MG-2013 is only 26 days away! · Corvallis, OR · July 17–21, 2013 · Visit MG-2013.com or the MG-2013 Forum for more info

MG Engine Swaps Forum

Welcome! Sign In Register
Please Sign In or Register to Search

Idle issue with Rover 3.9 EFI

Posted by cgill 
Jim Nichols
Sequim,WA, USA   usa
Chris, Jeff's advance curve would be fine. I'm thinking the stock vacuum can that was meant to use ported vacuum and allows way too much (probably around 22 degrees) vacuum advance at idle if attached to manifold vacuum(plenum)rather than above throttle body. My guess is the Rover distributor vacuum advance can only has 10-12 advance and is hooked to the plenum. That's why I suggested the Crane adjustable one. My experience with the Ford was similar with hunting, high idle. I wonder where Glenn hooks up his advance hose? He probably uses the Rover distributor. As for dwell the Delco window distributor is pretty easy to set with engine running with an allen wrench. If you don't have a dwell meter you can turn the allen wrench slowly untill engine rpm slows down both ways and set it right inbetween.



Edited 2 time(s). Last edit at 04/19/2012 12:36PM by 260mgb.

. You can hide this ad & support this site by upgrading to a Gold Membership ~ click here for more info.
Duncan Avatar
Duncan Cowen
Vancouver, Canada   can
Chris, perhaps it's information overload here, but there are some great suggestions here. For instance, unplug stuff! The ECU will continue to function based on factory-programmed "guesses," but it won't do any changes. Hey Simon, that means if you unplug the sensors, the EFI will work just like... a carb!! devil smiley

You can try this. Unplug ALL the sensors, maybe one at a time, then all at once. See how it runs. If it still hunts, when the ECU doesn't know or care what's going on (it will be blind and deaf with no nose!), then your problem is NOT with the EFI.

I had a friend's Jeep that had a high speed miss. It actually misfired enough that the cat overheated and burnt itself hollow inside! It finally began to stall occasionally, too. So, we unplugged sensors to diagnose, and you know what? It ran just as well (or poorly!) with no TPS, no IAT, no O2. Wow! Since my Jeep runs great, and it has the same sensors, we swapped known good parts in (including the cam position sensor, crank position sensor, cap and rotor, throttle body, etc). To no avail though.

We finally fixed it! When we took off the distributor to check something, we found a "rebuilt" sticker on the back side. We tossed in a ratty old distributor from the wrecker, and it's run perfectly for the last two years. Now, the distributor on that Jeep has ONE moving part. How the heck could it be a problem??? But it was!

I think, just like carbs, that most of the "fuel" problems are... ignition problems!! Go out and unplug EFI stuff to see if it still hunts.
cgill Avatar
Chris Gill
Salmon Arm, British Columbia, Canada   can
1971 MG MGB GT
1979 MG MGB "BOPR V8 Conversion"
Thanks guys. I am absorbing all the information you are throwing at me!

Here's a quick update. I dropped in on another rover guy in town and he loaned me his MAF and ECU from a 95 land rover 3.9. First I swapped in the MAF and it still hunted for the idle. Then I swapped in the ECU with his MAF and it did the same thing. The oxygen sensors also acted in the same way. So it's not the MAF or ECU. I checked the bolts on the manifold and they seemed to be tight although I don't have a torque wrench so it was just done by feel until I can get my buddy over on saturday.

So my next plan of attack is to go over the oxygen sensor wiring with my mechanic buddy on saturday. Then we will start unplugging things and see what happens. We will also do the carb cleaner spray around everything to look for vacuum leaks again. Plus we'll look at the distributor and check timing, and, and, and...



My other obsession...photography!

http://www.flickr.com/photos/77646803@N00/

. You can hide this ad & support this site by upgrading to a Gold Membership ~ click here for more info.
Simon Austin Avatar
Surrey, BC, Canada   can
Somebody mention carbs? Whadda need to know? smiling smiley

Chris, I'm envious you can access your ECU so easily. Wish the boys at Rover had thought to install it somewhere else rather than right above the AC evaporator.

