Distributorless ignition and Rover 14CUX EFi

[DaveEFI]

With the 4CU, Lucas did a 'Fast Check' box. You removed the ECU, and plugged into that. It generated a test ignition pulse. Its purpose to test the ECU in a dealer's workshop, without special electronics knowledge.

I adapted an MS Stim to do the same thing with the 4CU. It generates a low volt ignition (tach) pulse, and the 4CU read that happily.

From what I've seen of the hotwire wiring diagram, it uses the same 4K7 series resistor between coil and ECU (to limit the current from the coil higher voltage pulse) as the 4CU. Which suggests to me both the flapper and hotwire ECU use the same signal post that series resistor.

Because a car electrical system voltage varies so much (9-15v or so) most critical electronics operate off an internal regulated 5v supply - both the Nodiz and Lucas ECU do. Meaning any signal between them is going to be a low voltage one.

It's possible, of course, that there is a fault with either the tach output on the Nodiz or the tach input on the hotwire. You'd really need a 'scope to investigate both.

[DaveEFI]

Check your clocks, some will have a little potentiometer on them where you can adjust the rpm.

I know one of my old Ford clocks did.

Or if you're using Rover clocks...maybe hunt for a Rover tacho from a 4cyl car ?

Or this is a simple pulse adjuster ( I use it for my speedo...well the older DIY version. )

https://www.jaycar.co.uk/speedo-correct ... medium=web

Since we're talking a 50% error, it's likely down to seeing a wasted spark pulse. Which is half that of a dizzy. Most after market ECUs can double this wasted spark pulse (selectable in software) to give the correct frequency to drive an external ECU or rev-counter, designed to work off a single coil.

[stevieturbo]

Check your clocks, some will have a little potentiometer on them where you can adjust the rpm.

I know one of my old Ford clocks did.

Or if you're using Rover clocks...maybe hunt for a Rover tacho from a 4cyl car ?

Or this is a simple pulse adjuster ( I use it for my speedo...well the older DIY version. )

https://www.jaycar.co.uk/speedo-correct ... medium=web

Since we're talking a 50% error, it's likely down to seeing a wasted spark pulse. Which is half that of a dizzy. Most after market ECUs can double this wasted spark pulse (selectable in software) to give the correct frequency to drive an external ECU or rev-counter, designed to work off a single coil.

Well presumably his cannot, otherwise he would not be asking.

[GDCobra]

Check your clocks, some will have a little potentiometer on them where you can adjust the rpm.

I know one of my old Ford clocks did.

Or if you're using Rover clocks...maybe hunt for a Rover tacho from a 4cyl car ?

Or this is a simple pulse adjuster ( I use it for my speedo...well the older DIY version. )

https://www.jaycar.co.uk/speedo-correct ... medium=web

I’m not really too concerned about the tachometer, that issue is solvable, I’m more concerned about the signal to the EFi. Without this in place the 14CUX will either not run at all or nor run properly so I’m dead in the water unless I can sort that.

Thanks for the link to the pulse adjuster, that may also be useful to correct my road speed reading.

[stevieturbo]

So exactly what did the EFI get a signal from originally ?

Did it not have a crank trigger ? Or did it just take a signal from the dizzy ?

If the dizzy is still in place, then perhaps best route is to leave it's internals intact to retain that signal ?

[DaveEFI]

It gets the tach pulse needed for the coil negative. And because the coil is an inductor, that pulse measures at something like 50v. Using a dizzy just to generate a pulse for the OEM ECU would be silly, as the Nodiz already does just this - it couldn't produce a suitable spark without it. But that pulse has a smaller amplitude - about 12v.

My view is something is broken, or incorrectly wired. But so see what a pulse is doing requires an oscilloscope.

[GDCobra]

Unfortunately when I made my last resonse I hadn't realised a new page had started and there were more posts so I'll try and anwere the questions raised here and also clarify the situation.

There are 2 ways I can get an "ignition event"signal, either from the coils being triggered or from the ignitition units TACHO signal.

