Missfire / Running Issues - One For The Uber-Techies

[PW@Woodsport]

er,not the HT leads then.... oh well,G in that case you need to retrace every single thing you touched when doing this job,im 99% sure its something you have touched seeing as the problem was not there before the cambelt/dizzy change.

[System-G]

I found when putting the distributor back in......

it may seem aligned, but unless the cap is off and you can keep an eye on the rotor arm remaining where it should be, you can't tell that as the cog engages the cam shaft bevelled gear, it turns the distributor rotor arm out of alignment.

my tuppence worth of useless info.

On an SC, it's near impossible to put the dizzy between the engine and firewall/bulkhead without the rotor being knocked with the cap still on - the space is much less than on an N/A or N/A converted to SC. I think the engine sits further forward on the SC chassis.

[JMR_AW11]

You could try checking all electrical grounds for low resistance with a meter..

eg if the engine/chassis ground points are OK.

[PW@Woodsport]

if the engines cranking then the earths are A1 as thats the biggest draw ampage wise you can place on the engine earths.... so its not that.

[JMR_AW11]

if the engines cranking then the earths are A1 as thats the biggest draw ampage wise you can place on the engine earths.... so its not that.

There are various separate ground return paths to the ECU and ignition components on most engines (I've never looked at an SC)

These are separated to avoid crosstalk and therefore they all need to be checked or the engine sensors could start misreporting their readings.

I don't know how the wiring loom is arranged on an SC but I've seen a thread on MR2OC where someone with a tubby had poor running after forgetting to refit a ground link wire after engine work.

The car started fine but ran poorly.

He found the fault by measuring the DC resistance of all the various ground paths from the ECU and the engine. (after seeing that the grounds measured at different voltages when the car was running!)
IIRC he'd forgotten to fit a wire to ground the manifold.

It's probably not the grounding but SG sounds desperate so it might be worth bleeping out the ground paths with an ohmmeter.

[JMR_AW11]

Just for reference... The other way to potentially screw up the system is to randomly add extra ground wires from the stock ground return wires to various parts of the engine/chassis (causing sensor/ignition/injector crosstalk issues)

But I guess this doesn't apply in this case.

BTW (going off topic a bit)
I did once look at the Megasquirt ECU circuit board design and was amazed to see they didn't separate the various ground feeds to and from the sensors and ignition and injectors. The ECU had one common ground point for all these feeds at the ECU connector. Not good for avoiding crosstalk on the vital sensor readings (where even a few millivolts of crosstalk can adversely affect the ignition timing or fuelling)

Toyota ECUs have separate return paths for these grounds but this does have the downside in that they all have to have strong connections!

[System-G]

We had the Horn ground re-wired for an MOT not long ago, but they cocked that up and we were i the process of re-wiring it at the same time as the dizzy/cambelt change. But stopped due to the more pressing issue that we are trying to resolve. I doubt very much the horn wiring would influence idling/misfire

Any way some progress from last night.
I tested everything - AFM, Oxy snesor (although I only have a digital multimeter and it didn't behave as I expected), both Dizzys, both coil/ignitor packs etc... everything in spec

Put it all back together and tried started the engine, massive backfires was all
I figured that I was either a couple of teeth out on the dizzy or (more likey) hadn't checked the engine was at TDC before inserting dizzy.

Getting peed with all the backing & frowing, I decided to purposely put the dizzy a tooth out.
The engine started first time and the missfire/hesitation was not there
However, I dared not even check the timing just yet. For the first time this week I wanted to end my evening on a high and not feeling depressed and utterly useless.

I'll pick up on it again today

[PW@Woodsport]

Jeremy,my point was that all of the earths are present and correct,given that the separate earth paths are all doing their job.... there's a proper spark so the ignition system is earthed,and its injecting fuel,so the ecu earth paths are all correct.

So its not that.

Any sort of misfiring condition is down to bad timing be that incorrect camshaft timing or ignition timing in relation to that cam timing.If both cam timing and distributor are correct then the signal from distributor to ecu becomes suspect.

The only time poor earthing will account for a misfire is if the amplifier doesnt have a good ground,but we have established it has and theres a good strong spark.

