Idle OK, no load OK, drives horribly

[ashford]

definately noise on the tach input, is the signal wire shielded in any way? try more grounds. like ms to engine is the body well grounded to the engine? try routing the ignitin input wire differenly away from others.

[MS Spider]

Thank you!

Signal wire is shielded.

The MS grounds are NOT connected to the engine!!!!

Sometime during this long project I had, in my infinite wizard, decided to create a ground pod near the MS case to simplify my life. This ground pad is under my dash.

I have bundled the 6 wires into a single large gauge wire. I assume I can run the single large gauge wire to the engine?

I will try this first as re-routing the ignition signal wires is problematic.

Report to follow.

Thanks again,

Jordan

[ashford]

Thank you!

Signal wire is shielded.

The MS grounds are NOT connected to the engine!!!!

Sometime during this long project I had, in my infinite wizard, decided to create a ground pod near the MS case to simplify my life. This ground pad is under my dash.

I have bundled the 6 wires into a single large gauge wire. I assume I can run the single large gauge wire to the engine?

I will try this first as re-routing the ignition signal wires is problematic.

Report to follow.

Thanks again,

Jordan

correct, however it is preferred to have 2 ground wires for power and one for the sensors. ie tps, iat, mat and other sensors share one ground(or pair) FROM the megasquirt.

the ground from the edis module should be of quality wire and mounted on the SAME piece of metal the ms gets it ground(important) this is what can cause voltage offsets and false triggers in the ms. since the edis has no signal return(what ford calls a signal ground) the ms input has only 1 wire to use( all ground wires in the ms share a common ground plane) its the reason for needing a good direct ground.

also the shield on your tach input (assuming you have a coax wire) should be grounded on one side only either the ms or pin7 of the edis. connecting on both ends creates a ground loop.

[MS Spider]

Excellent! I feel like I am getting closer!

IAT, TPS, MAT and others have individual grounds from MS to sensor as per original instal.

EDIS is now grounded with a quality wire at the head. The combined MS grounds are now grounded on the head. Both grounds on same bolt on head. Previously they were both on the same bolt under the dash.

In the morning I will confirm the shield is only grounded on pin 7 of EDIS for both sensor and MS (I think this is done and I will check again).

On the road again for testing tomorrow.

Will follow up with a report.

[MS Spider]

Follow up from last night.

Confirmed that both the hal sensor and MS tach signal are in shielded wire and grounded to the EDIS module only. This was done correctly upon instal.

I have separated the PIP and SAW coaxial cables from all other MS wires as best I can. In the original install all wires were bundled together with zip ties in several locations. They are still close to the MS bundle in one spot as all wires pass through the firewall at one location. Is this worth re-routing?

Got a few hours on the road today. Driving has improved slightly although the fundamental problem still exists.

My log from today is 7MB so I have included a screen shot of the log viewer at a problematic point. The tach signal jumps up, plateaus, drops to 0 and then behaves erratically. Shifting gears will often, but not always, alleviate the problem.


Thinking about the ignition being a possible problem I notice that the EDIS module is approx 4" / 100mm from the EDIS coil pack. Can this proximity cause issues?

While the issue is somewhat random it almost always happens under harder acceleration and near 3K rpm. I have not been able to get the car above 3K long enough to learn any more.

Any more suggestions?

[ashford]

since you hava a middleman(EDIS) in the ingnition this is gonna be dificult.first we need to discern where signal is lost or getting fudged at. to me it looks like you are getting some sort of resonace on the tach input, i am leaning toward there being a problem after edis. but i need to confirm it first. the edis will continue to spark if it looses input from the ms, just that it will spark at min advance. your log still shows rpm (but not accurate) and is still injectinf fuel. i see you let off the trhottle right after the incident. did the engine just plain die(like the key was shut off) or did it just run funny like it had no power?

if it straight dies i would look at the edis itself, sensor gap, sensor wiring, etc.
if it just ran funny start looking between edis to ms processor.

things to try on the ms side is to change one thing at a time and see if it gets better of worse. id start out with noise filtering turn it off and see what happens, then change its settings. if no go then its time to start experimenting with hardware.

other things to look at are
does the coil pack still have the condensor?
spark plug wires savely away from signal wires.
plug wires arcing to sig wires? this happened to me but had far worse symptoms.

[MS Spider]

Post incident the car runs like it had no power. I am also leaning towards MS as the issue although mine is gut feel. I will plan a noise filtering exploration when I am next on the road.

Getting on the road is an issue with insurance as indicated previously. Any hardware checks that can be done in a driveway are much appreciated as I can get to one or two each evening.

Coil pack still has condenser and it is wired in.

Spark plug wires are at least 6"/150mm from nearest signal wire. However the EDIS module and coil pack are close to each other +/- 4"/100mm and mounted to the same sheet metal. Any concerns?

Hmmm how would I explore plug wires arcing to sig wires?

One more symptom. The issue appears to be much worse at cold start as the tach signal will drop frequently causing the fuel pump relay to open leading to low fuel pressure and therefore stumbling. I can hear the relay clicking regularly!

[ashford]

this seems (95%sure) after the edis. ms processor is getting poor signal, tps coolant and others look very good. since the car still has spark and runs during the incident. the other 5% is a bad edis module.

a few quic things to check
1. with the engine running and watching tuner studios(watch for the fart) hold it around the rpm it happens and with a screwdriver handle tap the module.
2. ohm out the tach input wire.
3. run the car in complete darkness, let your eyes adjust and watch for sparks under the hood during the conditions, sometimes the plug wires will have a bit of a glow to them

not so quick things.

