poor running, misfire?

[GreenV8S]

honestjohntoo said:

No power = no fuel.

Sort of, but only sort of.

If you turn the pump off the fuel rail remains at atmospheric pressure and since the injectors have manifold depression on the bottom fuel will still be sucked through under light load conditions. How much fuel would depend on how much depression you have and what the ECU is doing to the injectors. If you have a lot of depression or the ECU is being very generous on the pulse length then it's quite possible that the engine could run with the fuel pump powered off. Typically though you would feel it firing but not quite enough there to keep the engine running.

[adam quantrill]

Exactly, I had the fuel pressure gauge on there and the pressure did indeed drop to zero (rel atm), but the engine kept running!

I think that the injectors are somehow being jammed on full (despite the pulsing I see from the ECU) and the manifold vacuum is sucking the fuel in - wierd innit! At least it means there are no plenum leaks!!

So the next step is to try and debug this injector/wiring fault. Hopefully each injector runs a separate independent line to the resistor pack, so I'll unplug each end and check for shorts etc.

[honestjohntoo]

GreenV8S said:

How much fuel would depend on how much depression you have and what the ECU is doing to the injectors. If you have a lot of depression or the ECU is being very generous on the pulse length then it's quite possible that the engine could run with the fuel pump powered off.

You are right Peter,

I'd overlooked the science that says that the pressure differential can keep the engine turning at low load idle, but I had read into Adams description (wrongly perhaps) that the car was running and reckoned that as soon as the throttle opened manifold depression would decrease as would the differential, so stopping the engine very quickly.

But it begs another question. I'd be interested to know if removing the pump relay whilst the engine is idling has a useful diagnostic purpose?

For example, might it show that there were no minor fuel line blockages? OR perhaps that the pump non return valve had failed thereby releasing the fuel rail pressure?

[adam quantrill]

Yeah - the removal of power to the fuel pump intermittently is a very quick method to point to all sorts of overfuelling problems - though it doesn't discriminate between them.

I find that if overfuelling is going on, the engine will immediately run better when the fuel pump power is disconnected.

In fact a long time ago I had an intermittent engine temp sensor that would go open circuit during the warming-up phase, and before I twigged what it was I had a switch wired in place of the fuel fuse so I could open and close it. This would get me through the rough running phase and once the engine was warm it was OK. When the car was overfuelled it wouldn't warm up any further due to the cooling effect of the fuel!

Anyhow I've received my 50M operation throttle pot so I'll make up a bracket and fit it soon...

Edited by adam quantrill on Thursday 4th December 20:23

[honestjohntoo]

adam quantrill said:

Yeah - the removal of power to the fuel pump intermittently is a very quick method to point to all sorts of overfuelling problems - though it doesn't discriminate between them.

I find that if overfuelling is going on, the engine will immediately run better when the fuel pump power is disconnected.

Anyhow I've received my 50M operation throttle pot so I'll make up a bracket and fit it soon...

Well Adam, I have to admit, You got me, I am completely moggydawed.

The process is not described anywhere I had seen so I think it is merely a characteristic of the system that just happens to occur in certain rich conditions.

Because its indiscriminate perhaps spark plug color, exhaust smoke density and fuel smell will be a better guide.

When the TPS is seen to be fixed, you will be able to progress further, on hopes.

Edited by honestjohntoo on Saturday 6th December 15:11

[adam quantrill]

At last - I seem to e getting somewhere. Imanaged to get it in the 'good phase' again this morning, and I managed to measure the on time of the injectors, which was between 1ms and 2ms. So it seems that the earlier masurements indicate far too long an on time, and in fact it lapsed back into the too rich condition and the injector on time went up to 10ms.

So it seems that the ECU is calculating far too rich a setting, so far, so good.

Now as to why? Well the old TPS is a bit dicky and did result in some overfuelling in the 'good' phase, but that was only when I tweaked it.

Eventually I measured the voltage on the engine temp sensor and it seemed a little high at 1.1v. I added a parallel 330 ohm resistor and the voltage dropped quite substantially to about 0.4V. I need to do some more comparisons but I think that an intermittent connection at the engine temp sensor input is the root cause. The sensor itself seems OK, when fully hot it dropped down to 300-ish ohms, I think it's the wiring in between.