You'll get it sorted. thumbs up



"Speed costs........how fast you want to spend?"
Mark Thompson
Hungerford, United Kingdom   gbr
Dont do as the chap has recommended and unplug all the sensors, the ECU is smart enough to do a self test on them on power up, and if the signals from the sensors are wrong or missing if will fall back into fault mode. The only reason you can do it on the stepper, is there is no feedback from this so the ECU is none the wiser, and the lambda sensors as they don’t give an output all all when cold, so if the ECU does not ever see a signal it just assumes the probes are still cold and runs on the base map. It does not activate the lambda circuits until it see the first rise in voltage as the probe gets hot. As for air leaks, you can check for these by blocking the stepper feed, then winding down the base idle screw to shut off all the air supply, and the engine should die. If it does not air is getting in somewhere else, although it will idle really badly as its not going through the AFM so the fueling amounts will be wrong. Minor air leaks in the plenum don’t affect the high power running, just the low speed idle and pickup. If the idle is unstable with the stepper blocked you could either be locking at an air leak as you say, or an ignition issue where it may be running on 7 cylinders, then 8, allowing the idle to pick up. Its difficult to pick up HT faults without something to read the HT voltage down each line, but if you can you should see around 10-16 kv per line at idle, but this voltage changes depending on the type of HT leads you have or if the engine has more or less mixture going in. The voltages rise as the engine takes more air and fuel. If you suspect a cyclinder is not firing, you can use a cable tie on each of the exhaust headers when hot, and see if it melts, as a missing cylinder will run a lot cooler. It does not work that well on cast manifolds due to the rapid heat transfer through the casting.



Edited 1 time(s). Last edit at 04/20/2012 02:09PM by blitzracing.

. You can hide this ad & support this site by upgrading to a Gold Membership ~ click here for more info.
Duncan Avatar
Duncan Cowen
Vancouver, Canada   can
In reply to # 2045514 by blitzracing Dont do as the chap has recommended and unplug all the sensors, the ECU is smart enough to do a self test on them on power up, and if the signals from the sensors are wrong or missing if will fall back into fault mode.

Absolutely right that the ECU will run in "fault mode." That's the idea! If running in fault mode makes it smooth, then that tells us a sensor must be giving faulty information to the ECU (or the ECU interprets the info wrong), which helps narrow the field of possible problems.

If however the car still hunts at idle in "fault mode," then the problem has nothing to do with the disconnected sensors. Again, useful info (as with the Jeep diagnosis I mentioned above - we ruled out all the sensors by unplugging or swapping them).

Chris, as to vacuum leaks and an unstable idle, consider the vacuum leak caused by mixing manifolds and valve covers that I had on my car. Now, I think you had some parts machined, or swapped in orientation, to fit under the bonnet? That might be a good place to start a comprehensive investigation!





Edited 1 time(s). Last edit at 04/20/2012 01:08PM by Duncan.
Mark Thompson
Hungerford, United Kingdom   gbr
The problem you have doing this is lets say you disconnect the AFM or throttle pot the ECU has no data on airflow or throttle position- so it can only approximate the required fuel input- basically enough just to keep the engine running to get you home, and it also switches to a different fuel map in the process- it can never be described as running anywhere near correct this way, so you cant use this as a method of fault finding. Although these ECU's are pretty long in the tooth now, they are still pretty smart, and one of the biggest issues is they have a huge fuel trim adjustment so if something is out, it just tries to compensate with fuel trim, that simply confuses the isse when trying to trouble shoot the initial failing. On the UK wiring looms you can switch off catalyst mode and lambda feedback altogether with a "tune " resistor that pokes out of the wiring loom, but I suspect this has been removed on US looms. Point being here is you can run the map without the ECU altering the fueling, and you can also set the idle mixture with the "CO" trim on the side of the AFM, so you have a mixture stable platform to start with. The nearest you will get to this on the catalyst fuel map is to unplug the lambda's, so the ECU cant adjust.