[JMR_AW11]

Jeremy,my point was that all of the earths are present and correct,given that the separate earth paths are all doing their job.... there's a proper spark so the ignition system is earthed,and its injecting fuel,so the ecu earth paths are all correct.

So its not that.

Any sort of misfiring condition is down to bad timing be that incorrect camshaft timing or ignition timing in relation to that cam timing.If both cam timing and distributor are correct then the signal from distributor to ecu becomes suspect.

The only time poor earthing will account for a misfire is if the amplifier doesnt have a good ground,but we have established it has and theres a good strong spark.

I agree it's unlikely to be a grounding fault given the history, but you still haven't said anything to fully rule out a sensor grounding issue.

The sensors usually* have a separate ground path (to prevent crosstalk)
IIRC this is the E2 (E22?) connection on the ECU.

The ignition connections you refer to are related to the engine ground.
This is the E1 connection on the ECU. If this connection is healthy it doesn't mean the E2 sensor ground wire is OK.

There are other ground connections to the ECU eg for main power and the injectors.

Also there is the ground connection from the ECU case to the car body. This one would only fail if the ECU was unbolted from the chassis.

These grounds are all arranged like this to prevent crosstalk and also to protect the ECU from EMC/ESD problems.

From what SG is saying now it's probably a timing issue but it's still worth bleeping the various E1 E2 E22 E01 E02 ECU ground paths to make sure the sensors aren't riding on a false potential.

This could screw up the ignition timing and the fuelling if the AFM and coolant/air sensors started giving voltage readings with an offset. This is what happened to the tubby owner on MR2OC.

*I've never seen an SC wiring loom so I'm assuming that the loom will follow the usual (sensible) pattern of separate ground paths for igntion/sensor/injector signals. The ECU connector pinout suggests they adopt this approach.

[JMR_AW11]

Hope nobody minds me posting this but I had a quick look thru the ECU ROM code for the 89 JDM SC ECU wrt the service ignition timing (in case there was something odd that might screw up the 10deg service timing setting)

This mode is activated if you bridge the link from E1 to T.


What may be of some interest is the fact that the ECU can override the 10deg BTDC service timing even with the T to E1 link fitted.

This can happen if the car is moving (there's a speed sensor test in the ROM code just before it selects the 10deg BTDC fixed timing)
and also if the ECU thinks the throttle is pressed. (IDL contact test)
It instantly reverts back to the service timing once you lift off the throttle so this effect is not 'sticky'

I think it's safe to assume you are not testing the timing whilst driving but it might be worth checking the TPS IDL contact is functioning OK before trying to set the car into service mode.

Mind you, I guess you would already get an ECU error code if the IDL switch was faulty...

NOTE: It also appears the ECU can select the fixed 10deg BTDC ignition timing even without the T to E1 link.

This can happen if the rpm falls very low (eg 300rpm) or if the ECU is receiving the STA signal during engine cranking/starting.

[JMR_AW11]

The other thing to note is that the ignition timing in service mode doesn't care about the coolant sensor or the air sensor or the AFM reading.

As long as the T to E1 link is fitted and the throttle is closed (IDL OK) and the car is not moving then the timing *WILL* be set to 10deg BTDC according to the ECU so it should be safe to rotate the (correctly fitted) dizzy to give 10deg on the timing light and then take it from there.

What happens to the idle timing (on the timing light) when you take it out of service mode? Maybe this could give a clue.

Can you measure the DC voltage at the THW and THA and AFM sensors at idle when warm?

Is there a voltage difference if you measure at E1 and E2 (at the sensor end) wrt chassis ground with the engine running?

[greglebon]

Looking at the ECU wiring diagrams, and inside the ECU, the E1 and E2 grounding pins are internally connected.

E2 is the ground for the sensors, and this ultimately comes out of the ECU through E1.

E01 and E02 are the ground plane for the high-power transistors, etc and are on a separate ground plane from the E1 stuff... : there is a suppression diode which will ensure continued ECU operation if E1 is disconnected, by grounding E1 / E2 through the noisy earth if required, but the same does not apply the other way...ie: if E01 AND E02 are disconnected, you get zero fuel...period..!

However, they all common at the same ground point on the loom.

This point generally seems to be either the engine block, or inlet manifold...... (across several Toyota models.......)