Anyhow, I'm going to put in a workaround mod to the ECU breakout plug on my diagnostic ECU so's I can troubleshoot it and still drive the car, plus my new TPS looks the business electrically, so I'll be fitting a bracket for that too.

[GreenV8S]

The coolant temp sender is a thermister. The best way to test it is to disconnect the plug at the ECU and measure the resistance across the thermister. From memory it's 5KOhms at 0C and roughly halves for each 20C increase.

[honestjohntoo]

TEMP....OHMS....................TEMP....OHMS
-10°C...9100 to 9300............60°C....500 to 700
0°C......5700 to 5900............80°C....300 to 400
20°C....2400 to 2600............100°C...150 to 200
40°C....1100 to 1300

[adam quantrill]

Yeah well the temp sensor is brand new, and seems to check out OK with resistance vs temp, it's the wiring I was worried about (and/or the connections).

Anyway the rough running symptoms reappeared despite the extra mod to pull down the temp sensor to earth, so it isn't that. I did some tweaking of all the connectors to make sure they were all fine, but no improvement, so I skulked off to the barn to make up the bracket for the new TPS, which is now finished, connector wired up, etc. I'll give it a go in the morning (I don't expect it to cure anything apart from the spurious enrichment when it's running right).

The saga conntinues - hopefully it won't last as long as "wot's going on in your garage this weekend"!

Last thing I did was to check and clean up all the earth straps - some loose stuff found - but no improvement generally. Also I reflowed the solder joints on the 'spare' ECU - I'll be trying that one evening - really starting to scrape the bottom of the barrel now!

Edited by adam quantrill on Monday 8th December 14:44

[adam quantrill]

Well I pulled my finger out and have the new re-engineered throttle pot in there now. Unfortunately the new device works in the wrong sense - clockwise increases the output - so I had to put together a simple inverter circuit to reverse the voltage output. It seems to work really well with the voltage rising smoothly and linearly from 0.36V on the stop to 4.1V at WOT. I now get the enrichment clicks when I rotate the throttle rapidly.

Anyhow the ECU is still intermittently running rich or normal fuelling. It ran for almost 5 minutes completely normally after I installed the new pot, then reverted to the rich mode like someone flicked a switch.

I had another go running i without the fuel pump relay, and it definitely does still run. I can even disconnect the AFM and it still runs!

Both ECU's I have show the same symptoms. Either they have both failed in the same way or there's something strange on one of the inputs - but only intermittently.

[honestjohntoo]

adam quantrill said:

Unfortunately the new device works in the wrong sense - clockwise increases the output - so I had to put together a simple inverter circuit to reverse the voltage output. It seems to work really well with the voltage rising smoothly and linearly from 0.36V on the stop to 4.1V at WOT. I now get the enrichment clicks when I rotate the throttle rapidly.

Oops! One is tempted to observe that you simply had to reverse the connections to the potentiometer to achieve that effect. This generic drawing is very explict:



Introducing an inverter circuit adds complexity to the most simple of circuits, thus, more things to go wrong!

[honestjohntoo]

adam quantrill said:

Anyhow the ECU is still intermittently running rich or normal fuelling. It ran for almost 5 minutes completely normally after I installed the new pot, then reverted to the rich mode like someone flicked a switch.

I had another go running i without the fuel pump relay, and it definitely does still run. I can even disconnect the AFM and it still runs!

Both ECU's I have show the same symptoms. Either they have both failed in the same way or there's something strange on one of the inputs - but only intermittently.

As I mentioned previously, removing a component (Relay or AFM) from the system adds simply no diagnostic information to your enigma, as nowhere did Rover postulate what to do if such a symptom is observed. Sure its possible to guess how/why it continues to tick along, etc, but for what logical reason is beyond me, at least.

You are right, if both ECU's behave the same, it would be unusual to see them display identical faults.

Nevertheless, cross checking your system with a KNOWN good ECU from a friends car would eliminate that possibility.

Apart from the obvious checking of the sensors inputting data to the ECU, which I'm sure you have gotten covered from reading the gospels, you are now left with a couple of unusual but logical possibilities.