Mark Thompson
Hungerford, United Kingdom   gbr
Thinking about it- you have the ECU mate that will clearly show if one of the sensor inputs is changing to correspond with the change in idle, but you have already tried another AFM which would be number one suspect. You also know its not the ECU altering the airflow as it still hunts with the stepper blocked off, so that would only leave fuel under ECU control, and if you unplug the lambda's that wont change unless one of the temp senders is doing something odd.- you can check the readings with the ECU mate. The throttle pot is not a major part of fuel control at idle, but again the ECU mate will show if this is working anyway. Im going for ignition system, or a mechanical problem like head gasket.
Rich Truett
Berkley, USA   usa
Chris:

I am working through this very same issue with my TR8 4.0-liter w/ Glenn's fuel injection system. I am making great progress.
My car is almost there. Here's where I made the biggest improvement: the PCV system.
You CANNOT veer from the Rover set up.
That is, you need the small breather on one valve cover only to admit air into the engine.
The mistake I made was that I had an open orifice on the other valve cover and just put a small K&N filer there.
That caused a huge vacuum leak and all sorts of hunting.
If you have any PCV valves other than the stock Rover, eliminate them. That helped me, but did not totally solve the problem.
Next, I set the base idle. That really helped get me closer to the goal line. She does not stall and only hunts from 500 or so t0 750 once in a while.
Also check that the rubber O-Rings on the oil filter cap and flame trap are good. You could get vacuum leaks there. The dipstick also would be a source of a vaccum leak.
Despite what you may have read, if you are running a 14CUX ECU, you DO NEED the vehicle speed sensor. They are cheap and you can easily have custom speedo cables made, like I did. There will be no smooth idle w/o that. You will get stalling without it.

Later today, I am switching to the green tune resistor to see what that does. My ECU is a 14CUX runninbg the TVR Chimeara fuel map. I do have 02 sensors, but no cats.

Richard Truett
Berkley, Michigan
1981 Triumph TR8 4.0-liter

Rich Truett
Berkley, USA   usa
Here's my 4.0-liter engine in my TR8 that clearly shows the wrong way to plumb the PCV system. This orifice in the left-side valve cover is now completely blocked off. The Rover PCV is stock on the other valve cover. This fixed 90 percent of my idle hunting/stalling problems. There should also be a T piece in the vaccum line that runs from the flame trap to the throttle area. The T is fitted with a small vaccum hose that connects to where the wrong PCV is on my engine. This also really helped smooth things out. I feel Chris's problem is two or three things: incorrect base idle, no speed sensor and possible wrong PCV set up.



Edited 1 time(s). Last edit at 04/20/2012 03:30PM by Rover827.
Attachments:
Romo.jpg (48.9 KB) –
Romo.jpg
Rich Truett
Berkley, USA   usa
Rich Truett
Berkley, USA   usa
Glenn T...

If you are following this thread, you may be interested to know that tonight I finally solved my inconsistent idle/stalling problem. Now, my TR8 works just like it should.

Here's were we went off the rails...
(1) The PCV system has to be stock Rover, no deviations. At idle, the PCV system is what lets air into the engine because the throttle plate is closed. On the hose from the flame trap to the plenum, there needs to be a T-piece. The smaller hose goes to a fitting on the side of the plenum, right next to the power brake booster outlet.
(2) There MUST BE a vehicle speed sensor if you are using a Land Rover non-modified ECU.
(3) The base idle speed must be set. It was way, way off on my car. Perhaps because I am running a mild cam. In any case, I took off the tamper proof plug and set the idle speed by blanking off both sides of the hose to the stepper motor. Set the idle at 600 RPM. Then hooked everything back up. She idles perfectly now.
(4) I switched to the green tune resistor since I am running the TVR Chimeara UK Market ECU.

So, after about 17 months of shitty idling....hunting, stalling, etc. No more problems. The 4.0-liter engine is now silky smooth at idle.

