R400 D - Electrical problem.. i think

[aerobod - near CYYC]

Hi Tomiam, to get used to the software and understand the sensor outputs, it is worth starting the car and running it up to temperature while the car is stationary, you could then run your laptop while plugged into the mains.

While on the 2nd page, which should have about 25 panels displayed, look at the sensor values. All the temperature sensors should start out pretty close to ambient if the engine is cold and then steadily increase as the car warms to operating temperature. When fully warmed up your water temp should be over 80C and the fan should come on in the 90C to 100C range. Your air temp will be somewhat warmer than ambient when the engine is warm (will be dependent on heat soak and air flow around the filter), but will reach a stable value while the car is stationary.

With the car on but not running the TPS voltage should increase and decrease smoothly over the whole throttle range (I have the plenum intake not the throttle bodies, with my TPS voltage from 1.35V at idle to 4.75V at full throttle, which may not be the same for you). Throttle sites should also smoothly change from below 0.5 at closed throttle to over 14.5 at full throttle.

The battery voltage should be in the range of about 13.2 to 14.2V with the engine running at 2000RPM, but I have seen on my R400D that it can drop to about 12V at idle with the radiator fan on. I originally saw about 0.4V of noise in the battery voltage signal, but added some extra grounds to get this down to about 0.2V.

The lambda sensor short term control status will start at either "bad AFR", "waiting for timer" or " coolant temp too low". By the time the coolant temp gets above 60C and it is not in a "waiting for timer" state it should change to "all conditions OK" and you should then see the lambda data oscillate between about 0.8 and 1.2 between about once a second and once every 3 seconds under gentle throttle openings. Under aggressive throttle changes and/or high engine speeds you should see a lambda status of either "disabled by target map" or "TPS unstable"

At idle your throttle site should be between 0.0 and 0.5, the ignition advance will bounce around a bit from small positive to small negative values and your fuel final injection time when the engine is warm should be about 3ms per cycle, with the duty cycle under about 5%.

If you have the MAP sensor beyond one of the throttle bodies it should read (assuming close to sea level) 1.0 bar with the engine off, about 0.5 bar at idle, 0.2 bar when the throttle is closed quickly and 0.7 to 1.0 bar under acceleration. If the MAP sensor is before the throttle bodies, then is should just indicate atmospheric pressure at about 1.0 bar and not vary within throttle settings.

If all of the above look good, it is time to take the car for a drive to log data!

 

[Tomiam]

Thank you so much! Last night I decided to just rig it all up and go for it. Hopefully something will be obvious.

Your post will save me a massive amount of time searching the Web for what values I should be seeing.

Again. Thank you!

I'm off to the garage right now!

[Tomiam]

Ok... connected everything up.  Pretty simple to install the software and everything worked first time without any effort.

First thing Ive spotted. At both idle and when the engine is off, its reading "Throttle Site 1" as 1.1 not 0 - 0.5 as suggested above.

The TPS was replaced during the investigation into this problem.   Ive read on a few websites that you can reset its position by turning on the ignition and pressing the throttle pedal to the floor 5 times.   Havent googled beyond this yet.
 

Video: https://www.youtube.com/watch?v=0X5ObENv16E
​In the first bit you'll see me pressing the accelerator to get the rpm to 2000 with the throttle site shown as 3.4   However, after a few seconds the engine rpm starts to drop (by 300rpm) even though Ive kept my foot still and the throttle site is still showing 3.4
This would explain the stalling at idle, if this is happening then too.

In the second part of the video. after the engine stalls, I keep the ignition on and press the accelerator to the floor a few times.

[Tomiam]

Forgot to add... barometric reading is red. I'm going to check the part number on mine to confirm if it is meant to be a map or air temp/ barometric as suggested earlier in this thread.

Edit. Part no. Confirming it as a manifold absolute pressure sensor but when I disconnected it.. the air temp figure on easimap disappears. No. 1S7A--9F479-AC

[John Vine]

Confirming it as a manifold absolute pressure sensor but when I disconnected it.. the air temp figure on easimap disappears. No. 1S7A--9F479-AC

That makes sense. 

When I had my RBs fitted, CC Crawley transferred the existing sensor from the plenum, where it previously measured absolute pressure (that is, manifold vacuum, being downstream of the butterfly).  Tony at CC Crawley explained to me that, when used in the RB set-up, it operates as a combined air-temp / barometric-pressure sensor. 

Upgrading to RBs requires a revised fuel map to reflect this -- achieved by moving from a MAP-based map to a TPS-based one.

Looking forward to the next instalment of your investigation!

JV

[Tomiam]

Ive just done a logging session.  Engine runnning. Car sat still.    

Allowed car to warm up at idle.

Once warmed up and lambda saying "all ok" I pressed the accelerator and held it still at 2000 rpm.

