Testing EDIS Coilpacks

[geoffbeaumont]

I'm currently back on the dizzy, as the mystery misfire is considerably less severe on that. This may be no fault at all of the EDIS system, as there's definitely something else wrong as well (petrol will only run in limp home, gas is very lumpy and underpowered).

First job tomorrow is to swop out the lambda sensors and see if that helps (already tried the throttle position sensor, which is the only other one shared by both fuel systems), but I'd like to check out the EDIS system and see if I have got a dead coilpack or something.

Any suggestions on how to go about this?

Only thing I've come up with so far is to get some cheap HT leads (haven't got quite enough for dizzy and EDIS at the same time) and check the EDIS timing on every lead while the engine is actually running on the dizzy. No point trying to do timing with it running on the EDIS, you have to keep the throttle open to stop it stalling

[rtbarton]

The way I checked mine (Punto) was to use an old fashioned neon strobe light, the intensity of the flash depends on the HT voltage.

You can get simple HT voltmeters, they are like a pair of scissors with an inward facing point at each end of the "blades", connected to a plug lead & ground, the points are opened until the spark stops jumping accross them and the voltage read off on a scale on the "handle".

One plug was down, so I swapped the LT & HT leads over and the fault moved to another plug, ie stayed with the faulty coil.

Metering with an ohmmeter didn't show anything.

I suppose the easiest way is substitution.

[Hybrid_From_Hell]

Geoff,

I think it very odd that having swopped back to the dizzy you still have a misfire, as such I think the problem is elsewhere.

Can you explain what exactly the issues are, and what and when it does what ?

Have you check that the throttle pot is correct by ensuring that the accel cable is loose, the butterfly having 2-4 thou clearance when fully close and then set to .325 ?

Is the base engine HW or Flapper

If HW Disconnect at the AFM plug any difference ?

Post up and lets have a forum chew over it

Nige

[TheRecklessEngineer]

I had a similar problem with a ski boat (5.7 chevy underneath). All the coil packs checked fine with a multimeter, but only fired on all 8 consistently after swapping in new coil packs. Some sort of HT voltmeter would have been really useful - it took an entire day to figure that out.

[FridgeFreezer]

You could measure the resistance of each coil, all four should be very similar values.

As Nige says, if the misfire doesn't go away on dizzy then it sounds like a fault elsewhere. EDIS creates a MUCH bigger spark as you know, which creates much more efficient combustion and effectively leads to the engine running very lean (as Nige found out) compared to teh weak spark which leaves more unburnt fuel. This means the misfire could be "masked" on the dizzy setup a bit by the fact that it's effectively running a lot richer on the dizzy than it was on the EDIS. RV8's don't like idling lean so that is probably why it's happier on the dizzy.

In short, could well be a fuelling thing.

[jwriyadh]

I assume you are still running hotwire.

Before you swap out the lambda sensors why don't you put a voltmeter on the lambda sensor signal wire and watch what it is doing? When the sensor is hot enough and the ECU is responding to the lambda signal the voltage should swing back and forth between the rich and weak sides in just less than a second. The signal should swing from 0.5 to 1.0 volts If either is not swinging the ECU will try to react to that (lack of) signal.

Remeber to connect the meter negative to the correct wire, not just earth.

Check that they are both drawing heater current.

jw

[geoffbeaumont]

I think it very odd that having swopped back to the dizzy you still have a misfire, as such I think the problem is elsewhere.

Have you check that the throttle pot is correct by ensuring that the accel cable is loose, the butterfly having 2-4 thou clearance when fully close and then set to .325 ?

Is the base engine HW or Flapper

If HW Disconnect at the AFM plug any difference ?

I'm certain that there is a problem elsewhere, just want to check the EDIS out and make sure there isn't a problem there too.

Hotwire.

Throttle pot checked out okay - haven't checked the butterfly clearance though. The throttle cable is currently tightened to keep the idle speed up enough to stop it stalling on gas. On petrol it will only run at idle with the AFM disconnected, which raises the idle speed anyway.

Can you explain what exactly the issues are, and what and when it does what ?

Post up and lets have a forum chew over it

We already have, in the Range Rover forum.

You could measure the resistance of each coil, all four should be very similar values.

Makes sense - I'll give that a try.

As Nige says, if the misfire doesn't go away on dizzy then it sounds like a fault elsewhere. EDIS creates a MUCH bigger spark as you know, which creates much more efficient combustion and effectively leads to the engine running very lean (as Nige found out) compared to teh weak spark which leaves more unburnt fuel. This means the misfire could be "masked" on the dizzy setup a bit by the fact that it's effectively running a lot richer on the dizzy than it was on the EDIS. RV8's don't like idling lean so that is probably why it's happier on the dizzy.

In short, could well be a fuelling thing.

