3.0L Megasquirt w/ EDIS: Headers running at different temps?

[jangell]

I was never really sure what a ground loop was (but I’ve heard the term, and recall that it causes issues in old analog video recording tech, but that’s about it), but I admit that if Josh hadn’t set the harness up this way, I might well have used whatever ground was convenient instead of putting them all together. I mean, they’re all eventually going to the same place, right? It seems like it shouldn’t matter.

— Joe

[jangell]

This morning I checked the ignition timing. The VR sensor is correctly positioned 6 teeth before TDC, and a timing gun is flashing in roughly the right place (I can’t be any more accurate there because the RPMs vary a lot during cranking, usually below the 400 RPM for EDIS to stabilize at 10 degrees, so I’m seeing 6-14 degrees as it runs under 400 RPM to when the engine briefly exceeds cranking RPM and MS tries to boost the timing. The docs note that you really can’t reliably use a timing gun until the act is actually running).

I also removed the SAW plug on the EDIS to lock it into 10 degrees advance. Still won’t start, although it doesn’t seem to backfire.


I pulled each plug, one at a time, and tested them by cranking the engine with the coil connected to test the plug itself. All plugs fired.

Interestingly, with #5 out the car briefly ran (at a very low speed) before stalling out. I have no idea what that means, if anything. It almost started with plug #1 out too, running for less than a second before stalling.

Still, the backfires imply that either some fuel isn’t burning, or at least some plugs aren’t firing. Besides the orange flash visible in the throttle body, the engine would stop while cranking, I assume because the backfire is pushing the piston back down and countering the engine rotation.


I did get a new TPS sensor, so that’s calibrated and working now. At first I overtightened the screws and the sensor got stuck, but once I loosened them it worked properly. I’m wondering if that’s what killed the last one.

Anyway, I have to the the battery charge a bit before I do anymore tests. Any ideas about what else I can check? I’ll be back at the garage tomorrow to bash my head against the wall more.

Thanks!

— Joe

[jangell]

Today's update: Still doesn't work.

I tried rotating the coils (rotating through ACB to BAC to CBA), and it still won't start. This was just to confirm I really do have my coils set up correctly, although all I really know is that nothing starts. I didn't try other combinations like ABC, CAB or BCA.

I also tried just having pairs of coils installed (the "A" coils, the "B" coils, etc), just to see if it would run on two cylinders, and then again with four coils, but none of that had any effect. The only remotely positive thing is that with only two or four coils it didn't seem to backfire. It even ran smoothly enough at one point (with four coils) that I thought it was going to start, but it never actually caught. As you'd expect, there was a strong fuel smell out the exhaust when you have most of the coils out of the car, but at least I know some of the injectors are working.

I was hoping I could narrow down which cylinders are firing and which are backfiring, but I couldn't get it to backfire with pairs of cylinders installed, and it just seemed to crank and never actually fire a cylinder... I think.

And that's really all I did today -- shuffled some coils. I did enough cranks that the battery drained fairly quickly.

I've attached a cranking log with all the coils in place. The points where the RPMs drop to zero is where the engine backfires. I notice there's still noise in the CLT and MAT sensors (under 0.5 degrees, it seems), which is annoying, since the whole point of changing the harness routing (before which the engine actual ran) was to get rid of the noise.

-- Joe

edit: Oh, and I tried a spare EDIS module for good measure, but that didn't matter either.

2018-08-19_11.09.26.msl

[dn010]

You've gone through the wiring to ensure the coils are wired up the same as the wiring diagram. You've made sure the correct coil is going to correct plug as shown on the wiring diagram. You checked to see that the CPS is in position, did you check the air gap? Seems like you've got a good start in narrowing down the problem.

Get rid of the noise on your sensors, maybe you still have a ground issue. Instead of unplugging the SAW connection, you can set base timing in TunerStudio to whatever you want, try 20, try 10. In my opinion I'd use a different connector than the stock bulkhead ones. Those connectors are junk and I've had nothing but problems with them throughout the years running stock. My tail lights would go out, my frequency valve would stop working, and I'd run rough because some pins were just not connecting. It was a very, very good day when I cut down all those connectors and put in a single weather pack. Few more thoughts - did you ground your shielded wire, the actual shielding to the MS? At this point you might try picking up a cheap used EDIS coil pack and wire it in, see if that makes any difference.

I don't have my laptop at the moment so I cannot look at your tune, will look after work.

