8000 rpm/300 bhp 3.2 rebuild

[Bruce M]

Very true. Although oscilloscope will also show if the signal is noisy & corrupted.

I’ve had a read of the manual. If it was me.

1) disconnect the big plug on the ECU. On the plug, check there is 600-700ohms between pin 10 & pin 11. Triple check the pin out orientation!! (I’ve got it wrong countless times - is the diagram looking into the ecu connector or the loom connector which are obviously mirrored). The sensor has a ground & a shield, on the shield goes to a different pin (3). Swapping them (ground & shield) would be a problem. Also check pin11 goes the red wire at the sensor. If the sensor wire colour coding is not red / black / bare then all bets are off & the correct colour coding needs confirmed.

2) optional: check at TDC the sensor is pointing to the trailing edge of tooth 11, after the tooth gap.

3) disconnect the power from the coil packs to remove all electrical noise. Crank and check the status light blinks red/green (will not start obviously). If no change (solid Green), the HT system is not the problem.

[911hillclimber]

Thank you Bruce for taking the time to look so closely.
I will go through those ideas, did them yesterday, but I could be confused!

My nice new DVM has Hertz!
Just cranked the engine over, black dvm lead to black, red lead to white.

The result was 1.45 hertz.
Not exactly sure but the engine prob cranks over at 1.5 sec/crank revolution, so 1.45 cycles/sec is smack on.

I think this reasons the sensor is good.

The Clewett kit comes with the sensor housing pre-engineered to align the sensor accurately to the toothed wheel, you just have to set the air gap and pinch it tight.
The housing is fixed and I've assumed to be correct, this kit is sold as right for the 3.2 and it's 127mm dia crank pulley.
This is why i bought the kit, all engineered for the engine, so no faffin' about...

will try to confirm the position.

Will report back on the other steps.

[Bruce M]

If you are reading 1.45hz the meter is picking up the tooth gap and not seeing each tooth. If the engine cranks over at 1.5 times per sec, the meter should show 85-ish hz. Assuming the meter can read the signal.

You might try a few sensor gaps & retest with the meter. Up to 1.5mm if necessary. Not conclusive even if that works, as a good reading on the meter doesn’t mean the ECU will be happy.

EDIT: another thought- a noisy signal might cause the meter to only pick up the big tooth gap & not the smaller signals?

Reading electromotive’s unofficial forum, noise is the major cause. Good electrical separation of the starter motor power feed & solenoid feed is important.

I’d suggest testing with independent power to the ECU & disconnect the coil pack power feeds, just to see a good cranking status (red/green blink) to get a baseline.

[neilbardsley]

Very much what my Senior Advisor on DDK says too.
This absence of a spark anywhere is not a light problem, much darker, something which has a big effect, like no crank sensor signal.

Issue for me right now is what to do.
I'm picking at small things, now getting little help from Clewett; I feel I'm being kept at bay instead of some real engagment to a solution.

i hope Electromotive will respond very soon.

The fact that the XDi unit shows a steady green light when switched on is good.
Fact that the green light does not alternate with the red light is bad.
Fact that the green light stays on during starting is bad and then turns to red when you stop using the starter.

To the people who designed this device that light sequence must describe a failure, and according to their manual it is the sensor, but this is not what Clewett is saying, just lots of this and that produce electrical noise and that is the issue, even the sensor they supply could be wrong!

Now, I'm told the high torque starter is noisy and needs resistors in the wires to it...

This Kit is sold as suitable for the 3.2 air cooled engine, REALLY?
It's cost £2000 and does not work.

I have a month to fix it and get it tuned on the rolling road.
I think I will run this engine on 6 plugs and a distributor yet.

This always seem to happen when company A includes company B's product as part of a kit. Both A and B can pass the buck and the customer is stuck in the middle.

One of the reasons our products are sold at a higher price point is that we factor in the support and you are actually speaking to the designers of the products, not a salesperson. We spend at least 25% of our time supporting customers at all levels.

I'm sure this will get resolved but in my experience, setting up twin plug engines in particular is never straightforward. I would expect there to be other challenges beyond getting the engine started.