Hmm, that seems like a very plausible explanation. It would seem to point at the lambda sensors again, since I don't think there's anything else remaining that could have that affect on both hotwire and the LPG kit.

Before you swap out the lambda sensors why don't you put a voltmeter on the lambda sensor signal wire and watch what it is doing? When the sensor is hot enough and the ECU is responding to the lambda signal the voltage should swing back and forth between the rich and weak sides in just less than a second. The signal should swing from 0.5 to 1.0 volts If either is not swinging the ECU will try to react to that (lack of) signal.

That bit isn't in the hotwire diagnostics manual Would you still get that swing if the engine is consistently running on the wrong mixture, though? Surely it will stay at one end of the scale?

The main reason for swopping the lambda sensors out is because I have two spare ones in the garage and it's actually less hassle than checking them out thoroughly with a meter. They've passed the diagnostic manual tests, but as far as I can tell, that only tells you that the heater elements have the correct resistance.

Remeber to connect the meter negative to the correct wire, not just earth.

Already been caught that way - another omission from the manual. I think it was you that pointed it out to me?

Check that they are both drawing heater current.

If swopping them out doesn't work then I guess I'm back to the multimeter.

[TheRecklessEngineer]

You could measure the resistance of each coil, all four should be very similar values.

Just because the resistance is good with a multimeter doesn't mean the coil is good. The insulation breaks down at high voltages leading to a weak spark (see above). The only way to check a coil is with an HV multimeter or by swapping for known good units.

[jwriyadh]

Or fit the green (470 ohm) tune resistor and take the lambda sensors out of the equation.

jw

[FridgeFreezer]

Didn't some RR's have titania O2 sensors not the normal ones? These will not give a 0-1v reading. Normal lambdas read from 0v (massively lean) to 1v (massively rich) and the happy point (14:1 A/FR) is ~.45v

[geoffbeaumont]

Well, changing out the lambda sensors made bugger all difference, so it's time to get the meter out again tomorrow This would all be a lot easier if the Megasquirt was already installed...

Didn't get time to do anything with the EDIS, but from the sound of it I haven't got the right tools to measure anything useful anyway.

Did try and get some diagnostic software talking to the LPG kit, to see if that could tell us anything useful, but no joy. Not entirely sure if the CD I got with the connectors actually contains the right software for my kit, but since Jonathan couldn't even get anything out of it from Teraterm I doubt that's the problem anyway.

[jwriyadh]

Fridge, all LR hotwire systems use titania sensors. They are somewhat more resistant to the LR environment than zirconia types but titania have the disadvantage that they are only available in an M12 thread instead of the the more normal M18, plus titania are resistive not generative so the signal values read vary according to the whims of the ECU circuit designer.

geoff, faultfinding on any closed loop system is a PITA and substitution can lead you up blind alleys. Break the closed loop by taking the lambda sensors out with the green tune resistor so that your engine runs on just the ECU fuel map moderated by temperature, TPS and airflow.

For the full version of the hotwire tests try looking at http://www.users.zetnet.co.uk/barnes_firsnorton/lr_efi/

As fridge would say MS is the best move you could make.

jw

[jwriyadh]

Forgot to mention above, if you fit the green tune resistor to disable the mixture correction the signals on the lambda sensors are still valid, ie, just because you have instructed the ECU to ignore them the sensor output still tells whether the mixture is weak or rich (if they are working correctly).

More help maybe, http://www.roversnorth.com/web/downloads/r...P65_pg36_37.pdf

Found by trawling the net for "14CUX"

jw

[geoffbeaumont]

Okay, the problem with the LPG diagnostic link was that the software can't cope with Windows virtual COM ports - tried it under DOS today and it connected up fine, and immediately revealed a fault which we thought was going to be the answer (you can already see where this is heading, can't you ) - an intermittent connection to the lambda sensor (the LPG kit only uses the right hand one). We traced this to some dodgy wiring where the LPG kit has been bodged in - bodged being the operative word, the gas and coolant pipework is nice and tidy, but the wiring leaves a lot to be desired...

With that fixed we now have a nice steady signal from the lamda sensor (at least to the LPG kit - didn't get time to check at the 14CUX end), but...still a misfire. It's better than it was though, so reset the ignition timing now it's got a more stable idle (it was too far advanced still). It doesn't need the throttle cable adjuster overtightened to stop it stalling on LPG now, but still won't idle with the AFM connected on petrol, and both are still very rough at all revs.

Sniffing the exhaust it's definitely running very lean on petrol when the AFM is connected - barely any smell at all to it.

I guess next step is to try jwriyadh's suggestion and take the lambda sensors out of the equation. If that doesn't work, I guess invest in a compression tester and see if there's something up mechanically, like a worn camshaft.