5 minutes is sufficient time for the engine to get get warm/hot so you might want to look for a heat related issue/quirk with your local Efi wiring connectors to and from any/all of the various components thereabouts. Vigorous wriggling of all the wiring and connectors as the fault occurs might reveal a dodgy connection, where simple visual/continuity checks do not.

5 minutes is also sufficient time for the state of battery charge to change after the prior cranking loads have drawn loadsa current, and the amount of charging current through the engine/chassis earth strap(s) to vary considerably as the battery recovers.

If that connection is dodgy then the voltage drop across it will vary considerably with charging current, and because chassis earth and engine earth may be different even in a well maintained system its worth checking all the key earth connections.

This is also particularly true for the Efi system earth on the grounding stud(s) behind and below the L/H rocker cover. Its whole purpose in life is to maintain a rock steady earth upon which the rest of the Efi system can rely.

Easy to dismiss these suggestions a trifling, but believe me, small changes in voltage on any of the sensor inputs is all it takes to upset the equilibrium of an otherwise perfect/balanced system.

Edited by honestjohntoo on Saturday 31st January 20:55

[GreenV8S]

honestjohntoo said:

small changes in voltage on any of the sensor inputs is all it takes to upset the equilibrium of an otherwise perfect/balanced system.

On a different system, I cured a persistent ECU reset problem by providing the alternator with a good sound earth. It's remarkable how indirectly some of these fundamental faults show up.

[adam quantrill]

Yes the symptoms came and went again this afternoon. I've already gone round the earths and cleaned them up but there's no harm in taking another look. I'll take a careful detailed look at battery voltage, too. I have the thing on constant trickle charge so it's being topped up all the time.

I changed the ignition amp (no difference) today as these can be unreliable, and I'mm thinking if it isn't one of the 'steady state' signals, then the ignition pulses might be the culprit. The pulses at pin 1 show a 5ms 10V drop followed by a spike about 2ms after the pulse finishes. If this is a normal waveform then fair enough.

Too answer your suggestion about reversing the connections, unfortunately this isn't a passive device I'm using, so the connections aren't reversible (without blowing it up)! A contactless device is much more reliable but yes there is the reliability of the inverter circuit to factor in.

[honestjohntoo]

adam quantrill said:

Yes the symptoms came and went again this afternoon. I've already gone round the earths and cleaned them up but there's no harm in taking another look. I'll take a careful detailed look at battery voltage, too. I have the thing on constant trickle charge so it's being topped up all the time.

It would seem that the suggestions dont apply in your case then, so moving on:

adam quantrill said:

I changed the ignition amp (no difference) today as these can be unreliable, and I'm thinking if it isn't one of the 'steady state' signals, then the ignition pulses might be the culprit. The pulses at pin 1 show a 5ms 10V drop followed by a spike about 2ms after the pulse finishes. If this is a normal waveform then fair enough.

I dont know from observation what the normal waveform should be (never had to question it) but hopefully you have the 6,800 ohm resistor in place. Various debates have called into question its purpose, but bearing in mind that coil negative might suffer as much as 150 volts of back emf on the LT side of the circuit, I personally think it would be prudent for the resistor to be in place on the basis of protecting the ECU from potentially damaging voltage spikes.

adam quantrill said:

To answer your suggestion about reversing the connections, unfortunately this isn't a passive device I'm using, so the connections aren't reversible (without blowing it up)! A contactless device is much more reliable but yes there is the reliability of the inverter circuit to factor in.

I'm probably out of my depth here but are you telling me it is a dynamic rotary position sensor (guessing now) that inputs an externally derived voltage signal into Pin #2 of the ECU instead of drawing upon the 4.3 Volt output potential between Pin #3 and Pin #18 of the ECU? If so you might care to examine the role of your missing 5,000 ohm potentiometer and the effect that missing load has on the input impedance of the acceleration enrichment circuits. Also, strangely, there appears not to be a direct earth connection from Pin #18 to "Engine Earth" but instead through some sort of fusable link. Another protection device perhaps?