P.s. It wouldn't be the first time I have seen a new sensor dead on arrival or the wrong type shipped.

I'm not sure promoting your products is helping Graham?

Sent from my M2004J19C using Tapatalk

[neilbardsley]

I'm getting to suggest something stupid have you tried running the wheel/something back/fwd Infront of the sensor (without the engine running) to see if a signal is generated. Perhaps putting an multi meter inline?

Sent from my M2004J19C using Tapatalk

[911hillclimber]

Thank you all, and I've done a few 'tests' this morning, so a summary is in order to answer several suggestions offered.

Everything I have done below has changed nothing from a week ago.
When you switch the power on, the ecu shows solid green light.
When you crank the engine over, the green light stays solid (it should flash red/green/red/green etc
When you stop cranking the engine is shows solid RED.
Disconnect the power and the light sequence repeats.

Clewett says this show the crank sensor signal is not present, and that is due to interference from the leads and coils and starter motor.
Or the sensor is broken.
I have at their request sent pictures of the layout of the coils and the ecu. I am expecting this to be condemned as they have repeatedly stated the sensor cable must be at least 8" from any of those noise sources. Most significant noise source is the starter motor. The stock starter is ok (!) the high torque starters are dirty.

This is what i have done and the result:

All wires are continuous and carry 12v nominal to all places they should.
All wires are in the correct location in the ecu 23 pin plug.
The crank sensor has been used with a 35 thou air gap, and this is now 48 thou max is stated to be 50 thou.in the Clewett manual.
As far as I can tell, the sensor is timed correctly to the toothed wheel as per instructions and as engineered by Clewett.

I have tested the sensor using the Hz setting on my dvm, 1.45 Hz after cranking the engine for 15 seconds.
I will repeat this, but the sensor lead id still close to the plug leads/coil pack remember, 2" clear at best. (not 8"+)

My next sensor test is to remove it and track Hz away from the car. This will answer the frequency count, though the meter could still be the problem.

I have powered the 3 feeds to the ecu and the 2 coil packs with a n independent 12.7 volt battery. Made no difference, BUT cranking drops the volts to 9.7v from 12.5, so ecu and coils do not get 12 volts without the auxiliary battery.

I have disconnected the power to the 2 x coil packs and note that the light sequence has not changed in this state.
I can only surmise that with the coil packs disabled this way there is no noise from them or the leads so the sensor signal should not be corrupted?

So, with all this done where am I?

I will do the sensor out-of-the-car test. That will take a bit of time, done today using my lathe as a rotation source, easily 1000 rpm.
I am sure the wiring is all correct.
The only outstanding issue is the plug lead/coil pack locations, so this means clearing the engine and re=wire all the plugs again using as many of the cut leads as I can.

The coil pack looms will reach to the gearbox end and reach the ecu with ease, a few wires will need lengthening, trivial in the scale of things.
Before i clear the engine looms I will await Clewetts response to my pictures, but i know what he will say.

finally, to clean the wires to the starter (the thin ones, not the cable) needs a 1.5K ohm resistor in line.
How or rather what do i use for this please?
Is it simply a resistor from (say) RS? soldered and sleeved in-line with the existing cables? Could do with some guidence.

[Bruce M]

<<edit>> .

I'm not sure promoting your products is helping Graham?

True but he makes a fair point I think.

Designing electronics for an automotive environment is hard. Nasty place for electronics. Doing that job for an OEM where you have total control over all aspects such as wiring, connectors, strain relief, abrasion protection is bad enough but doing that job for an aftermarket solution that could end up being used in less than ideal installations is much worse and customer support is going to be a cost factor.

I picked up a couple of pointers from the unofficial forum (where engineers from EMC do post). Graham is unfortunately not going to like these…

1)If the environment is noisy, you can change the VR sensor & replace with a hall sensor which less sensitive to noise (generates a 5volt square wave signal).
2) with the VR sensor, you can try adding a 1k ohm resistor across the signal & ground wires of the sensor as it suppresses some noise. (That fix worked for a Pantera owner).

