Clewett XDi issues!

[Trackrash]

It's too bad you can't bench test it like I did with my dizzy setup.

[Gordo2]

Sorry to hear of your troubles Hillclimber - I hate electrical problems and was relieved when my XDi installation came together without at hitch.

Here's a link to my XDi installation / wiring thread that includes a few comments and questions on things I could have installed / wired incorrectly:

http://forums.pelicanparts.com/911-engine-rebuilding-forum/812816-engine-harness-multi-pin-connector-carb-conversion-83-sc.html

A while back you posted a thread that helped me ID a hesitation problem I was having with my engine with PMO carbs - a gasket between the PMO intake manifolds and heads developed a leak (which seems to be somewhat uncommon). Greatly appreciate your efforts to document your troubleshooting and posting your findings and outcome.

Good luck - wish I could help more... Keep at it and don't let it beat you.

Gordo

[911hillclimber]

Thank you!
We can all learn something every day.

I will be going through all the wires today while the package is with FedEx.
Hope i find something silly.

Out of interest while I wait, here is a video of the car 12 months ago, 267 bhp, and the scope trace of the VR sensor @ 10 thou gap.

https://youtu.be/c8pzuNoSi0w

[mysocal911]

Originally Posted by 911hillclimber

Thank you!
We can all learn something every day.

I will be going through all the wires today while the package is with FedEx.
Hope i find something silly.

Out of interest while I wait, here is a video of the car 12 months ago, 267 bhp, and the scope trace of the VR sensor @ 10 thou gap.

https://youtu.be/c8pzuNoSi0w

Perfect crank signal (missing tooth TDC) like from a flywheel used on a 964/993, but not a 911 3.2 flywheel which has separate TDC signal.
So where's the scope trace of your cam position sensor signal for your ECU input?

[911hillclimber]

There isn't one.The XDi has only the input of the crank sensor coming to the XDi.
Nothing else on the engine is monitored.
Power comes to the XDi separate to the coil packs.
It is a very simple system as what it is doing is simple, just an electronic distributor.

[r lane]

I did not see this mentioned and if it was, apologize for a duplicate. The contact pins in the multipin plugs at the ECU, can lose their spring tension and lose contact. A 1mm drill bit down the center of the pin should have a slight drag on it. If not, the tiny vanes can be repositioned to help regain good contact. I have had to deal with this and find it a major pain in the ass. How can a company that manufactures an expensive complicated engine management system, have inherently faulty wiring harness plugs. To me it ruins the whole thought of wanting that system. Enough things to worry about at the track without a built in failure.

[mysocal911]

Originally Posted by 911hillclimber

There isn't one.The XDi has only the input of the crank sensor coming to the XDi.
Nothing else on the engine is monitored.
Power comes to the XDi separate to the coil packs.
It is a very simple system as what it is doing is simple, just an electronic distributor.

In reviewing the image of your engine, it appears there's no distributor just coils driven by an ECU, correct?
If so, please explain how your system develops ignition timing without a cam timing signal.

[Focker]

Originally Posted by mysocal911

In reviewing the image of your engine, it appears there's no distributor just coils driven by an ECU, correct?
If so, please explain how your system develops ignition timing without a cam timing signal.

It's an XDi system. There's a toothed wheel on the crank.

Not sure if it was your intention, but you came across as kind of a dick.

[911hillclimber]

I imagine it is the way the XDi unit 'clocks' the TDC via the 2 missing teeth on the crank pulley wheel.
The missing tooth gap is clocked by the crank sensor when the crank rotates and the count then starts. The command to spark to a given set of plugs in the firing order than commences.

You could do the same using the cam shaft, just the timing count is 1/2 the crank speed. However, the XDi needs about 900 rpm to find the crank position.

It is this bold simplicity of the (2005) XDi system that appealed.
They have made these for a lot of years, I know to 3 in England, all 6 plug 911 engines, and all fired up straight away. All 3 running mechanical injection. One has been running untouched since set-up for 10 years.

Today:

Interesting day on the Lola today.

Ran the XDi box's grounds direct to the battery (helps kill some noise)
The crank sensor is removed and the loom modified for the Hall sensor that's coming.

Emailed the starter motor supplier, Brise, yesterday and they have replied. Not a known problem but they want to know how I get on ( )
A suggestion is to add a dedicated earth cable from the starter to the battery ground direct.