Lets take the TACHO signal first as it's the easiest to eliminate.
When I feed this signal to the tachometer either directly or via a relay-coil/transistor circuit the meter shows half the correct reading indicating this is giving 2 signals per engine revolution.
The beta software which I origninal used (now removed) had buttons to set the TACHO output per rev' to 1, 2 or 4 however this button did not work, pressing it would result in the communications between PC and NODIZ failing. There appears to be no other way to set the TACH pulse/rev' output. I was hoping that setting the trigger wheel to "36-1 8 cylinder" would assume a 4 pulse/rev' output but it does not.
It seems the ignition unit generates it's TACHO signal via the RPM calculaiton from the trigger wheel NOT from its coil triggering routine, which would be fine if the amount of pulses could be controlled.
I also tried feedinng this signal to the 14CUX both directly and via the relay-coil/transistor circuit (and both with and without the 6K7 resistor in line), this gave no respone, not even a cough from the engine and the fuel pump did not re-start (which it normally will as soon as it sees a valid input on that line). I don't know if this is because the pulse in not of the correct form or whether the 14CUX requires a minimum frequency which is not being reached due to only running at half rate.
All of this is irrelevant though, as the signal is only 2 pulses/rev' it is never going to run the EFi correctly, the pulse adjuster suggested by Stevie may be useful but looks like it only goes up to 99%, I'd need 100%, so don't think this would work although it may be possible to do something similar.


So what about using the coil triggers. Obviously changing from single to multiple coils it is necessary to sum the trigger pulses from the coil -ve together to get a single, 4 pulse/rev' signal, I originally did this using 4 regular diodes and a zener, the signal was good enough to allow the 14CUX to run and also gave the correct reading on the tachometer. However when I ran the car under load I kept getting a major "stumbel", felt as though the engine had been turned off for a split second.
I reinstated the distributor & single coil setup, sending the HT output to ground and using the coil -ve to trigger the 14CUX (which worked), at this point keeping the diodes linked to the tachometer. I still got the "stumbe".
This demonstrates that the problem is with the ignition not the EFi (the engine ran cleanly from the single coil and EFi is now getting same signal).
I removed the diode setup completely from the loom, now the engine ran better, still an occasional, minor "stumble" (which we are currently putting down to 'noise' on the MAP signal) but nothing like as bad.

As my only chance of success with this is to use the coil firing signals I'm going to try the diodes again but this time separate the output from the diodes from the ignition loom (take it out of the opposite side of the coil pack) in the hope that this will eliminate any interference between this signal and the wiring between the NODIZ and the coils.

Questions
From DaveEFi - "It's possible, of course, that there is a fault with either the tach output on the Nodiz or the tach input on the hotwire. You'd really need a 'scope to investigate both."
I'm happy that the input to the 14CUX is fine as it works perfectly with the distrubutor system. There is a problem with the NODIZ TACHO signal, it runs at 2 pulses/rev'.

Also from DaveEFi "Since we're talking a 50% error, it's likely down to seeing a wasted spark pulse. Which is half that of a dizzy. Most after market ECUs can double this wasted spark pulse (selectable in software) to give the correct frequency to drive an external ECU or rev-counter, designed to work off a single coil."
The 50% frequencey error is on the TACHO output from the NODIZ, I don't believe this is generated from ignition events, more likely created as a fucntion of the RPM (monitored from the trigger wheel) and output at 2 pulses/rev' hence nothing to do with wasted spark (which cumulatively has 4 coil triggers per revolution same as the distributor setup, only difference is those pulses are sent to individual coils rather than the same one.). Unfortunatly the "pulse/rev' value" does not appear to be controllable.

From Stevie "So exactly what did the EFI get a signal from originally ? Did it not have a crank trigger ? Or did it just take a signal from the dizzy ?"
The -ve terminal of the single coil - A pulse for every ignition event. No crank signal.

From Stevie "If the dizzy is still in place, then perhaps best route is to leave it's internals intact to retain that signal ?"
I have currenlty been running the car in exactly this way, with the HT output being sent to ground via a spark-plug. This is only for testing, I'm not happy to run that route for any length of time partly due to this high-energy spark and any possible interference it may cause but also due to still having the (full) distributor which I want to remove.

I'm hoping to get the NODIZ back this coming week and will report back if I get better behaviour with my re-routed diodes (or if the re-flashing helped), meanwhile I wish Geordie Jim better success than I'm currently having.