I kept the accelerator still. Didnt move my foot. Checking both the throttle angle and throttle site readings which remained constant.

Engine rpm (at 2000rpm) starts to go lower and lower.. down to 1300 rpm ish. At this point there is occassional THUNK! noise coming from somewhere up front.  After a couple of THUNK!'s I've taken my foot off and the rpm drops to idle.

Ive linked the log file in the hope someone can have a look and say "it looks like this".   Its from about 4mins into the log (is obvious when looking at the data).   Prior to that i was unplugging MAP / Lambda to check connections etc.

EDIT: MUCH BETTER LOG FILE - FURTHER IN THE THREAD.
Dat file:  http://www.filedropper.com/miker4009am28hnovlsn

​Lsn file: http://www.filedropper.com/miker4009am28hnov

​re: MAP sensor (reconfigured as Air Temp / Barometric sensor according to John). I did try to source a new one locally but no one had one.
Im unsure whether its outputting everything normally , and the panel which is showing red on easimap (barometric pressure) is fine because the data that sensor is sending is showing up ok on 'another panel'.

 

[Tomiam]

Update regarding MAP sensor.

I believe it's working and the fact that easimap shows a barometric pressure panel in red is a red herring.

If I disconnect the sensor then the panels for air temp and manifold pressure and manifold comp are blank.

If I put a pipe over the sensor and suck/blow on it then the readings change. The air temp reading goes up as the engine warms up too.

Conclusion. Sensor is working.

 

Steep learning curve this lol.

I will get a clean log of starting up from cold and upload later this evening.

 

[Tomiam]

EDIT: double post.

[Tomiam]

Today has been a good learning experience.

Summary of situation & findings:-

Problem gets worse the warmer the engine is.  Perhaps linked to the reduction of fuel adjustment as it warms up?

Spark plugs were black & sooty.  A powdery soot on them.  Have cleaned and re-inserted.

With the engine up to temp (60+ deg C) holding the accelerator down to get 2000 rpm results in the rpm continuously slowly dropping.

Starting the engine from cold. The engine runs smoothly and is audiably smooth.  Rpm is about 1000.  As it warms up, it becomes less and less smooth with the rpm dropping suddenly and then back up as if it was about to stall. Rpm drops to around the 800-900 mark.   After the engine has fully warmed up, it eventually stalls.

Whaat I did on log file: (below)
Car is started from cold.  Left to idle on the spot.   Eventually stalls.   I then restart it and hold the rpm at about 2000 twice.   First time no issues (lambda on start up timer?) and second time the aforementioned rpm drop happens.

(on log file) TPS is showing a value of Throttle Site 1.1 on easimap at idle.
(on log file) Engine Speed starts off very consistant.  As time goes on it fluctuates a lot more including a number of big up/downs.
(on logfile) I dont know what it means but when I graph the data, the data for Oxygen A (Raw) drops just before the engine stalls.  It also disappears from the graph for a time period.   Lambda data also (which normally fluctuates) stops fluctuating during this time and returns a flat line which I assume is because the number its sending back has reached a maximum?

Log file: http://www.filedropper.com/sateveninglsn

If someone is able to have a look at it and advise I would be VERY grateful.  

 

[colindavies56]

whats the gap on that plug.

[Tomiam]

0.8mm
Since Ive had the issue, Ive changed the spark plugs to a brand new set to try and rule them out.

[colindavies56]

sorry mate,must be camera angle.

it just looks too large,and worn out.

[Tomiam]

No need to apologise mate.  More than grateful for anyone trying to help ;)
I see what you mean by that photo :D

[John Vine]

Are your plugs the standard Ford ones - part 5M5G-12405-AA, marked TR6AP-13?  If so, the gap should be 1.3mm.  (Although I doubt that has much bearing on your problem.)

The evidence so far seems to point to over-fuelling.  Why this should happen has yet to be pinned down....

JV

[aerobod - near CYYC]

Hi Tomiam, I could only open your first log, as I couldn't see the matching LSN file for the second log. In the first trace though, your lambda values are not oscillating correctly when short term lambda status is OK. The oxygen sensor voltage stays around 0.25V, but should oscillate around 0.5V, as in this diagram: http://n.b5z.net/i/u/6137562/i/narrow_band_o2_voltage_graph.jpg

This could be caused by a number of factors, but if all the plugs are evenly black I would rule out fuel injector problems for over fuelling. As the target lambda is 1.00 (that correlates with a narrow band oxygen sensor voltage of approximately 0.5V), a lambda value above 1.00 tells the ECU that the engine is running lean, so trim fuel is added to try and push the engine to a lambda value below 1.00, but that would require the oxygen sensor output to reach a voltage higher than 0.5V, which it doesn't seem to do.