1. open up the ms, take some marco pic of the board both sides.
2. touch up all solder joints dealing with the ignition inputs.
3. maybee change some things around.

[MS Spider]

Well, tried a couple more things tonight with no success. Frustration is mounting.

Relocated EDIS coil pack capacitor to intake manifold from body. No change.
Changed noise filtering from 40 to 97 in 10 unit increments with no change.

A couple of new thoughts.

1. I am running lowZ injectors. There seems to be MANY threads about this being extremely problematic to many people although I have not found a clear solution. Can anyone suggest a source for the latest thinking on this? I will review my original build notes (using 2008 documentation) to the latest documents and see if any changes have been made.

In my case, it appears that the tach input to MS is the primary culprit and all other readings follow appropriate MS responses to a bad tach signal. My feeling is that if I had LowZ noise issues I would see issues on other signals.

2. The shield for tach signal from EDIS to MS is grounded to the EDIS module. I will investigate changing this to the a ground on MS although I fail to envision how this can be an issue. Again documentation is not really clear on the best set up.

3. The coil pack is powered directly from the alternator. A couple of different diagrams indicate this should be powered from the same source as the MS. Again, I fail to see how this could generate any interference.

Over 600 views! Thanks all for following along. Any suggestions are much appreciated.

A big Thank You to Ashford again for continual assistance.

Cheers,

Jordan

[ashford]

didn't know you had low-z injectors, try changing % if you havent and see if there is a change, but i doubt it all other signals look good.

im more interested in seeing pics of the board and components on the tach input.

[MS Spider]

Pics of board attached.

I will start reviewing the current assembly steps and see if my board is correct.

Any input is much appreciated as always.

Current board configuration showing components.

Front of board small.jpg (233.23 KiB) Viewed 307 times

Current board configuration showing jumpers

back of board small.jpg (208.71 KiB) Viewed 307 times

[MS Spider]

From current manual;

Hall sensors Only (EDIS, VAST, luminition, etc):
a) Link OPTOOUT to TSEL - This is correct
b) Solder a wire between TACHSELECT and XG1 - Tach select is currently connected to Optoin - XG1 is connected to XG2
c) For 5V inputs fit a 470R resistor into the position of R12
For 12V inputs fit a 1K resistor into the position of R12 - I assume EDIS is 12v? - current resistor is 340K
d) For 5V inputs solder a wire between the top hole of C30 and 5V on the proto area.
For 12V inputs, solder a wire between the top hole of C30 and s12c pad on the main board. - C30 is empty and s12c is connected to js9

WTF? My original build notes are not accessible right now but I will check and see if I was following some earlier configuration. I was fairly diligent in my original build as i was pretty nervous.

My plan is to make changes as above although it seems like a big change and I would be more comfortable if I understood what was happening.

Any thoughts or comments?

[ashford]

From current manual;

Hall sensors Only (EDIS, VAST, luminition, etc):
a) Link OPTOOUT to TSEL - This is correct
b) Solder a wire between TACHSELECT and XG1 - Tach select is currently connected to Optoin - XG1 is connected to XG2
c) For 5V inputs fit a 470R resistor into the position of R12
For 12V inputs fit a 1K resistor into the position of R12 - I assume EDIS is 12v? - current resistor is 340K
d) For 5V inputs solder a wire between the top hole of C30 and 5V on the proto area.
For 12V inputs, solder a wire between the top hole of C30 and s12c pad on the main board. - C30 is empty and s12c is connected to js9

WTF? My original build notes are not accessible right now but I will check and see if I was following some earlier configuration. I was fairly diligent in my original build as i was pretty nervous.

My plan is to make changes as above although it seems like a big change and I would be more comfortable if I understood what was happening.

Any thoughts or comments?

the edis module puts out more power grounding the output wire than it does when puttin out 5v. i believe it just has a pullup resistor and then grounds it for its signal. the new configuration uses the input when grounded for "lighting up" the opto isolator instead of the 5v signal. it is more reliable. since it now runs off of 12v instead of 5v it has more resistance to noise and a more robust signal input. dont worry about no c30(it needs to be gone) it is just a place to put constant 12v to the opto isolator where the 5v signal used to go.

also you will need to change your capture edge when done.

[ashford]

make it as is on this page http://www.msextra.com/doc/ms2extra/MS2-Extra_EDIS.htm

[MS Spider]

mods complete as above.

Instal and "driveway" test tomorrow.

Cautiously excited that we are getting to the bottom of this!

[MS Spider]

Well, I think we are making progress but I don't think we have a complete success. Post-board-mod data log below showing tach signal mess very similar to original issues.

The car started OK and would accept throttle input much better that previously. I only did a quick start up and no driving.

Screen shot 2012-08-23 at 8.53.37 PM.png (157.06 KiB) Viewed 254 times

I suppose I could try the noise filter settings again?

Thoughts?

[ashford]

what did you put in c12?

[MS Spider]

I changed R12 to 1K.

I did not change C12

[ashford]

I changed R12 to 1K.

I did not change C12

from what i see in the pic there is no c12, populate it and see what happens

[MS Spider]

C12 is in. 0.01uf.

Same tach spike is present.

Screen shot 2012-08-24 at 10.32.21 PM.png (158.79 KiB) Viewed 221 times

I would say the car "feels" a little smoother on my very short driveway test.

Is the shape of the disturbance significant?