MS3X sync loss problems
 

Looking for some ideas and help here from any MegaSquirt experts. I seem to be drawing a blank on the MS forum.
My 3.6 project has come to a bit of a halt while I try and figure out its MS3X tuning issues.
The current problem seems to be sync loss during cranking. I actually managed to get it running about a week ago but had sync loss at 2500rpm which acted like a rev limiter at that speed.
I fiddled with a few things to address that problem but made it worse - now it won't start at all.
My fiddling involved adding a couple of resistors into the crank sensor wiring to clean up the signal. This consisted of a 12K resistor in series in the crank sensor circuit (by replacing the jumper at JP1 with resistor) and a 2.2K resistor from the sensor wire to ground (at pin 2 of the DB37 connector).
Today I took out the resistors but that didn't help either.

I am aware that there are two pots on the MS3X card for conditioning the signal from the VR crank sensor, but I can't do much with them until I get the engine running.

Currently, the spark outputs fire correctly for the first few seconds of cranking then they go haywire with random firing. I can clearly see this happening because I have a spark box with LEDs for each spark output from MS3X and for each signal to the COPs. With initial cranking, the LEDs flash in the correct firing order sequence, then they just seem to lose the plot and go random when sync loss kicks in.

http://forums.pelicanparts.com/uploa...1401714552.jpg

COPs are from Toyota Highlander with 350 ohm input resistance.
Starter is a new HiTorque unit.
Crank sensor is standard 3.6 part running on a 3.6 flywheel (60-2 teeth).
Cam sensor is Clewett hall sensor on LH camshaft.
Injectors are standard 3.6 Bosch 0280 150 786.

Below are current ignition settings.
If you have any ideas, I'm all ears. :confused:

http://forums.pelicanparts.com/uploa...1401715104.jpg

Bill,

I am running MS3X board on my install. I am using a 60-2 wheel on the crank (Electromotive) and a home made cam sensor mounted on the cog belt pulley/turbo scravenger pump.

The only differnce I see in your setting is "Input capture". You are using "Falling Edge", I am using "Rising Edge". So, you may want to change that setting and try it. What are you using for a cam sensor?

When first trying to start the engine the VR sensor pots R52 and R56 are to be fully counteclockwise. The adjustment screw will not stop when turning but if you listen closely you will hear a faint click when it is fully counterclockwise. I would disable power to your injection or fuel pump, then crank the engine while doing a composite log in Tuner Studio. You will see a pattern as shown the Mega Manual for the composite tooth logger. If you are not syncing it will show an irregular spike/pattern. That is when R56 and R52 will need to be adjusted. If your VR sensor puts a lot of voltage you may have add back your inline resistor, but that does usually come into play until after the engine is running at a higher RPM.

Have you verified the sensor/wheel gap and/or rigidity of the sensor mount?

Going to be hard to give advice without seeing alterations beyond the MS manual stock wiring diagrams. Normally sync loss is bad grounds... the manual changed about 4yrs ago to move the sensor grounds to the ECU collectively instead of the engine block. This will cut down on noise but if you have a harness that was built before the change (if you upgraded from MS1 or MS2) then you'll have some fiddling to do.

You have the second wheel enabled (dual wheel missing tooth) - do you actually have a cam sensor in place? If no cam sensor then set this to single wheel.

4MS dwell is high for most coils. This value may need to come down. Most are happy between 2.5 and 3.0 otherwise they can overheat.

How many spark outputs are you using for coil on plug? Are the highlander coils 5v or 12v source? I ask because many newer coils are 5v and they'll run for a bit on 12v but eventually fry.

If this were my car, I would return the wiring to when it ran and confirm it matches the MS diagram especially for the crank trigger and sensor grounds.

Please post up the engine and sequential settings, general settings, and other screen shots... if you have uploaded your MSQ to the Megasquirt boards then post it here. If you tell me more about your setup I can dig through and see if anything stands out.

What cam sensor? Have you tried running it in wasted spark mode? Have you tried running without the spark box?

Two things you can try;
1. as someone else suggested, use "Rising Edge" for spark capture
2. Do you have a cam sensor? if you don't, change your trigger wheel setting to "single wheel with missing tooth"

Do Number 2 first if you don't have a cam sensor. Otherwise try #1

There is also the possibility the coils don't like any noise (the 5v Audi/Vw coils are extremely sensitive) so you will need to set to timing based crank fire while cranking. This would be discovered if the engine fires up with a roll start instead of a starter crank (which soils the electrical frequency cloud on and around the engine). This is rare but I have encountered it before even with a perfect wiring installation.

Very early on I identified that the coils overheat very fast with "rising edge", so "falling edge" is correct for these coils.
I haven't checked or attempted to adjust R56 and R52 yet. I suspect they are related to the sync loss at 2500 rpm, but as you say, I can't see how they would be the cause of the current no start problem.

Crank sensor, sensor mount and flywheel are all stock parts set to the factory gap. Can't remember what it is now, just remember setting and checking it.

Wiring diagrams are based on the MS versions for all the basic stuff. It's a shame we can't post files here, but I have posted the complete engine wiring diagram and spark box on the MS forum.
I am happy that grounds are all good even though they all run to the engine block. All the MS diagrams that I can find all show ground to engine, so that's what I followed. The nature of the sync loss, originally consistently at 2500 rpm and now after a few seconds of cranking, just doesn't feel like bad grounds.

Clewett cam sensor installed on LH camshaft.

