Mixture adjustment on a Rover 3.9 V8 without Lambda Sensors.

[Kiwibacon]

Hi guys. Newbie here so be gentle.

I have a rover 3.9V8 in a NZ/AU spec rangerover and I need help getting it to not bankrupt me. I see the expertise in these engines is spread right across this forum so I figured I'd post in the neutral technical zone.

The NZ/AU spec 3.9 V8's in rangies are 9.35:1 hotwire open loop control. They have no cats or lambda/O2 sensors. This leaves them up to the mercy of the airflow meter for mixture control.
I am a diesel guy, other petrols I've owned were shopping carts that never needed tinkered with. The tuning tools I have are a wide-band A/F ratio gauge (innovate motorsports on a temporary connection at the end of the exhaust) and roverguage.
I also have EEPROM programming tools, which I've used on diesel ECU's, but surely that is not necessary.

My current problem is simple, it runs very rich. Fuel economy between 20-26 litres/100km (10-14mpg). I'd like to see 15 litres/100km (19mpg).

A/F readings first round.
Idle (cold) around 11.5.
Driving cold 12-12.5
Driving warm 12.5-13
Acceleration drops below 11 then misfiring brings the A/F ratio at the exhaust pipe back up to around 15.

I left the battery disconnected for a week to reset the ECU and it did a bit better:
Idle is now around 12:1.
Driving when warm I did see 14.5:1 A/F ratio breifly on a few occasions, had it above 14.0:1 for a few km of flat road but mostly high 13's.

Diagnostics I've done so far:
Water temp sender checks out fine.
Thermostat replaced.
Fuel temp sender checks out fine.
Idle stepper checks out fine.
Set the trim-pot in the MAF to around 1.1V with engine off.
MAF voltage at idle checks out fine (did this weeks ago, can't recall exact voltage).

All of this has helped. But it's still too rich and so thirsty it just doesn't get driven. If it had O2 sensors then surely the ECU would adjust the fuel trim back and it'd all be good. But with no O2 sensors, how can I acheive that?

Current thoughts:
Installing O2 sensors is a job I just don't want to do.
Intercepting the MAF signal may do it, but this would be a last resort. Is there any way to further calibrate the MAF?
Writing a new fuel map shouldn't be necessary.
Is it possible to manually/electronically alter the long term fuel trims? But if so, would this need re-done every time it loses battery power and resets?

Cheers
Dougal from NZ.

[Kiwibacon]

Pumaracing said:

Well the fuel pressure being too high would be an ideal candidate for causing these sort of problems.

Ah yes, I have measured the fuel pressure, first at the pump (solid 37psi) then at the rail (32psi raising to almost 40psi when revved). This fits nicely with ~2.5bar over manifold pressure.

I've been tinkering with this for a year or so now, I keep forgetting what I've actually done. It was in a sad state and I know exactly why the last guy sold it.
It also has new cap, rotor, leads, vacuum advance tested, centrifugal advance free and moving, plug gap correct etc.

And ~100,000 miles genuine. So the cam shouldn't be too far gone.

[Kiwibacon]

Thanks guys, just answering these points in order. I really appreciate the help.

Pumaracing said:

A compression test should be the first thing you do. No point trying to sort an engine that's knackered. An adjustable fuel pressure regulator would be the easiest way to wind the mixture down if you can't find the real cause of the problem. If the engine is ok then I'd say it's either the MAF or something gone tits up in the ecu.

I don't have a compression tester of my own, but I might be able to borrow one. The engine shows no conventional signs of mechanical ill-health. Exhaust is clean, burns no oil. I do know about the reputation for these eating cams and I'll bet at 100,000 miles it's lost a bit of compression due simply to valves no opening that far. The MAF reads the volts it should. Which suggests (let me dream a bit) that airflow and cam isn't far from normal. But if it were reading lower (no different to being at higher altitude) shouldn't the MAF read the lower air-flow compensate for that? I mean what's the point of having a MAF if it doesn't?

At some stage I'll pull a rocker cover and measure how much the valves are moving to gauge cam wear.