HTH.....?

[JMR_AW11]

Looking at the ECU wiring diagrams, and inside the ECU, the E1 and E2 grounding pins are internally connected.

E2 is the ground for the sensors, and this ultimately comes out of the ECU through E1.

E01 and E02 are the ground plane for the high-power transistors, etc and are isolated inside the ECU from the E1 stuff...

However, they all common at the same ground point on the loom.

This point generally seems to be either the engine block, or inlet manifold...... (across several Toyota models.......)

HTH.....?

Yes, also the ECU case ground is usually not connected to its E1 or E2 pins or E01/E02 if the ECU is removed and I assume they reference to some 'cold' point on the engine like the manifold.

What happened to the tubby owner was that the ground wire had lost this connection at the manifold. He'd disturbed the connection after a headgasket swap IIRC.

It meant all his sensor voltages were offset by something like 0.3V once he started the engine.

[greglebon]

Just had a look at a few diagrams and the ECU I have here..

Cheers, J...did not realise that the ECU case is not actually grounded with any of the above circuits.....its sole earth is through the mounting bracket.....

So, my setup with ECU fixed to the rear panel in the boot with VELCRO means my case is not earthed at all.......

Even though it all works fine, I'll sort this out.....

I'm not sure exactly at what point on the loom the 3 main grounds are commoned, but it seems to me that joining them as soon after the ECU plug as poss and running a nice meaty cable to the engine block is the optimal......?

Is this how Mr.T does it, or are they joined at the block end?
My diagram does not show this..?

[JMR_AW11]

Just had a look at a few diagrams and the ECU I have here..

Cheers, J...did not realise that the ECU case is not actually grounded with any of the above circuits.....its sole earth is through the mounting bracket.....

So, my setup with ECU fixed to the rear panel in the boot with VELCRO means my case is not earthed at all.......

Even though it all works fine, I'll sort this out.....

I'm not sure exactly at what point on the loom the 3 main grounds are commoned, but it seems to me that joining them as soon after the ECU plug as poss and running a nice meaty cable to the engine block is the optimal......?

Is this how Mr.T does it, or are they joined at the block end?
My diagram does not show this..?

TBH I'm not the right person to answer that question as I don't ever mess with wiring looms (my knowledge really stops at the ECU)

But they do join together once all the wiring loom is fitted and the ECU case is screwed to the chassis. If a connection is lost somewhere in all this then it's possible to get small potential differences once the engine starts and this could upset the sensors (in theory at least, and the guy on MR2OC seems to have suffered from this)

With my ECU daughterboards I have to make sure that the ground on my board doesn't ever touch the ECU case ground. So all mounting screws are isolated from the groundplane on my board. This prevents a direct connection from the sensitive ECU inner grounding and the ECU case.


Having said all the above, the fault still sounds like an ignition problem even though so many things have been ruled out.

Could the spark plug outer insulator/body be dirty on a couple of plugs? swapping the HT leads over wouldn't cure this as the spark could leak down the side of the plug body.

[greglebon]

Yeh...grounding can be fun...

IMHO, though, you cover all bases by making sure that the engine block is well earthed to the chassis, the chassis is well connected to the battery, and the ECU is well connected to the block......

(Also, on my conversion, there was the small matter of the metal bracket containing the igniters requiring grounding to the block.......but thats another story... )

In fact, you can add igniters-to-block to the above list........

[JMR_AW11]

Yeh...grounding can be fun...

IMHO, though, you cover all bases by making sure that the engine block is well earthed to the chassis, the chassis is well connected to the battery, and the ECU is well connected to the block......

(Also, on my conversion, there was the small matter of the metal bracket containing the igniters requiring grounding to the block.......but thats another story... )

In fact, you can add igniters-to-block to the above list........

Agreed but it's the E2 sensor ground wire that is the most sensitive to crosstalk.

That's why I said earlier that it's dodgy to add extra grounds in certain places. IMO the E2 sensor ground return wire should be a clean path straight back to the E2 pin on the ECU. If it had a connection to another ground halfway along then I would have thought it would be prone to transients/crosstalk.

I suppose I could go out and look at my NA mk1 to see how the E2 wire is grounded.