Also - if you notice a difference with COP 1 & 5 pulled, 1 & 5 are both on the same wire. Check that wire, make sure it has a good connection. If you wired it according to the wiring diagram you posted in your other thread, it will be the gray/yellow wire going to the blue bulkhead connector. Again, I hate these connectors, I've tried in the past to bend the side of female connectors of these in where the split is in the metal a little, thus making a better connection and it has worked somewhat.

[jangell]

You've gone through the wiring to ensure the coils are wired up the same as the wiring diagram. You've made sure the correct coil is going to correct plug as shown on the wiring diagram. You checked to see that the CPS is in position, did you check the air gap? Seems like you've got a good start in narrowing down the problem.

I haven't directly measured the air gap; I was figuring that since it had a good RPM signal that it was fine, but as my higher-level attempts fail I start doing more and more tests like that, so I'll check it out in September (I'm out of town this weekend, so I have to wait a bit longer to do any more checks).

Get rid of the noise on your sensors, maybe you still have a ground issue.

I'll have to do another trace of the wires, I guess. I'm having trouble figuring out where the grounding issue could be. Apparently routing away from the coils wasn't the solution.

I'm starting to wonder if there's a problem with the passenger side injector wiring. In the setup I got from Josh, the passenger side uses the two blue wires run from MS (as is normal) and one large orange one from the 5A fuse. All three of the of the orange wires from the injectors are then soldered to the orange one from the fuse, and all three blue wires to the two blue ones from MS.

When I lengthened these wires to get enough slack to reroute the harness, I cut the blue and orange wires, and soldered in two new blue wires and two new orange wires (the two orange wires to account for the smaller gauge wire I was using relative to the larger original wire).

In theory, if one of those wires isn't connected sufficiently, there wouldn't be enough currently to reliably open the injectors. While test mode seems to click them open and closed OK, it may be that in actual operation they can't operate smoothly or reliably enough.

This is a wild guess, and would only account for the injectors on the passenger side of the engine, but it's the only thing I actually changed before this whole problem started.

Instead of unplugging the SAW connection, you can set base timing in TunerStudio to whatever you want, try 20, try 10.

Ah, cool. I knew that some of the timing controls don't apply to EDIS, but I hadn't looked at the base timing.

This harness is one of the things I got from Josh (sans DB37; I had to get one and solder it up, and I went through the whole harness at the time to make sure I understood it, and to crate my diagram). He had installed a removable plug of some sort (almost looks like a fuse, but without actually being a fuse) that I can just pop out to disable SAW. I have no idea where that plug is from, but it's pretty handy. I got the impression that EDIS ignored any SAW input under 400 RPM anyway, but I may be misremembering.

In my opinion I'd use a different connector than the stock bulkhead ones. Those connectors are junk and I've had nothing but problems with them throughout the years running stock. My tail lights would go out, my frequency valve would stop working, and I'd run rough because some pins were just not connecting. It was a very, very good day when I cut down all those connectors and put in a single weather pack.

I'm also not a fan of the original connectors. I used them in part because Josh's harness had them, but also because I had some silly idea of trying to remain somewhat original, before I realized just how much of the electrical system changes for EFI. I've seriously considered replacing them with GM weatherpack ones myself -- I just haven't wanted to pull everything back out to do it. I also want a decent way to mount them and seal up the holes for the old connectors in the bulkhead. That, and my connector kit maxes at four or six pins. What did you do for yours?

Few more thoughts - did you ground your shielded wire, the actual shielding to the MS?

The crank sensor wire shielding is grounded only to EDIS pin 7. It is not connected to the engine ground, MS, or to the crank connector -- It is only connected to the EDIS. This is how I understood it from the docs (I made a note about it in my wiring diagram so I wouldn't forget).

At this point you might try picking up a cheap used EDIS coil pack and wire it in, see if that makes any difference.

Worth a go, I guess. The thing is, I know the coils work -- I went so far as to pull each plug, one at a time, and saw the nice blue-white spark provided by its specific coil while cranking.

Somewhat relatedly, For some reason my coils don't really want to sit firmly in the socket, and tend to pop out a bit, which is a bit annoying. I cut a some aluminum stack to make a bracket to hold them down more securely, although all I probably really need to do is cut down the conductor springs so the they simply make contact with the tops of the plugs and don't try to force the boots out of the engine. I'm honestly surprised that these boots stayed in on the original Monaco/Premiere setup.