[GDCobra]

It gets the tach pulse needed for the coil negative. And because the coil is an inductor, that pulse measures at something like 50v. Using a dizzy just to generate a pulse for the OEM ECU would be silly, as the Nodiz already does just this - it couldn't produce a suitable spark without it. But that pulse has a smaller amplitude - about 12v.

My view is something is broken, or incorrectly wired. But so see what a pulse is doing requires an oscilloscope.

This is part of my problem, I don't know what the original pulse looks like.
I have seen indication of this pulse being several hundred volts, don't know how true that is.
Why would a pulse taken from the single coil used to fire 8 cylinders be different to a pulse taken from a coil used to fire only 2? I know it's firing only 1/4 times as often (hence suming all of them using diodes) but I'm just wondering obout the individual pulse 'form'.

I'm (reletively) happy the wiring is OK as the system works when the diodes are disconnected.

I wish I had a silly-scope!

[DaveEFI]

Can you remind me of which Nodiz version you have? According to the installation instructions I found online, it only produces the correct number of pulses on the tach output to drive the OEM rev counter. In other words, if the device is configured for an 8-cylinder engine, (and works OK with that) the tach output will be correct.

My scope is a large old mains one, and lives in the workshop. So I've not actually looked at the pulse from a single coil with it, and no longer can anyway. However, I have looked at the pulse you get from one of those relay coil based pulse amplifiers. And that is about 50v.

Thing is, it seems nonsense to me to amplify the pulse from the Nodiz merely to attenuate it afterwards with that 6k8 inline resistor.

You do, obviously, need the correct number of pulses (8 every two engine revolutions) to the Lucas ECU, as engine speed is one of the parameters used to set the fuelling.

BTW, you can check for a pulse present with a small LED. With a say 1K series resistor for a 12v pulse. At low revs you will see it flash with each pulse. At higher revs it will look to be on continuously.

[stevieturbo]

Really we need to know exactly what signal the OEM injection ecu expects or needs.

Regardless of what voltage spike the coil trigger might see....it is unlikely it needs to always need a high voltage trigger. It may just need 5v, 10v, 12v, whatever.

It may simply be a grounding/open it needs for the injection ecu ? And not a high voltage spike at all. I assume you were using a relay coil to try and get a back emf spike ?
Perhaps just try using a relay as a a simple set of points for a test and ground switch for the injection ecu ? rather than using the coil itself

But that trigger detail needs to be known. Electronic setups tend to be different where a coil charge will indeed be a ground event, but a stop charge will often be a 12v applied as opposed to open circuit of a mechanical switch. ( ie, 12v ends up on both sides, no potential difference = off state. )

Although really, there should be no issue leaving a single coil in place and just grounding the output. It's not ideal, but it is workable and would let you scope the neg trigger side to see what is happening and if then playing with signals into the ecu get something workable

Or as said form the start....I'd have opted for a full injection ecu, which although more work, would at least leave you in the position you could go full aftermarket now, rather than half and half.

Are there any workshop manuals giving details on the ignition/injection system ?

[stevieturbo]

file:///C:/Downloads/RANGE%20ROVER%2014cux%20Fuel%20Injection%20system.pdf

This seems to suggest the injection ecu gets a 12v signal from the coil trigger. Which seems plausible.

file:///C:/Downloads/1314CU_14CUX_Systems%20info.pdf

this suggests it may seem 12v at the coil, but is reduced to 7v for the ecu. This makes a lot of sense, as most ecu's do not like high voltages.

Hook up a relay for now ( baring in mind switching limitations ) and rig it so your Nodiz tacho output fires it at 8 times per cycle as the rev counter would have had, but simple wired so it sends 12v through the contacts to the injection ecu, switching 12v, open, 12v, open etc etc. ( or via the 6800ohm resistor to drop to 7v )

If it seems to work then you could replace the mechanical relay with a simple pnp transistor that can handle any switching speed.