I have been through the entire installation.
First was to machine a small tool turned to push directly on the female terminals in the 23 way connector.
No terminals push back (ie when you push the 23 way connector into the XDi box).
All wires in the right pin location ticked-off in the manual, so all good.
The jumper lead added for the Hall sensor coming (linking pins 4 & 5 together) in the 23 way plug.

All power leads (3) belled through from the +ve battery terminal to the XDi box and the two coil packs.
All grounds tested good.

The closest the crank sensor cable is to any of the system on the engine is 10".
Inevitably, the lot all come intimate at the 23 pin plug.

Electromotive have not given me a FedEx tracker ( :roll: ) so have no idea if it is on it's way or gone to the moon.
Clewett has stopped answering emails.....

Emeralds ECU people enthusiastic, Megasquirt Dealer very energetic! DTA still there waiting.

I really cannot see where there is a flaw in the installation.
End of next week might be interesting.

[mysocal911]

Originally Posted by 911hillclimber

I imagine it is the way the XDi unit 'clocks' the TDC via the 2 missing teeth on the crank pulley wheel.

Same as a 964/993.

Originally Posted by 911hillclimber

The missing tooth gap is clocked by the crank sensor when the crank rotates and the count then starts. The command to spark to a given set of plugs in the firing order than commences.

So the XDI system basically develops arbitrary cylinder sparks until it sees a change in RPM, at which time it synchronize the timing to the TDC signal.
This method is used by Porsche in the Cayenne engine.

Based on your problem and how the XDI functions, the ECU is not detecting any timing match to begin proper timing, even though it's receiving a crank timing signal.

[mysocal911]

Originally Posted by Focker

It's an XDi system. There's a toothed wheel on the crank.

Not sure if it was your intention, but you came across as kind of a dick.

Sounds like you're knowledgeable about the XDI system, right? Please fully explain how it functions.

Thanks

[911hillclimber]

Yep, the XDi box, now tested by Electromotive in Virginia, and deemed trouble-free on testing did not 'see' the VR sensor. Never has in my hands.

You can see the VR sensor output. It looks good to my untrained eyes.
Wires all there and in the right places and connected, tested from the 23 pin plug that would fit to the XDi box through to where the wires go in the system.

What can be wrong?

I am no further forward than i was 44 hours ago/4 weeks and $400 worse off.

The lack of detection has been attributed by Clewett to my installation, so I changed everything to their suggestions, then he starter is noisey, so I added suppressors everywhere and will on Sunday add a direct ground from the starter to the battery.
Then the circuit has been damaged by me because the crank sensor cable routing was too close to everything.

I then at their suggestion returned the XDi to Electro at my cost, and my cost to ship back as it is all my faulty.

The side suggestion is to convert to a Hall sensor.

The conversion lead is in the Clewett kit, bit telling that imho.

Overall, i think the design of the XDi is past it's sell by date, is very sensitive to EMI and why not sell it with the Hall sensor rather than the noise prone VR?

Electromotive gave me a $16 discount mind on the $165 sensor.

Costs so far are around $2500 and not a spark in sight.

If this last step fails, i have an (expensive) plan B.

This will have cost me more than my Carrillo rods!

[Porsche 935]

I know this is a stupid question but when you hooked up to your laptop did you choose your sensor type and make sure changes are being programmed into the unit?

[mysocal911]

Originally Posted by 911hillclimber

Yep, the XDi box, now tested by Electromotive in Virginia, and deemed trouble-free on testing did not 'see' the VR sensor. Never has in my hands.

You can see the VR sensor output. It looks good to my untrained eyes.
Wires all there and in the right places and connected, tested from the 23 pin plug that would fit to the XDi box through to where the wires go in the system.

What can be wrong?

Did you read post #30? Did I describe how the XDI functions? If not, what's your understanding? Did XDI describe how their system functions?

If my functional description in post #30 is correct, the XDI ECU is receiving the TDC signal, but is unable to lock-on to the correct timing.
Most likely it's not detecting a change in engine RPM during the setup time, while the engine is cranking.

The key question is whether you get any spark. If your don't, then the ECU doesn't detect the crank sensor signal. You must have a spark to all cylinders while cranking.