A lot of this 3.0L/EFI conversion was me reading a bunch of docs, getting confused, taking parts that seemed to work for other people and putting them together, and then re-reading the docs again, which now actually made sense with context and my new-found experience. A simple example is the springs springs in the spark plug boots -- I just thought they needed to be that long because that's how they came, but now that I properly understand their purpose it's clear that they only need to be long enough for a solid electrical connection.

I don't have my laptop at the moment so I cannot look at your tune, will look after work.

No problem -- thanks! I forgot to mention that I also tried Josh's known good tune, on which my car had previously run. Still didn't work.

Also - if you notice a difference with COP 1 & 5 pulled, 1 & 5 are both on the same wire. Check that wire, make sure it has a good connection. If you wired it according to the wiring diagram you posted in your other thread, it will be the gray/yellow wire going to the blue bulkhead connector. Again, I hate these connectors, I've tried in the past to bend the side of female connectors of these in where the split is in the metal a little, thus making a better connection and it has worked somewhat.

Good point -- I'll do a more thorough test on #1 and #5 next time I'm at the garage.

Thanks!

-- Joe

[Bitsyncmaster]

For best electrical noise reduction:

Sensor grounds should not conduct any current. So it's best to run all the sensor grounds to one point. Then run power grounds with their own wires to another ground point. It's normal for shielded cables to only have one end of the shield tied to ground, that way no current can flow in the shield.

[jangell]

For best electrical noise reduction:

Sensor grounds should not conduct any current. So it's best to run all the sensor grounds to one point. Then run power grounds with their own wires to another ground point. It's normal for shielded cables to only have one end of the shield tied to ground, that way no current can flow in the shield.

The sensor grounds ("sensor return" in the MS docs) for CLT, MAT and TPS all run to black or black/white wires that meet at a bulkhead connector, then run to a single pin on MS. They aren't connected to anything else other than the individual sensors themselves.

The shield is set up exactly as you said.

So it sounds like I've done it right. I'm not sure where the noise is coming from.

Thanks!

-- Joe

[dn010]

The gap may appear to be fine but the signal my drop when you begin getting higher or lower RPMs, simple enough to pull out some feeler gauges and check it to rule it out.

Odd you still have noise if you used the sensor grounds from the MS. On mine, all wires are bundled together, injector wires, TPS, temp, etc – all of them are together except the CPS which I sent the shielded wire on its own into the hole for the vacuum line and down in the left side pontoon, away from anything. All connections are soldered, heat shrink tube covering them and taped up so that there is certainty they’re connected and the splice won’t break. I have to look at how I grounded the shielded wire – as you and Dave already have said; it is only grounded at one end. I think I grounded the shielded wire from the EDIS module at the MS unit and connected the shields at the module with the CPS wire.

You should double check your injector wiring but when in doubt, drink a 6-pack. Take the empty bottles and put the injector in them & then run your test modes. They should all spray similar amounts after a few rounds of tests. If one bottle is less then you’ve got problems.

Because I felt the same about keeping the car with the ability to go back to ‘original’, I left the ECU connectors in place and stuffed the bulk head connectors back behind the plate only cutting the wires I needed a few inches down from them. The only one I actually and removed completely was the black one for the tail light harness. I then used a 22 pin weather pack bulkhead connector (Amazon!) and custom made my own connector plate to replace the original one with all the connector cutouts. Majority of my wiring is in a separate harness going through the firewall directly to the MS or wherever without having a connector between because I didn’t want to deal with any weak points. Because of this, 22-pin was plenty for me. For you, you could use a few 22-pin ones or buy one that is closer to the pin count you’d require. This is a photo from when the swap was first up and running so the plate looks a little crude, and you can see the wiring going through the firewall at the left side of the frame brace or whatever you want to call it.

Few things for what they’re worth:

I am running my conversion on a mass air flow sensor which saved me the pain of having to create fuel tables. I’m not sure of anyone else basing their fueling off a MAF but I love it.

I did the same as Volvo in their Bertones and ran ground straps to frame from each head.

[jangell]

The gap may appear to be fine but the signal my drop when you begin getting higher or lower RPMs, simple enough to pull out some feeler gauges and check it to rule it out.

A feeler gauge -- I should have thought if that. I was going to try to figure out how to get a micrometer in there. Mine is mounted fairly low, so it's a bit of a pain to get in there.