Rich Truett
mgv8glen Avatar
Glenn Towery
Dover De., USA   usa
Rich, yes you need to leave all the Rover PVC/flame trap & T with the .100 restrictor hole that goes to the bottom of the trumpet stack plate. Yes this all has to be stock/left alone!!! On mu cars I run on the other valve cover from the flame trap the small open K/N filter. Rich, you are the ONLY person that I have sold a speed sensor to, to put in with the Rover F/I & I am sure I have sold 12 to 15 now that are in use. My motot might have a search of 50 to 75 RPM every now & thin. I wish I had writed down what year I put the first F/I in but I am sure it has been more than 12 years now on 3 different M.G.s of mine, a 3.5 a 3.9 & a stock 4.2 & a VERY HOT 4.2. & I am sure I am close to 300,000 miles PLUS with F/I now. I have run Mr. Jeffs dizzys for a lot of years. I did change the location of were the vacum was put on the plentum for the weight of the
Rover the hole was farther away from the butterfly & at 60 mph in my little M.G. I was not getting my vacum to the dizzy (butterfly was not to the vacum port) for were I would get my fule mileage. (will post a pic. tommorrow). Rich, I read your post and Chrises & I see were the 2 of you have the same thing that might be your problem (mild cam) I tried to use a Isky 270 cam in the first engine that I sold with F/I, RAN LIKE STINK but would not Idle, the stock cam has a 103 deg. of (center line I think) & with a mild cam you MUST HAVE 112 deg. I had Isky make a 270 cam with the 112 deg. & worked like a champ. & would do a 700 rpm idle. MORE FOOD FOR THOUGHT!!! FWIW

Mark Thompson
Hungerford, United Kingdom   gbr
A a quick bit on switching fuel maps between green and white tune on TVR maps- you have no way of knowing if TVR ever modified any of the other maps other than the cars original fitment, I suspect not, so you would switch from say a 3.9 TVR map to an unmodified Range Rover 3.9 non cat map. The biggest difference between the maps is the airflow scaling where the map has been stretched on the TVR map to allow for the larger AFM voltages for an engine that will reach 6000 + RPM, whilst the Range Rover one tops out at around 5400 rpm. The top speed limiter is also removed on the TVR maps, but this may not be part of the fueling data. If you run the green tune you need to make sure the engine does not run lean as the RPM rise, and it cant be run if physical catalysts are in place. If you have the Mark Adams Optimax chips, both cat and non cat map are modified for the engine spec'.
Rich Truett
Berkley, USA   usa
There are no cats on my TR8. But I do have 02 sensors. I just put 50 test miles on my car. WOW! That's all I can say.

The green tune really added smoothness and a bit more power in the mid-range. The 4.0-liter does have a mild cam, but she idles perfectly smooth at around 800 RPM.

MY ECU is one I bought off of eBay UK sometime ago. It is a modified Range Rover ECU, reprogrammed to run the TVR fuel map. So it seems the green tune resistor is the one that matches the ECU.

As for the idle: It is now perfect. I have never had a perfect idle -- that is, without hunting and stalling -- in the 17 months the EFI system has been on my TR8. In that time, I have learned that it is critical to set up your PCV system as per Rover, get your timing and base idle correct and make sure you connect up a vehicle speed sensor.

Rich Truett

Mark Thompson
Hungerford, United Kingdom   gbr
Check the date stamp on the ECU you have, if its mid 1992 or later it will be a TVR catalyst fuel chip. The reason for the nice idle and smoothness is the Rover engine likes to run slightly richer than lambda 1 to get it smooth. The catalyst lambda switching really does bang the mixture around quite a bit at idle, especially if the cam is longer duration than the original which causes the rough idle.
Rich Truett
Berkley, USA   usa
Thanks, Mark, for the advice.
If I do have the later Cat ECU, should I switch back to the white tune resistor?