[911hillclimber]

The XDi has a light.
Limited, but it starts on a solid Green on Switch ON. Crank the engine and the solid greed should go Green/red alternating.
On my unit the light stays solid Green until you stop the starter, then the light goes solid Red.
In the diagnostics section of the manual this fault is described.
It concludes this situation means the sensor is faulty and/or the Xdi is not catching the sensor signal.
As you can see, the signal is good the resistance of the sensor should around 620 ohms, mine is 618. Clewett state this is ok.
This light condition/sequence has NEVER changed no matter what I try/have done, yet Electro state it tested 'Trouble-Free'.

There has never been a spark anywhere, anytime.
P935:
There is no laptop hook-up. There is no remote facility to fix anything, all is sealed.

So:
All wires safe and sound
XDi box checked at the factory, trouble-free.
VR sensor sending out a great signal.

But no spark.

[Focker]

Originally Posted by mysocal911

Sounds like you're knowledgeable about the XDI system, right? Please fully explain how it functions.

Thanks

Do your own research.

[Focker]

Originally Posted by 911hillclimber

The XDi has a light.
Limited, but it starts on a solid Green on Switch ON. Crank the engine and the solid greed should go Green/red alternating.
On my unit the light stays solid Green until you stop the starter, then the light goes solid Red.
In the diagnostics section of the manual this fault is described.
It concludes this situation means the sensor is faulty and/or the Xdi is not catching the sensor signal.
As you can see, the signal is good the resistance of the sensor should around 620 ohms, mine is 618. Clewett state this is ok.
This light condition/sequence has NEVER changed no matter what I try/have done, yet Electro state it tested 'Trouble-Free'.

There has never been a spark anywhere, anytime.
P935:
There is no laptop hook-up. There is no remote facility to fix anything, all is sealed.

So:
All wires safe and sound
XDi box checked at the factory, trouble-free.
VR sensor sending out a great signal.

But no spark.

Xdi box is grounded as well, correct?

On my xdi the only issue I ever had with no crank sensor was when a wire pulled out of the plug.

How did you go about testing the existing sensor? Is it possible there is a problem between the plug on the sensor and the plug on the main harness?

My gut says it's a wiring issue.

[Focker]

Sorry, just a thought, but what are you using for the toothed wheel? Did it come from Clewett as well?

When on the oscilloscope, was that the sensor as installed with the engine cranking?

[911hillclimber]

Yes, that trace was on the engine, 10 thou gap, normal cranking speed.
Scope was on the connector at the end of the sensor lead.
We also tried the same with the polarity reversed, no difference.
The wave form got rougher at 48 thou but generally same shape, less pk to pk volts.
Clewett said engine would not run at 10 thou, 40 is best.

The pulley wheel toothed wheel came from Clewett as part of the kit, 127mm dia., 60-2

I think the XDi is grounded by the large ground wire coming out of the unit, but we did ground the outer heat sinking case to chassis.
Made no difference.

[Focker]

Originally Posted by 911hillclimber

Yes, that trace was on the engine, 10 thou gap, normal cranking speed.
Scope was on the connector at the end of the sensor lead.
We also tried the same with the polarity reversed, no difference.
The wave form got rougher at 48 thou but generally same shape, less pk to pk volts.
Clewett said engine would not run at 10 thou, 40 is best.

The pulley wheel toothed wheel came from Clewett as part of the kit, 127mm dia., 60-2

I think the XDi is grounded by the large ground wire coming out of the unit, but we did ground the outer heat sinking case to chassis.
Made no difference.

Sounds like you have a good ground so that won't be an issue.

Try doing the same oscilloscope test that you did, but with the sensor harness plugged into the main harness. This would help you eliminate the possibility that there is a fault in the main harness that connects to the XDi box.

Repeating the test that way would tell you if you have a crank signal getting to the connector on the XDi box.

IIRC, the 23pin connector is a snap together design to allow for the addition of an extra coil pack harness. I had to open up my harness when adding the extra coil pack harness when I went from single plug to dual plug.

I really think the problem is in the main harness connecting your known good sensor and the tested Ignition controller.

The XDi is not a very sophisticated piece of electronics and I would have no reason to not believe that it bench tested OK at Electromotive. I wouldn't bother even pugging in the coils at this point, as the issue is 100% the crank signal.