[QUOTE]Odd you still have noise if you used the sensor grounds from the MS. On mine, all wires are bundled together, injector wires, TPS, temp, etc – all of them are together except the CPS which I sent the shielded wire on its own into the hole for the vacuum line and down in the left side pontoon, away from anything.[QUOTE]

Mine are all bundled together, too, although they separate into bulkhead connectors before going back into a bundle to get to MS behind the driver's seat. My CPS is running through a removed bulkhead connector along with everything else, with headphone-style connectors (shielded, of course) to keep everything clean.

Sensor noise appears to only be in the TPS, CLT and MAT. I have to re-check the TPS, since that sensor was broken anyway. And the noise I'm seeing in the sensors is on cranking -- it seems possible that the starter load might be messing with the readings? It'll be easier to check this stuff once the engine is running.

[QUOTE]All connections are soldered, heat shrink tube covering them and taped up so that there is certainty they’re connected and the splice won’t break.

Mine too -- everything is soldered and (mostly) heat shrinked. I wrapped everything up in self-fusing tape to keep the wires together and safe, although I cut that off while trying to diagnose this problem. I'll re-wrap once it starts again.

I have to look at how I grounded the shielded wire – as you and Dave already have said; it is only grounded at one end. I think I grounded the shielded wire from the EDIS module at the MS unit and connected the shields at the module with the CPS wire.

Mine's grounded at EDIS but the CPS to EDIS shield and EDIS to MS shield are contented to each other, making one long shield.

You should double check your injector wiring but when in doubt, drink a 6-pack. Take the empty bottles and put the injector in them & then run your test modes. They should all spray similar amounts after a few rounds of tests. If one bottle is less then you’ve got problems.

I thought about that, but I wasn't sure how well the injectors would stay in the fuel rails while the pump was running. If that will work, I'll be happy to try it.

On a somewhat related note, I know my car is low on gas (I put in a few gallons from a gas can), and the fuel sender is reading low, but the fuel pressure is good so I'm pretty sure "out of fuel" isn't a problem. Still, I should probably through in a couple more gallons of gas. I mean, I the last fuel I put in was back in December...

Because I felt the same about keeping the car with the ability to go back to ‘original’, I left the ECU connectors in place and stuffed the bulk head connectors back behind the plate only cutting the wires I needed a few inches down from them. The only one I actually and removed completely was the black one for the tail light harness. I then used a 22 pin weather pack bulkhead connector (Amazon!) and custom made my own connector plate to replace the original one with all the connector cutouts. Majority of my wiring is in a separate harness going through the firewall directly to the MS or wherever without having a connector between because I didn’t want to deal with any weak points. Because of this, 22-pin was plenty for me. For you, you could use a few 22-pin ones or buy one that is closer to the pin count you’d require. This is a photo from when the swap was first up and running so the plate looks a little crude, and you can see the wiring going through the firewall at the left side of the frame brace or whatever you want to call it.

"22 pin" and "bulkhead" where the keywords I was missing. Searching Amazon for "weatherpack" mostly finds 2-6 pin connectors. Of course, I need 27 pins if I want to do it all in one connector, and the 29 pin one is more than twice the price of the 22 pin one, but it exists, so that's good. Doesn't cover the black and red connectors (I have an automatic), but I'm OK with leaving those as-is for now. Maybe I'll spend a few hours converting the MS stuff over to that, and eventually 3D print a plate or something to fill the reset of the original bulkhead connector space. I'd always planned to do some mods (digital dash and the like) and I've already re-routed the air ducts into the knee pads for a double-DIN radio install, plus all the ECU wiring I removed, so it's not going to be quick to go back to original anyway.

Few things for what they’re worth:

I am running my conversion on a mass air flow sensor which saved me the pain of having to create fuel tables. I’m not sure of anyone else basing their fueling off a MAF but I love it.

I'm not too worried about the fuel tables, since I'm mostly just going to let auto-tune figure it out. I've gotten the impression that it's pretty good, as long as you make sure to hit all the useful cells in the table when tuning.

I did the same as Volvo in their Bertones and ran ground straps to frame from each head.

I might look into that. I have the one ground strap at the passenger engine mount, and my ground bus to the top of the passenger valve cover. I don't have anything on the driver's side, but I guess that shouldn't matter. Although now I'm wondering how electrically insulating valve cover gaskets and RTV are...

Cool. I want to get one of those fire suppression things at some point. Never quite got around to it.

Thanks!

-- Joe

[dn010]

Be careful with autotune, when I tried it before I went MAF, it did more harm than good.