As you have replaced the oxygen sensor, I would look carefully at the supply voltage, to see if it is high enough to operate the sensor properly, if possible back probing the sensor connector on the sensor wiring side. The black/yellow wire on the loom side of the connector is the power ground, the green/orange wire is the 12V supply, the blue/brown wire is the oxygen sensor output to the ECU and the brown/black wire is the signal ground to the ECU. The issue may be bad power or signal grounds or a ground loop between the sensor and the ECU. It is also worth checking the resistance from the ECU 36 pin connector to each of the oxygen sensor leads. Pins 5, 23 & 24 on the connector are signal grounds, pins 6 & 7 are power grounds, pin 13 is 12V switched supply and pin 29 is the oxygen sensor signal. The resistance with the battery disconnected should certainly be less than 1 ohm between a given pin on the ECU connector and the matching connection on the oxygen sensor, preferably close to 0.1 ohm.

Here is the ECU connector pinout (the 992 and 9A4 ECUs use the same pinout: http://www.sbdev.co.uk/Info_sheets/MBE/9A4-PinoutIssue%20F.pdf

Just an observation on the throttle site and TPS voltage, site 1.1 being the minimum may cause an oscillating idle, depending on how the map was written. Adjusting the TPS rotation if possible or closing the throttle idle stop could get you into the 0 to 0.5 range. Normally the throttle site would either be interpolated between adjacent maps or adjusted to the closest map, in this case maps 0 and 1 are relevant. Map 0 would normally be optimized for idle parameters, but map 1 would have parameters more appropriate for non-idle conditions. The TPS voltage seems to decrease with  throttle opening, this is a Cosworth TPS configuration. According to SBD, Cosworth accidentally switched their TPS wiring many years ago and have not followed the convention of low voltage being low throttle ever since. The R400D with plenum uses the conventional low TPS voltage equals low throttle opening, but the MBE ECUs can be setup to handle either increasing or decreasing TPS voltage with throttle opening.

 

[Tomiam]

Hi Aerobod, 

Thank you once again.   I know I keep saying this but the help is appreciated!

Testing of the wires.  Battery disconnected.  Paper clip + multi meter on ECU connector end, multi meter on lambda connector plug at end of loom (lambda unplugged)

Power ground,  Black  / Yellow.   Pins 6 & 7 tested fine. No resistance.

12V Supply.  Green / Orange.   Pin 13 tested fine. No resistance.

Sensor Output.  Blue / Brown.  Pin 29 tested fine. No resistance.

Signal Ground to ECU.  Brown / Black.  Pin 5 tested ok.  No resistance.   Pin 23 couldn't get any reading from this at all.   Pin 24 is missing from my ecu plug (guessing its not used),  

-------------------------------------------------------------------------

Lambda sensor testing.  Loom side connector plug for Lambda sensor.
Green / Orange is receving 12v without engine running & ignition on.   Is receiving 13v with the engine running.

--------------------------

Lambda sensor connector plug, on lambda sensor.   Lambda sensor unplugged on its own, and a continuity test is done across pins.
I have continuity between the power ground (black/yellow) and 12v supply (green/orange) pins.
I no continuity between the sensor output (blue / brown) and signal ground to ecu (brown / black).  This normal?

------------------------------

Lambda sensor connected.
12v power (green / orange) tested by putting safety pin into the back of the connector on the lambda sensor side.   12v being received,

New logs attached.   As before, I started the engine. Let it warm up.  It then stalled after a while.
I then restarted and kept my foot on the accelerator at a fixed position.  Whilst it was during 'waiting time' for the lambda, the engine revs remained quite constant.  Once the waiting time passed, the revs increased before continuing to drop.

Is it possible that the lambda , even though its been replaced, is broken?

Dat file: http://www.filedropper.com/sunamlsn
Lsn: http://www.filedropper.com/sunam

Another thought... fuel pressure regulator?  But i guess that wouldnt explain the Lambda saying "lean!" when we believe its actually running rich.

[Sim]

Do you have a leak in the exhaust before the lambda sensor ? eg exhaust manifold gasket. -  That could fool the lamba sensor into believing you are running lean and causes the ECU to richen the mixture.

Simon

[Tomiam]

Spent the day under the car.

Bruised fingers, aching head and a minor case of "seriously??...." later... I can confirm that Ive been over every connector plug that I can see and can not find anything amiss.  I've scanned everywhere I can see and can not find any damage / frayed wires etc.

I've taken the lambda sensor out.  It was covered in black soot.   Inside the exhaust looks the same.

I can not see any damage to the exhaust or manifold. All connections look secure and as far as I can see everything is bolted in place properly.

Going by the data logs, is it possible that Ive replaced a duff lambda sensor with another duff one?


 

 

[Jonathan Kay]

It is possible to test some types of lambda sensor in situ with an oscilloscope (or possibly a multimeter). We probably have someone who knows the details... otherwise I'll try and find them.