In the absence of any info to the contrary, I think I just left this at the default value. I can try a shorter dwell, but coil temps are OK.

Coils are 5v. Toyota 90919-02246 made by Denso.
I am using six spark outputs, one for each pair of coils.
When I was have trouble getting first start twelve months ago (yes, its been a long project!), it was identified on the MS forum that the MS spark outputs wouldn't drive one let alone two of the Toyota coils due to their low impedance (350 ohms). To work around that problem, I built a spark splitter box which effectively takes the six spark outputs to six pairs of mosfets (one for each coil). See pic in first above and wiring diagram below.
The mosfets switch a high(er) current 5v supply to each coil. (The mosfet is just acting as a switching relay to provide a high current signal based on the low current MS spark output.) So I could visualise what was happening, I added LEDs to each function so I can easily see that I have good 5v and 12v power as well as outputs from MS3X and 12 signals to coils.

Over the weekend, I reversed the last changes so I could at least get it running again, but after doing that, I am now getting sync loss at start. :confused:

Let me know if you want to see any other screen shots from TunerStudio.
Link to MSQ on MS forum.
Link to log on MS forum.
http://forums.pelicanparts.com/uploa...1401758854.jpg

http://forums.pelicanparts.com/uploa...1401758959.jpg

Steve, cam sensor is Clewett hall sensor on LH cam.
Yes, I tried wasted COP setting. It didn't make any difference.
I haven't tried without the spark box since a year ago! I can easily make up an adaptor to by-pass its internals so that spark outputs go directly to each pair of coils. I may try that just to eliminate the spark box as a problem source.

http://forums.pelicanparts.com/uploa...1401757329.jpg

What part number fuel injectors are you running? I have split the spark a, b,c on the main board of ms2 before for 5v waste spark coil on plug. The coil drivers are a Motorola 5v specific driver. What part type are you using for that duty and what resistors inline for each?

I'll take a look at the msq tonight.

2N4033 is the 5v transistor.

Here are my notes:
http://msruns.com/viewtopic.php?f=88&t=23022

You would need 12 of these resistored ~340ohms each if you're running 12 individual coil on plugs. Not sure the chip would like that kind of load or if these would still fire.

I'm curious about how you have this wired if you're running dual plug. Are you waste spark? If so, are you firing 4 coils per spark output or did you find a way to run this as a 12cyl waste spark with creative wiring...

Injectors are standard 3.6 Bosch 0280 150 786 (22.15 lb/hr).

The diagram below might answer some of your queries about wiring of the coils.
Inside the spark box are two separate circuits for LH and RH banks.
Each spark output is sent to two mosfets, one for top coil and one for bottom coil.
The mosfets switch the high current 5v supply from the 12v/5v converter.
Diagram shows only cylinder #1 and does not show LEDs.

http://forums.pelicanparts.com/uploa...1401766751.jpg

I had tons of sync loss issues until I turned one pot a full turn and voila, nice square waves without the extra waves.

Yeah, crank both pots all the way cc. Don't do anything else until that is done. They don't often tune after that setting.

Are you running any resistor on the yellow line? 6 total spark outputs, correct?

When splitting, the older chipsets liked it.

http://i458.photobucket.com/albums/q...022905100H.jpg

No resistors on the spark outputs.
I am definitely no electronics expert, but I would expect that a series resistor in this circuit would just limit current to the coil. This might protect the ECU but may not be enough current to fire the 350 ohm coils.

Correct. It would limit the output from the ecu on that line. Might be worth a try... One resistor to each driver after the line split. I only have experience with the 2n4033 and none with the 2n7000... I stocked up on the Motorola before supply was consumed. Digikey should still have some if you buy bulk worst case.

xbmwguy has a high boost twin plug setup running 2 Toyota Denso coils directly off each MS3X channel and reports 4 others doing the same despite the low impedance so it might be worth a try.

Yeah, I'm watching that thread and it really puzzles me how that can be possible unless the Toyota coils they are using are high impedance. I think their part numbers are slightly different to mine.
The experts on the MS forum were quite clear about MS3X spark outputs not having enough capacity to drive one, let alone two low impedance COPs so I deferred to their expertise.
I originally had my coils wired up directly to the spark outputs and sparking was all over the place and completely random. I assumed this was because it was taking a few cycles for the spark outputs to get their act together because they were over-loaded.
Anyhow, I am planning to make up a jumper to effectively by-pass the spark box as I really need to be able to remove it from the system to be sure of its effects.

Bill,

when you have sync loss does the rpm signal drop to zero, just become erratic or spike to high numbers?

When you look at the tooth log what do you see? Does it show the cam signal and crank signal? These will help you diagnose if the signal is incoming to the ECU or outgoing to the ignition configuration. Can you borrow an oscilloscope?

The one thing I don't like about COP is that its very difficult to use a timing light. A timing light would tell you if you have synched ignition.

I would be tempted to wire the COP as 3 channel wasted spark and disable the cam signal to get the car to fire. This would allow you to setup your input signals and make sure they are clean. Then go back and switch to sequential spark.

Bill,

The first thing you need to get resolved is the sync-loss issue. If you are not synced, MS will not be sending signals for spark. That is why it is important to do a composite log of the crank and cam signals in Tunre Studio. As mentioned before, make sure R52 and R56 on the MS mainboard are fully CCW as a starting point. I can't remember if you will need to remove your MS3X card for access or not. If you have access to an o-scope that can be helpful looking at signals. You can get a USB o-scope to work with your laptop for about $160.