Lot of info here....but not of use to this problem from a glance.

http://www.g33.co.uk/pages/technical_fu ... ction.html


or again if the Nodiz is sending out 8 pulses per cycle to ground, hook it up to the 12v line via a 6800ohm resistor as a pull down which should then give you an 8 pulse 7v trigger for the ecu

[DaveEFI]

As I said before, I know of a couple of SD1s running the ignition only MegaJolt. And the tach signal from that drives its Lucas 4CU with 6k8 resistor removed happily. So safe to assume the 14CU expects the same signal, since it uses the same 6k8 resistor.

Nodiz don't seem to publish a schematic for their unit. So can't see how they derive the tach output. It's just possible it needs a pull up or down resistor, but surely the Nodiz supplier has experience of this setup?

Grounding is also important. Both the Lucas and Nodiz should have the same ground point. And unless they are very close together, I'd also use screened cable between them.

I'd advise against connecting a relay to the Nodiz pulse output. I very much doubt it is designed for that job. (MegaSquirt uses dedicated drivers for relays, unlike the other ports you can use to feed external ECUs, etc) But by all means try a switching transistor.

[GDCobra]

Can you remind me of which Nodiz version you have? According to the installation instructions I found online, it only produces the correct number of pulses on the tach output to drive the OEM rev counter. In other words, if the device is configured for an 8-cylinder engine, (and works OK with that) the tach output will be correct.

It is a NODIZ Pro Gen 2.
The problem here is there is no such thing as an OEM rev' counter wtih respect to this car as the vehicle is made up from a collection of disparate parts (including the NODIZ) not a production model same as a few thousand others which left the same factory to a production specification. In any case the tachometer is not the important thing, its only relevance here is the diagnostic information it gives me. It has demonstrated that the TACHO signal is 2pulse/rev', This is no good to run the EFi so not really much point in trying to do anything further with it.

On my car the tachometer, and more importantly the EFi get a pulse from the -ve terminal of the ignition coil hence 4 per crank revolution so that is what I need to replicate for the EFi.

My scope is a large old mains one, and lives in the workshop. So I've not actually looked at the pulse from a single coil with it, and no longer can anyway. However, I have looked at the pulse you get from one of those relay coil based pulse amplifiers. And that is about 50v.

As I mentioned above this in one area where I'm completely in the dark, from what I've read the signal is going to be a "spikey" form peaking at several 10's or hundreds of volts. The only information I have found is on a diagram here:

http://www.g33.co.uk/pages/technical_ig ... ystem.html

About 1/4 way down the page, however this is described as a "typical" signal not specific to the ignition system I'm using so amplitude may differ but presumably the form (with an initial spike) should be similar. It seems possible/probable that a smaller (relay) coil would produce a smaller amplitude than a large ignition coil but this is outside of my area of expertise.
I'm assuming that my "relay-coil & transistor" circuit is giving a similar pulse form from the 12V TACHO square wave but neither gets a response from the EFi system, however this could be due to speed. Obviously I'm not expecting this to run the engine (at least not correctly) as I know the frequence is wrong but it does not even re-start the fuel pump but this could be due to the program requiring engine speed to be above a specific threshold, as its only seeing half RPM this may not be getting reached.

Thing is, it seems nonsense to me to amplify the pulse from the Nodiz merely to attenuate it afterwards with that 6k8 inline resistor.

That may well be correct but at the moment I'm trying to keep this simple by simply replicating the original output, that is the flyback from a coil -ve, OK so I'm summing the signal from 4 coils now via diodes rather than taking it from a single coil but that should only be a detail change.
It's also worth bearing in mind that when triggered in this way the EFi is working, the problem is in the ignition system. I've proven this by running the 14CUX from the 'old' ignition system with it's HT output "thrown away", the problem still existed.
At that point the diodes were still in the circuit and driving the tachometer. When the diodes were snipped from the loom the problem was removed.

[GDCobra]

Really we need to know exactly what signal the OEM injection ecu expects or needs.

Regardless of what voltage spike the coil trigger might see....it is unlikely it needs to always need a high voltage trigger. It may just need 5v, 10v, 12v, whatever.

It may simply be a grounding/open it needs for the injection ecu ? And not a high voltage spike at all.