Other than that temporary swaps have been offered by generous Members...

Jonathan

[Tomiam]

Hi Jonathan, 

Yea I did a search of the interwebs this afternoon to see if I could find a reliable way to test the lambda.  Couldnt find anything reliable which didnt involve an oscilloscope.  Which I sadly dont own.

From my notes I believe the lambda sensor number is: ​MHK10006 LSH 24, 0 258 003 229, 9923032, 12v
Which I think is.. HERE
This is of course assuming that Aerobod, or someone else, doesnt look at the logs and tell me Im talking rubbish :)

[John Vine]

If it helps, I'm happy to lend you my lambda sensor for your tests.  At least that would eliminate one variable.

I'm abroad right now but will be back home by Thursday 3 Dec.

JV

[aerobod - near CYYC]

I think it is a bad lambda sensor or a problem with the wiring between it and the ECU (although your continuity and voltage tests seem to show the wiring is OK). The oxygen raw voltage at initial start with a cold sensor starts at 5V as expected until the sensor heats up. It is interesting that the oxygen raw voltage flatlines at 0V a couple of times, indicating a potential intermittent wiring or total sensor output failure issue, but when it isn't zero, it tends to 0.25V or so once warmed up.

At the 6 minute mark when the coolant hits 60C, the lambda status changes to "OK", then within 3 seconds the adaptive fuel starts increasing until it hits about 70% extra fuel at the 8 minute mark, when the engine stalls, after engine restart the adaptive fuel is back at 0% for 45 seconds while the lambda status is "warm up", then quickly starts to increase again. The way the adaptive fuel increases is consistent with the perceived lambda status and continued lean reading.

i checked the narrow band lambda sensor part number from my car and it is the same as yours (Bosch 0258 003 229), the heater element resistance on mine is 3.3 ohms across the two white wires. The black wire on the sensor should connect to pin 29 on the ECU (blue/brown wire), the grey wire to pin 5/23/24 (brown/black wire) and the two white wires will use pin 13 and pin 6/7 (black/yellow and green/orange wires).

One thing I also thought about, is your lambda sensor in the standard position just after the 4 into 1 transition in the exhaust but before the catalyst (if fitted)? Also, if you have a catalyst fitted, have your checked that there is no obstruction in it or the rest of the exhaust system? As someone else mentioned, an exhaust leak can be problematic, I used high temperature copper silicone sealant to remove some small leaks where my 4 into 1 pipe joins to the header pipes and the silencer (I don't have a catalyst). The leaks were indicated by some soot near the pipe connections.

[Smiffy69]

One other thing to look into is the spec of the lambda sensor and at what temperature it will start to work as they need to "up to temp" to start to read properly.  This is normally OK from a cold start as the ECU ignores the sensor until the coolant gets above 60C and by which point the ehaust gas has heated up the sensor.  If you get a scenario where the coolant is 60C + but the exhaust / Lambda probe have cooled to a point where its not in its operating zone then you will get rough running and possibly it will stall one th warm up timer expire. 

IMHO the warm up timer is not set long enough (mine is 40sec) by CC recommend that it should be around 60 seconds.  I can see from my Lambda trace when it come up to temp and switches on.  Try looking at your trace with coolant / Lambda / warm up timer and see if it corresponds to my findings 

This is a feature as I can't change it but if i could I would set it to around 80 secs which I am convinced would solve this issue.

[John Vine]

A wild long-shot...

If the lambda sensor is OK and your loom is OK, could the ECU itself be playing up in some way? You said that the problem started after heavy rain. I'm wondering if water has penetrated somewhere. You could try removing the ECU and leaving it in the airing cupboard overnight.  

JV

[Tomiam]

Thank you all :)

Today:-
Examined exhaust.  Small amount of soot around the 4 into 1 pipes at the 4 pipes end.   
Purchased copper silicone sealant.  Dismantled exhaust.   Cleaned up where the pipes connect.  Applied sealant around pipes and the bits they go in.  Reassembled exhaust.  Applied sealant as much as I could around the joins of the pipes.

Lambda sensor checked to ensure all pins go into the correct pins on the connector on the loom. All fine. Have previously checked from the loom back to the ecu.

Heater element resistance measured at 3.9ohm

Started engine.   0.6 was initial lambda reading.  Very quickly got to 1.01 and then climbed to 1.2 before engine stalled after enough time for coolant to get to 77 deg C

Have ordered a new lambda sensor based on the reference code given above: Bosch 0258 003 229   Should be here in 2 or 3 days.
Thank you for the offer of yours on loan John, but its about £37 quid all in for a new one.  Buying a new one saves arseing about with P&P etc.
Btw... if you wanted to borrow my cable from SBD, I know you were considering one, then once my car is up n running you are more than welcome.

Thank you Aerobod, John Vine, Smiffy, Jonathan and Sim.  :)