I guess on many of these cars the wiring will be non standard because of aftermarket alarms and immobilisers but these could cause problems if they connect to the E2 wire.

[greglebon]

Hmm..I don't get that....!

Assuming all the main grounds are tied together solidly, then they will all be at the same potential..?

Technically, you could ground E2 at the same point on the block as the rest, as long as they're all the same, it matters not......

E2 is connected to E1 when it enters the ECU case, anyway.....

[JMR_AW11]

Hmm..I don't get that....!

Assuming all the main grounds are tied together solidly, then they will all be at the same potential..?

Technically, you could ground E2 at the same point on the block as the rest, as long as they're all the same, it matters not......

E2 is connected to E1 when it enters the ECU case, anyway.....

In theory if you replaced the wiring loom with ultra ultra heavy guage wire you could probably keep all the ground potentials within the acceptable noise/error limits of the engine sensors.

But that adds a lot of weight. Ever felt how heavy a stock skinny wiring loom is?

FWIW on a relatively stock loom I think it does matter that the E2 wire is isolated from other ground wires.

The ECU needs to measure the sensor thermistors and the AFM very accurately in order to prevent glitches in the fuelling and ignition.

So I would expect (and I could be very wrong) that the E2 ground return will live on its own quiet little wire from the ECU.

The reference point for measurement for the ECU is its own inner groundplane. That's all it knows. This contains E1 and E2 as you say but it is a large groundplane and there are no large varying currents to/from the ECU that connect to this plane. (i.e. the injectors don't return their ground to this plane) So as long as the sensor ground potential tracks with the ECU groundplane potential then accurate and low noise measurements can be made.

The ECU PCB groundplane potential is the one that really matters for sensor measurement.

Example:
The ECU needs to measure the THW sensor resistance to measure the coolant temperature.

It does this by using the THW sensor as a variable pulldown resistor and measures a voltage change to predict the THW resistance.

Therefore it needs to know the voltage directly across the THW sensor.

It has to assume there is a very small and consistent volts drop in the E2 wire or the THW measurement will be wrong if the far end of the E2 wire is at a different potential to the ECU groundplane.

The THW sensor is on the end of a long piece of wire and if anything else was connected to this wire (other than the other stable low current sensors) then it could cause a volts drop or a glitch in the E2 wire.

This would cause an incorrect reading of the THW sensor.

it's even more critical for the AFM.

Eg if someone fitted an extra cooling fan and grounded it to the E2 wire then it would cause havoc with the sensor readings every time it switched on.

(because of the volts drop in the E2 wire when the fan came on)

Also, if someone linked the E2 wire to another part of the loom or engine/chassis metalwork then the E2 wire would lose its uniform potential along its length.

This is because the E2 wire would current share with the E1 wire and would have more noise across it due to a less uniform potential along its length. Not good for accurate sensor measurement.

The simplest and lightest solution to this is a unique E2 wire that only goes to the sensors.

Otherwise, if you join everything together you would have to do it with wires so thick that there was virtually no volts drop anywhere across the whole car engine/chassis even with the injectors pulsing. i.e. just a couple of millivolts drop anywhere in the grounding. Impractical IMO.

The separate E2 ground wire is much easier to fit.

At the end of the day the key issue is to ensure that the potential at the ground end of all the sensors is the same as the potential at the ECU inner groundplane where the ECU measures the sensor voltages.

This is achieved with the separate E2 wire as long as nothing else connects to it that can be at a different potential to the ECU groundplane.

See the image below for how I think the E1 and E2 grounding will be arranged. E2 only goes to the sensors, not the engine ground or any other ground returns. (to avoid crosstalk)

Of course, all this could backfire on me if actually Toyota wire their sensors in a different way , but the way I've described it makes the most sense.

[greglebon]

Gee...this IS getting a little TOO technical......

Bottom line is, and getting back on post, E1 and E2 are at the same potential: I suppose if someone grounded the loose end of E2 in your diagram to a dodgy earth point, then you could get interference / ground loop wierdness going on..?
The E2 connection on the ECU is an INPUT: all the sensors round to a common point in the ECU, and it is this common which EXITS the ECU via E1 to ground.

None of this is really relevant to the OP, though, is it......

I reckon we've covered enough ground (pardon the pun! ) about earthing.......?