Unfortunately I have no way of knowing what signal the EFi needs wtihout investigating its signal conditioning circuitry which I don't currenlty have the capability to do.
What I do know is that in its standard configuration it gets this signal from the -ve terminal of the single coil which I'm told is a "high" voltage "spike" type signal form, similar to the diagarm shown half way down this page: http://www.g33.co.uk/pages/technical_ig ... ystem.html

In which case it is a voltage signal of relatively high magnitude

I assume you were using a relay coil to try and get a back emf spike ?
Perhaps just try using a relay as a a simple set of points for a test and ground switch for the injection ecu ? rather than using the coil itself

Yes. I did the following.
Wired one side of the relay coil to +12v, the other side to ground via a transistor. The base of the transistor was wired to the TACHO output of the NODIZ.
The "output" was taken from the junction of the coil and the transistor. This was successful in running the tachometer.

The diagram can be seen on page 3 of this thread "viewtopic.php?f=11&t=15611&start=30", 7 posts down.

This is no good for firing the EFi as the frequency is wrong. However what you have suggested has given me an idea. I could substitute this in my distributor system in place of the large ignition coil and see if that runs the EFi.
OK, I still wouldn't want to run with that full time but if it works I would at least remove the worry of having this large HT pulse around which I currently have and would also prove out if the circuit I'm made up is "good" for the EFi.
It's possible that the output voltage may be lower that with the larger coil which would mean I may have to reduce or remove the 6K7 resistor from the input, but I see that as a good thing anyhow.
This would all be good but ultimatly of limited use unless I can get a 4 pulse/rev' low voltage signal from NODIZ.
Although there would be the possibility of replicating that in miniature to create a "crank pulse generator" as a separate entitiy. I'll do some thinkning on that!

Although really, there should be no issue leaving a single coil in place and just grounding the output. It's not ideal, but it is workable and would let you scope the neg trigger side to see what is happening and if then playing with signals into the ecu get something workable

I'm OK with that in the short term, particularly if I can replace the ignition coil with a relay coil as described above but not really what I want in the long term.
Although I accept that, in the words of the great prophet Jagger, you can't always get what you want!

Or as said form the start....I'd have opted for a full injection ecu, which although more work, would at least leave you in the position you could go full aftermarket now, rather than half and half.

I'm not really sure that this is the right time for "I told you so" (:-)), and I don't really have any intention of changing the EFi system (which works perfectly well) so I'm not sure I would be better off, I may find that whatever system I chose would not give me a usable, 4 pulse/rev' output, when I initially investigated NODIZ I did not expect this to be a problem.

Are there any workshop manuals giving details on the ignition/injection system ?

I have manuals (and a lot of additional information) on the 14CUX and a manual for the NODIZ however there is nothing which covers both systems.

Let's see how things go with the re-flashed unit and modified diode wiring routing, the problem may simply vanish.

[GDCobra]

file:///C:/Downloads/RANGE%20ROVER%2014cux%20Fuel%20Injection%20system.pdf

This seems to suggest the injection ecu gets a 12v signal from the coil trigger. Which seems plausible.

file:///C:/Downloads/1314CU_14CUX_Systems%20info.pdf

this suggests it may seem 12v at the coil, but is reduced to 7v for the ecu. This makes a lot of sense, as most ecu's do not like high voltages.

Unfortunately it looks like those links are to a location on your C drive so I don't see them, which is a shame as they sound interesting!

Hook up a relay for now ( baring in mind switching limitations ) and rig it so your Nodiz tacho output fires it at 8 times per cycle as the rev counter would have had, but simple wired so it sends 12v through the contacts to the injection ecu, switching 12v, open, 12v, open etc etc. ( or via the 6800ohm resistor to drop to 7v )

If it seems to work then you could replace the mechanical relay with a simple pnp transistor that can handle any switching speed.

That's pretty much what I have done but the problem is I can't do anything with NODIZ to modify the tacho output per rev'. That is ultimately the issue I have with that output.

Lot of info here....but not of use to this problem from a glance.

http://www.g33.co.uk/pages/technical_fu ... ction.html

I'm quite well aquanted with Mark Blitz's site, I think anyone doing any work with a 14CUX is (or should be), he's a top bloke and done a lot of good work with the RoverGauge system, he makes the cables to link the diagnostic socket to a PC.