel hoses (though I checked that several times).
2. The wideband is out of calibration (I touched the tuning knob after calibration), reading too lean.

Another possibility, though after reviewing how it's wired it should be fine, is that the ignition box is no longer getting an ignition on signal and the engine is trying to run on trailing only.

Okay, the ignition box and the fuel hoses are fine. I'm increasing fuel on the VE table.

It catches, dies. No black smoke that I can see--but it barely runs for a second or two. It's like it wants to run but then it just seems to stop injecting fuel (despite the fact that I see inj 1 duty running until it stalls).

Okay, I didn't use the req_fuel value. Will try later.

Last night, I went back through my msq and found a few things in Engine and Sequential Settings that were wrong; the req_fuel paramter is in that dialog.

I had a few things wrong; engine size (2000cc when it should be 1300cc), and number of injectors (2 instead of 4).

I re-calculated for req_fuel and noticed the initial calculation was for a 350 V8; changed it to reflect the rotary, which is 1300cc, 2 cylinders, 430cc (rx8 injectors, primary), afr 14.7).

One other question: I'm running alpha firmware on the MS3 (this was back when I was troubleshooting the SD card logging) and ended up sticking with the firmware. Could the alpha firmware be an issue?

Probably not, but it wouldn't hurt to update to the latest 1.1 (actually beta 24).

My guess is just that you're not getting enough fuel.

Ken

Progress. I increased the req_fuel number and the engine has gotten easier to start and will run for seconds longer. But, it seems like it stops getting fuel and stalls out. Also, the wideband is indicating that it's very rich.

Yes, I'm monitoring fuel pressure and when it starts it runs, but then fuel pressure drops and the engine stalls.

If I manually activate the pump, the engine will continue to run. The "fuel" wire is what controls the fuel pump right? The higher the fuel_req number, the longer the engine will run but it got too rich.

But, that's progress--I have more solid numbers and I know the Wideband is working correctly. I'll revisit it next week; cousin's wedding.

I'm not sure but isn't the Engine and Sequential Setting supposed to be 2600cc and 4 cyl. for the rotary. Ken correct me if I'm wrong please. Just seem to remember reading something about that. I was raised in the late 60's and early 70's so being in a fog is kinda my normal state LOL.

In a differant thred you said the car has been setting for awhile could it be that the fuel pump screen is clogged?

In old firmware such as the 1.0.x firmware and ms2/extra firmware, this is true.

In the MS3 1.1 firmware, you should set "stroke" to "rotary," Cylinders should be set to 2, and you should use 1300 cc. This *should* get you a req_fuel between 7 and 10 with stock or close to stock injectors depending on NA vs Turbo.

Ken

I store the car during the winter (and work on it), and I use fuel stabilizer to prevent the fuel from going bad; plus marine fogging oil on the engine internals. Besides, when I jumper the pump relay, the engine runs for as long as I want it to.

The in tank sock was last replaced with the fuel pump (a walbro 255) during the winter of 2004-05. Since then, roughly 30-40k miles have been on it since.

Okay, I updated the firmware to beta24, but I'm having the same fuel pump issue. The fuel pump cuts off just after the engine starts.

Pin 37 is the fuel pump wire, and on the DIYAutotune harness, it is a purple wire marked fuel on the main MS3 plug correct?

What would cause the fuel pump wire to stop grounding the circuit?

Correction. It seems that something else momentarily grounds the pump relay then stops. In diagnostic mode, I get no volts from the fuel pump output on the MS3. Seems that the issue may be ECU-related.

Though I do notice something strange on the pump relay. I may end up rewiring the fuel pump and use a new relay, but it seems that the MegaSquirt isn't grounding its fuel pump relay ouput.

I pulled the ECU panel out, hooked it back up to the JimStim and found that the FP LED is not lighting up. Apparently, I never noticed that during testing. So, the problem is that the ECU is not switching the fuel pump path to ground.

The software most likely is. I would suspect hardware failure unless it is turnig on the fuel pump at some point.

Ken

Something else temporarily triggers the fuel pump relay in the car. That was what was throwing me off; the stock relay would come on independent of the ECU then shut off. If not for that, I would've noticed that the ECU wasn't switching its fuel output to ground.

I pulled the unit and hooked it up to the StimJim and got no FP LED; not in normal operation nor in output testing mode. I referred to some old pictures of the build and found that the FP LED was never lit and I never noticed it.

I did some more searching and found that a few others have had the same issue.

http://msextra.com/forums/viewtopic.php?f=4&t=44469

The parts are cheap enough to buy and the vendor is only a few states away. It's a quick repair too.

That said, if I have to, I'll install the stable firmware if the hardware fix fails. Everything else on the ECU--the ignition outputs, fuel injector outputs, AC, and other i/o--works as it should. Unless the alpha and beta firmware has a software bug that affects the FP output that hasn't been documented.

I'll bet 10 cents that someone reading this thread is shaking their head at my attempts to troubleshoot and fix the problem between the devious factory fuel pump relay and my not paying attention to the fine print.

The thread URL I previously posted is for the MS1, not the MS3 that I have. The components mentioned in that thread have little to do with the fuel pump circuit and aren't even being used in my application.

The relevant schematic diagrams are now printed and sitting on my desk. I'm going to check for bad solder joints and continuity tonight. When the (correct) replacement parts arrive, I'll swap those in and test.

The schematics for the V3.0 board are here; annotated for the MS3 and expansion.
http://www.msextra.com/doc/general/ms3v3schems.html

The fuel pump circuit is actually very simple. I could be dealing with bad solder joints or a bad transistor. That's my best theory until I can get back home and work on it.

There aren't any known bugs in the 1.1 firmware having to do with the fuel pump.

The 1.1 firmware should actually be better for the rotary than the "stable" firmware, but you have to make sure you're using the latest beta. The 1.1 firmware is actually about to go to RC status (Release Candidate) so it's close to release now.

Ken

Well guess what? I'm a complete fool. I got back to the MegaSquirt expecting to find a bad solder joint or a improperly installed part. Guess what it is?

Q2 was never installed. Yeah, the fuel pump output wasn't working because of an assembly error. I go back through my parts bags and find Q2 still in its bag. I assembled this unit months ago and had the JimStim hooked up long before I ever installed it in the car and missed this one part.

Installed Q2, hooked up the JimStim, Fuel Pump output works as it should.

Lesson learned: Double-check your work and pay attention to the details.