Mark Thompson
Hungerford, United Kingdom   gbr
You need to do a simple test- get a test meter on you lambda sensor output with the leads back into the cabin so a passenger can read it (analogue is best as they dont jump around as much). Then hold the car flat out and let the revs climb to peak RPM (3rd gear will do) and check the lambda voltages dont drop below 1 volt. If you can get 1.2 volts even better, but its all a bit variable as the voltage to mixture range above the switching point at lambda 1 is a bit variable as the probes age. I spent an age trying to extract useful data from the narrow band probes at the richer end of the engines fueling range, but the results where never consistent across different probes- Anyway If it drops below 1 volt (its lightly to suddenly drop to zero as it drops below lambda 1 an A/F ratio of 14.7:1) its really too lean for full power use, where you want an AF ratio of about 13:1 for peak power. Its not all lost if it does drop, you can boost the fuel pressure a few PSI to try and compensate, but then it tend to run rich lower down. You can compensate with the CO trim adjustment to an extent, but this only alters the mixture up to around 2400 rpm, so you end up with a mid range area you cant really alter if the full power fueling is now correct. If the basic map is wrong, then this bit of free software has the ability to grab the mapping data from the prom, so if you can copy it from a non cat map, and reprogram the Eprom, but that an issue for a different post. :-) Its a brilliant bit of software work this, I got it up and running last week, no real bother on a laptop with a USB serial adapter.

http://www.pistonheads.com/gassing/topic.asp?h=0&f=13&t=1134539&mid=56164&nmt=Graphics+interface+for+the++14CUX
Attachments:
Rova non cat.JPG (55.1 KB) –
Please Sign In or Register to view additional attachments.
cgill Avatar
Chris Gill
Salmon Arm, British Columbia, Canada   can
1971 MG MGB GT
1979 MG MGB "BOPR V8 Conversion"
Away from home right now but reading all the comments. I will digest and respond when I get home. Thank you for all the help!



My other obsession...photography!

http://www.flickr.com/photos/77646803@N00/

cgill Avatar
Chris Gill
Salmon Arm, British Columbia, Canada   can
1971 MG MGB GT
1979 MG MGB "BOPR V8 Conversion"
Okay I will need a day or so to absorb all this information. Thank you for all the input!

Today I took the car for a two hour cruise. It idled high (well around 900 rpm) but nicely with no 'searching' about 99% of the time. Once in traffic it started to hunt for the idle and sputtered a bit until I gave it some throttle and then it was fine. Under hard acceleration from low rpm (2000) the car sometimes hesitated or 'surged' so the acceleration was not uniform if you know what I mean. Other times it pulled like a freight train to redline with no hesitation. So the symptoms (searching idle, surging/hesitating under acceleration) are not there all of the time.

Also, I was on the highway for quite a bit and the car ran very well at 50-75mph with no missing or hesitations. Because of the inconsistent symptoms I don't know if I would blame the cam.

Weird....a carb or a mega squirt system is sounding really nice right now...

Chris



My other obsession...photography!

http://www.flickr.com/photos/77646803@N00/

Add your reply here, or post your own questions!

Members Sign In if you've already registered, or
Register a New Account
Registration is free and takes less than a minute

Having trouble posting or changing forum settings?
Check the Forum Help File (FAQ) or contact the webmaster.
View the archived version of this thread.
Built using Phorum Open Source Software


Join Our Club

Sign In to post questions or share your photos!

MGExp Menu

Welcome

Forums ->

MGB & GT

MG Midget

Buy, Sell & Trade

Vendor & Group Buy

MG Engine Swaps

Original MG

MGA

MGC

MG Magnette

1100 & 1300

T-Series & Prewar

Modern MGs

MG Motorsports

MG-2013 Event

Member Meetup

Other Vehicles

Off Topic

Clubs

Forum Search

Latest Posts

Journals

Calendar

Membership

Tech Library

Car Registry

Cars For Sale

Model Info

Motorsport

Directory

Member Map

MGExp Store

Search

Advertising Info

Smartphone quick link
mgexp.mobi

Adjust Text Size

Larger Smaller
Reset Save