[Bruce M]

Graham, did you try removing power from the coil packs to see if you get a good signal status on crank? If the coil packs are powered down, the whole HT system is not a noise source, leaving only the starter motor & possibly the alternator. Certainly worth trying before you rewire the whole HT set up.

I suspect the resistor is to suppress back emf from the starter solenoid (big coil), but I’m not sure. You could remove the solenoid wire and use a jumper wire direct to the battery wire, to avoid feeding any emf back to the ECU, as a test.

[PeterK]

When trying to relocate the plug wires, you could just wire in the ''normal' 6, gaining yourself some longer wires for the testing.

[911hillclimber]

Bruce
Yes, tried that yesterday, made no difference.
Peter
Have a little trick to try later
Jonnie
I have no probs with any comment from you about this bloody engine

[Bruce M]

Okay, so for the time being, the entire HT ignition can be ignored. No need to change plug leads, cut up wires or anything related to the coil packs because they are not the problem (currently).

[Bruce M]

Another thought. You try “clip on ferrite” chokes. Cheap from Amazon next day & you probably have at least one or two in your house (used on power cables to sensitive electronics). Although we are back to guessing, rather than measuring.

[911hillclimber]

Had a reply from Clewett to my 4 installation pictures today.
As i expected, he said with this installation I've done I have no chance of it working, and the lot needs to be re-routed.

This means moving the coil pack to the bell housing and feeding the plugs from the clutch end top and bottom. The looms from the coil packs will run along side the carb on the one side and to the current ecu. The sensor lead will follow them.

I will mount the coils far away from the starter, and need to find a way to suppress the wires to it.

i went to the rolling road this morning to discuss the car going there.
Ben there (who tuned it a year ago) agreed the noisy hitorque starter can stop it working, he has had several cars in the shop where the starter noise killed the management signals.

I checked the sensor off the car today.
Set up a bar in my lather with a 6mm bolt screwed into the side of the bar. The sensor would count the times it sees the bolt as the spindle rotates. Meter gave 19.4Hz over about 1 min of running.
The other end of the spindle is a small gear with 19 teeth on it and the same diameter as the bolt in the bar.

Holding the sensor to about 1 mm gap gave 194.6 Hz. The sensor would be counting 19 teeth /rev, the other 1 'tooth' a rev. Lathe speed the same.

I think this shows the sensor is responding, i tried these two conditions several times, and the results were the same give or take a few revs/sec.

Finally, Electromotive have responded, and they will consider all I've summaries and respond early next week.

I've run out of ideas, run out of tests, run out of any additional clues, so this leaves 2 things:
Re-wire
Suppress the starter

Rolling Road BOOKED for 17 March!
No time left, need to take some actions.

[chris68]

Some great advice already from fellow DDK'ers. Here are my thoughts ....

1. Individually check operation of sensors by a process of elimination
2. Don't rip everything apart because it won't necessarily reveal the issue
3. Check power feeds to make sure supply transient dips aren't causing an issue
4. Physically segregate noise sources from sensor wiring as far as possible
5. Use screened twisted wiring to sensors. Terminate screens to chassis at the controller

It might be that you need a scope to 'see' what's going on while cranking the motor. Can bring meters and a scope if you're not too far away.

Don't panic though, something stupid is stopping the system from working. Just need to find it, then fixing should be straight forward. Always seems to be that way with electrical gremlins ..

Cheers,
Chris

[911hillclimber]

Thank you, but ripped it all out.
The sensor (there is only one) is screened and terminated to chassis ground via the ecu ground (I think).
Clewett is convinced it is the proximity of the sensor cable to the rest that is killing it..

I've taken the lot off the engine to start again. Clewett happy with my plan.
If this does not work then I'm in trouble deep.
I'm on the edge of Wolverhampton, not sure if that is too far Chris?

Found some good mounting points to use to mount the weighty coils. I initially thought of this location for convenience but it puts a weight right at the rear of the car hence going to the front of the engine. Note it is far away from the starter as I can.
Fortunately, in the nick of time remembered there is an anti roll bar to go in the same place as the coil pack carrier frame.
Sure is UGLY.

Still, it's not a fast car...

Hope to get a lot done this weekend if Mrs HC is ok.

What a crap project this has turned out to be.