Sardonicus said:

This may save you a lot of time Mark/Blitz is mustard with the Lucas CUX ECU http://www.g33.co.uk/fuel_injection.htm

I feel I've harassed him too much lately, so I'll let him join in at his own accord. I have one of his rovergauge cables and have read through that and all other 14CUX introductions.

paintman said:

Compression test as already suggested.
Depending how its been treated at 100k miles the cam shaft could be completely shot.

ETA They don't like cheapie pattern rotor arms & dizzy caps either. I had occasion to have to use a pattern cap on mine - on holiday & nothing else available locally - & had an occasional misfire under load. Re-fitting the genuine one having managed to source a replacement carbon brush for it cured the problem. Comparing the two, the genuine pick-ups are insulated apart from the side facing the rotor arm, shielding mouldings inside the cap, the lead turrets are much taller.
I also know someone (and I'm glad to say it wasn't me)who fitted a pattern rotor arm & then spent a LOT of money trying to cure a misfire to no avail. Until he refitted the old rotor arm.....

Edited by paintman on Tuesday 17th December 19:37

Cam is certainly a big question. This was a executives company car which was dealer serviced right up to about the 150,000km mark (all docs came with it) so the service history is as good as it could possibly be. If I have to do the cam and lifters etc then I will (mental preparation for that started a while ago), but I do want to rule out everything else first.
The thing is, the MAF reads the volts it should. Which suggests (let me dream a bit) that airflow and cam isn't far from normal.

A slightly worn cam should also only affect airflow at full throttle. At part throttle it should make very little difference? (yes this is a question).

The cap is genuine lucas, the leads bosch. Rotor arm came unpackaged, I couldn't find even a part number for a genuine one, if you have a source for known good ones I'd love to know.
But the misfire I believe is something else. Because just for one drive I managed to get it running perfectly with no misfire. This was after doing the cap, leads, rotor and fuel filter. On that drive I got it stuck in some mud and had to waterblast everything including the roof when I got home. The misfire has been with me ever since.
I live in a currently hot and dry climate and that was 4 months ago so everything should be well and truely dry by now.

stevieturbo said:

If you have a trim pot that lets you adjust output voltage from the MAF, then regardless of what a textbook says...try adjusting it until you see more favourable mixtures.

I haven't been able to find a good explanation for what that trimpot does. With my A/F meter I couldn't find any difference in mixture when winding it several turns either way, but idle became a lot more stable when I set it to the book values.

Edited by Kiwibacon on Tuesday 17th December 20:36

[Kiwibacon]

paintman said:

Depends how badly worn the cam is. Much discussion on what their life expectancy is, but not unusual to see 70/80k suggested. If you do change you need to do the cam and the followers. If you don't then the new bits won't last very long.

Good rotor arms are getting a bit problematic, avoid any that are riveted together, if you were in the UK I'd suggest this chap:
http://www.simonbbc.com/rotor-arms

Might also want to have a look through this:
http://www.carelect.demon.co.uk/rrind.html

Mine's the older 3.5EFi with 4CU & flapper AFM. I've had 3 4CU ECU's go over the past 20 years, one ran like a bag of nails, one went very rich - would just about run & the third was similar to but not as bad as the first. I've also had a couple of misfires that were sorted by renewing the plugs. I'm on LPG as well as petrol so the ECU faults were easily diagnosed as switching back to LPG resumed normal running.

Not got a tracking problem with the cap? You mentioned the issue started after you got it wet. Check inside as well, I've had condensation inside caps before now.

Edited by paintman on Tuesday 17th December 23:44

If I do get into the cam I'll be doing it properly, a mate is a landrover mechanic and knows all the pitfalls, I should ask if he's got a spare ECU for testing. But when a cam wears out, how does the ECU respond? I can imagine with O2 sensors the fuel trims keep reducing with airflow until they run out of range. So if they all wear evenly (unlikely I'm guessing) it would be a steady reduction in airflow.
But is this MAF completely useless? It doesn't seem to control much at all.

Thanks for the tips with the rotor arm. It's no problem at all to get parts from the UK. Most of my rover parts come from there already. Definitely no condensation now, it's been hot with 20-30% humidity for weeks and just to be paranoid I checked again yesterday. I'm wondering if I filled up the coil or amplifier or such with water instead!