As for a contactless device being more reliable, maybe so, I cant argue with that, providing the substitute is compatible with the original circuit design. I'm sure it is, as you clearly have everything covered and as I have not come across the signal system you have appended I guess as well as being out of my depth I also need to be silent and take to the lifeboat.

[adam quantrill]

The new TPS is indeed an active device, but is still powered by Vref and in seems that the output voltage is reasonably proportional to Vref, so I think this part of it is panning out quite well.

GreenV8S said >> On a different system, I cured a persistent ECU reset problem by providing the alternator with a good sound earth. It's remarkable how indirectly some of these fundamental faults show up.

Well as part of the earth checking procedure I found that alternator case -- engine block was about 180mV when running. This looked a bit on the high side given that the thing is bolted to the block. So I cleaned up one of the bottom bolts (surpriing how much stanless steel corrodes against aluminium!) and the voltage drop is much better, about 20mV or so. The car also runs _much_ better, in fact since I cleaned this up I think it hasn't gone into rich mode in ten minutes of running, so fingers crossed.

If this turns out to be the cause then I can only theorise that the alternator was introducing noise spikes iinto the entire system, plus also the coil earths around here so a bad earth might introduce nasty spikes into the ecu via pin 1. I think perhaps the spikes would fool the ECU into extending out the injector pulses far too long, hence the rich running, either by introducing false extra pulses into pin 1 or interfering with the steady state inputs such as the coolant sensor etc.

Anyway, let's see if this is the problem, I'll take the car to work and back (but probably not tomorrow, 4WD day by the looks of it!!!)

[honestjohntoo]

adam quantrill said:

The new TPS is indeed an active device, but is still powered by Vref and in seems that the output voltage is reasonably proportional to Vref, so I think this part of it is panning out quite well.

To improve my education will you put up a diagram and explanation of the circuit please.

[honestjohntoo]

adam quantrill said:

as part of the earth checking procedure I found that alternator case -- engine block was about 180mV when running. ---- so I cleaned up one of the bottom bolts and the voltage drop is much better, about 20mV or so. The car also runs much better, in fact since I cleaned this up I think it hasn't gone into rich mode in ten minutes of running, so fingers crossed.

I first came across these "milli-voltage variances" when I bought my Efi V8 in 1993 and have since offered "Bad Earths" as a last resort cause of unfathomable Efi issues.

Mostly, Efi probs are resolved by other means, but given how "common" earth contact corrosion must be, one wonders if there should be a "mandatory sticky" on all Efi V8 forums to highlight such underlying commonality.

Just how many cars, one wonders, are variously under-performing because of these small but significant situations? Fingers crossed indeed.

[GreenV8S]

adam quantrill said:

Well as part of the earth checking procedure I found that alternator case -- engine block was about 180mV when running. This looked a bit on the high side given that the thing is bolted to the block.

My setup is fairly convention with a V drive belt which is tensioned by swinging the alternator. So two of the alternator mounts are actually pivots rather than bolted - the third is bolted to a support bracket. The pivoting bolts might provide an earth - or they might not. The support bracket was covered in paint and rust and needed to make a good contact at both ends to provide an earth; it might do - or it might not.

Even worse, the voltage drop depends on engine vibration, electrical loads and even the injector pulse width. It was all over the place.

To avoid having to get the mounting bolts clean and keep them clean, my solution was to add a short earth cable bolted to the alternator body at one end and bolted to the engine at the other. This bypassed the various pivoting / rusty bolted connections and gives a consistent earth. This transformed the ECU from a problematic flakey thing, into something that just worked.

PS my car also has a earth 'ring' system fitted i.e. battery -> chassis -> engine -> battery. At least two earths have to fail to immobilise the car. Earth faults are less common now than they used to be but can be an utter s*d when they happen so it's worth making sure they can't.

Edited by GreenV8S on Monday 2nd February 22:47

[dickymint]

GreenV8S said:

To avoid having to get the mounting bolts clean and keep them clean, my solution was to add a short earth cable bolted to the alternator body at one end and bolted to the engine at the other. This bypassed the various pivoting / rusty bolted connections and gives a consistent earth.

Appears to be standard on my SEAC - even the AFM seems to have an original looking earth strap.