[Kiwibacon]

paintman said:

Without Lambdas the ECU has no idea what the mix is & will just go on the info its getting from the AFM, coolant temp sensors, throttle pot etc. If the idle revs drop the stepper motor should come into play to adjust the air flow.

The suggestions in one of the links is to substitute known good items, but that's only really practical if you can borrow them as changing parts at random will empty your wallet but may not solve the problem.

I had another car - not RR - occasionally cutout on idle & all diagnostics showed nothing as it was intermittent & no fault codes were stored. When it finally stopped altogether it turned out to be the coolant temp sensor. The ECU thought the engine was cold when it was hot & fuelled accordingly.

Indeed, so how did they account for natural variability in MAF's and the rest of the system? Are there any possible tweaks or is a modified fuel map the only sensible way?
The coolant temp sensor was one of the first few things I checked. When I got the car the thermostat was dead, it ran cold and would paint the concrete black with exhaust at idle. But with a new thermostat the coolant sensor measured fine on a multimeter and now with roverguage I know the ECU is reading it correctly.

I do have a chip burner and the tuner-pro 14CUX data that someone has helpfully mapped out makes this seem very easy compared to hunting down values for my EFI diesel.

But first. Does anyone know what "map 2" in rovergauge refers to? Does this mean my tune resistor is selecting the wrong map?

[Kiwibacon]

Thanks guys.

Yes I do need to check/clean all the leads for the misfire. But I've got that as a lower priority than the obscene richness. A mate tells me that a dark (but well ventilated) garage is a great place to look for stray sparks.
HT leads (bosch) are only a few months old. Definitely have the firing order right, I'm not above checking things like that.

I will definitely try the MAF adjustment again. Right now I'm in holiday prep mode on my other rangie. It's got a 3.9 turbo diesel and gets a solid 28mpg on road trips. Got 31mpg once. Only once.

[Kiwibacon]

I changed the MAF setting to 0.7v at idle. No change in A/F ratios when running, but now idle is hunting more. This isn't the steppers fault, this is the engine rpm being unstable enough that the stepper can't keep up.

Testing the vehicle with another ECU seems to be the best plan for now, but sourcing one may take a while. Yesterday it was driving in the range of 12.5-14.5. Today it's driving in the same 12.5-14.5 on the way to a job but in the range 10-13 on the way home.

[Kiwibacon]

So I've been probing the 14CUX connector with a meter and can't get anything useful for tune resistor ohms.

Now I think I should have a red resistor, 180 ohms according to this table and many others: http://www.pistonheads.com/gassing/topic.asp?t=118...

But I'm reading very high ohms (megaohms) and Rovergauge says "map 2" which would be the 470ohm with cats.

Time to go resistor hunting I think.

[Kiwibacon]

Found the resistor. It was green and measured 470ohms. Mine should be red and measure 180 ohms.

I put a 330ohm in parrallel to get 193 ohms and that was enough for Roverguage to tell me I'm now on map #1 as it should be. Map #1 is for no cats and no O2 sensors.

Huge improvement too. Misfire is gone and it revs freely past 4k rpm. It's running high 13's to low 14's on the open road so I'd still like to go leaner. But for now it's actually drivable.
I had another play with the CO adjustment on the MAF. I get a clean and quite stable idle at 1v. Even 0.1V either side of this and it starts to stumble.

[Kiwibacon]

blitzracing said:

Glad the new map has helped- but recheck this figure- use RoverGauge to turn the fuel pump on with the ignition on, but the engine not running. This means the plenum chamber should be at atmosphere so you should see 37 PSI at the rail- 40 PSI is a bit high. If its genuinely running at 40 PSI, then check you dont have a blocked or kinked fuel return rail. You can check the pressure regulator with an airline and gauge in the input, (having removed it) and notice when it opens. Dont worry about it if it seems to leak air as the pressure drops, its not a fault, I dont think it likes air instead of petrol! Also even 150 mv extra on the AFM output is enough to make the engine over fuel, so double check the AFM readings. Dont forget Rovergauge shows you what the ECU sees, not necessarily whats at the back of the sensor. You can get a high reading if there are any earth issues between the AFM ground wire and the ECU ground. Every sensor has its own ground wire back to the ECU to prevent odd chassis voltages altering the readings the ECU sees. You can digitally rescale the AFM output on the open loop map using a "UAFC" unit made by a company called 14.7, if you want to mess around with fuelling without remapping, but the one I had did not prove to be reliable, and kept corrupting the maps. I think its very sensitive to voltage spikes.

http://www.14point7.com/products/ultra-air-fuel-co...

Thanks Mark

The fuel pressure didn't hit 40psi, just high 30's, but by that stage I had spilt some fuel on the plastic face and it went all cloudy and hard to read. I wouldn't trust that gauge to be accurate to more than +/-1psi anyway.

With the tunerpro definitions for the 14CUX it's going to be easier and cheaper to write chips than to shag about intercepting MAF signals. But given the time of year it'll take a few weeks to get some chips. I don't have a UV eraser so I'm hoping to find some compatible EEPROM's rather than the UV EPROM.

[Kiwibacon]

blitzracing said:

28c256 EEprom

Brilliant. Thanks Mark and have a good new year.
I've snagged some off ebay for about $US12 incl postage.

Edited by Kiwibacon on Monday 30th December 22:36

[Kiwibacon]

I've pulled the ECU for a look. No socketed chip or lucas cover on my one. Just a 28 pin IC soldered down and a label that's long since faded to unreadable.

So I've ordered a 28 pin ejector socket from Ebay just like Steve has here: http://www.pistonheads.com/gassing/topic.asp?h=0&a...

ECU AMR1424, Date Code 0893 so appears original to the vehicle and unmolested.

[Kiwibacon]

Misfire sorted.

I replaced the Lucas ignition module (electronic points for the old school) with a Bosch BIM024 ignition module. This is as simple as mounting the module, cutting the lucas plugs off and fitting spade lug terminals.

It now revs cleanly to the rev limit both in neutral and loaded. It even sounded different backing it out of the garage. This let me check the A/F ratio under full load, generally around 12.2:1 which isn't too far wrong.

Otherwise I'm still waiting for the chips and socket for the remapping.

[Kiwibacon]

Chips and socket for remapping arrived.

[Kiwibacon]

It appears mine was originally mounted on the distributor and had the genuine upgrade remote mount kit later. The original plug is there hanging freely by the distributor. The extension lead and plug to move the module to the aluminium heat-sink plate bolted under the coil all look like factory lucas with the correct plugs etc.

So while I'm sure this was an improvement over factory it clearly still had some issues. I wasn't going to gamble with more lucas parts for the 60 quid eqivalent the bosch cost me.

[Kiwibacon]

I had a crack at unsoldering the chip from the ECU board today. No good.
As well as being soldered down, it's glued down with the resin that coats everything.

I have a backup plan that involves bribing an electronics tech I know. He's got a nice targeted heat-gun that can melt and release without going all "scorched earth".

[Kiwibacon]

Max_Torque said:

Generally, if you haven't got access to the "right" tools it's best to sacrifice the chip and not the entire ecu. By that i mean, use mechanical methods(sharp, small, side cutters) to cut the legs of the PROM, then remove it using careful leverage, preferably with a plastic object to avoid damaging any tracks on the pcb. After that, you can individually un-solder and remove the remains of each "leg" from the individual through holes of the pcb, which is easily done with a normal soldering iron and some tweasers.

That exact thought had passed my mind. I've saved a ROM dump from rovergauge, but I don't know if that is exactly the same image I'd get from removing the chip and reading it.

If I can extract the chip whole, then I can test my socket soldering with the stock chip and one or two less variables.

[Kiwibacon]

I've had good luck (with the help of that electronics tech) removing stuck down chips before. He has a targetted heat gun which melts all the resin and uses tin-foil to shield the surrounding parts from the heat.
So we'll try that again, if it doesn't work then it's butchery time.

[Kiwibacon]

ECU is back. Brilliant job with the original chip preserved and the quick-eject socket on the board with a notch to fit around that big wafer thing.

[Kiwibacon]

Well it starts and runs with the original chip. So all good with the socketing job.
But it's playing "clack-clack-clack" with the AC compressor clutch when the first tune attempt is installed.

Hmmm. This could be time consuming.