Just had a play with the MBE Diagnosis tool

Just had a play with the MBE Diagnosis tool

Author
Discussion

s3dave

Original Poster:

208 posts

165 months

Friday 5th November 2021
quotequote all
Been fascinated reading the thread “Intermittent start up problem”. I have had my car for 2 ½ years and only plugged the diagnosis in once, and then did not really know what I was doing, certainly did not understand the adaptive trace so much talked about. So I thought it best to learn and have a play. The engine is running well, but been high with emissions at MOT’s. I have 2 observations from the trace below: why is my bank 1 adaptives consistently lower than bank 2? And why is Lambda 2 trace much less stable? Which Lambda is behaving correctly?

I suspect I need to change one of the Lambda sensors




As an aside (and irrelevant to the trace above) I have noticed that when the water temp was reading 86 degrees, my dashboard was reading 96 degree is this difference normal?

subseamatt

80 posts

74 months

Friday 5th November 2021
quotequote all
Yes, the temperature difference between dash and ecu is normal.
Presumably the lamda reading doesn't necessarily mean one of the sensors is faulty, just that one bank isn't running as 'cleanly' as the other for any number of reasons? I wouldn't rush out and change anything until you've got more info from those that know more about these things, like the other thread (I didn't understand most of that thread but it was definitely interesting!).

shalmaneser

6,046 posts

202 months

Friday 5th November 2021
quotequote all
Not tvr specific but 11% adaption is very high indeed and would point to some rough running issues. There is lots of info online about trim adaptions and what can cause them, as well as further diagnosis that you can do.

nawarne

3,098 posts

267 months

Sunday 7th November 2021
quotequote all
S3Dave....Agreed if you're completely "green" then wait for Joolz to comment.

However, my understanding that with a 'positive' adaptive the ECU is adding fuel to the engine to get into the correct Air/Fuel ratio (AFR).
Whilst your adaptives show there is some imbalance between cyls 123 and 456 (as sensed by the lambda probes), I don't believe your displayed values are MILES out!
My (inexperienced) understanding is that adaptives above +/- 10% are not the best values, and anything more than +/- 20% is a real concern.

I guess your screen shot was done with the engine 'hot' having run for ~ 20 minutes? I find that all the sensor readings and inferred data settles down/steadies after everything has warmed up....the lambdas certainly need a bit of time to 'wake-up'

....My inference from your screen shot is that the throttle bodies need balancing using a syncrometer. If your happy doing that, one of the German Tuscan/(Speed 6) owners has produced an annotated photo on the best sequence to do this. It's not a 1/2 hour job though!

Nick

s3dave

Original Poster:

208 posts

165 months

Sunday 7th November 2021
quotequote all
Thanks Nick

I will admit I am "Green" when it comes to the inner workings of the MBE and any advice is appreciated. Yes you can see from the temperatures that it had been running for a while.

I balanced the throttle bodies with a syncrometer only a couple of thousand miles ago, but will give it a go again; I had some minor hesitation but it is now running well with new fuel. It does seem sensitive with E5 older than 6 seeks or so (I now have a new tank full of ESSO E0). I lent out my syncrometer to a friend so will have to balance them next weekend.

Its the general shape of the to 2 Lambda sensors that feels strange. One appears much more active and different in appearance - you could be right a rebalance of the throttle bodies, especially between bank 1 and 2 may change this. I'll give it a try next weekend and report back

Joolz mapped the car when it was in CliveF's ownership and obviously any advice would be appreciated.

spitfire4v8

4,017 posts

188 months

Sunday 7th November 2021
quotequote all
Nick above had pretty much nailed it ..

the obvious things to see are that your tickover is a bit too high, which will cause a net richening of the mixture, and because the lambda trim for the rear three cylinders is higher then that suggests that its the rear three cyls which need winding back slightly.

Remember that after every change of airflow you MUST reset the throttle through the software on the sp6 cars.
My routine is, make slight airflow change, reset throttle, reset adaptive maps, wait and check what effect that has had.

The other thing is that, whilst all things being equal if you have equal airflows you should have equal adaptives, that is not always the case. Injector flow variations (the lucas injectors are only balanced to something like 5 percent*) and pure mechanical differences between ports and combustion chambers will all affect the overall mixture. Do not get too hooked up on trying to balance the adaptive trims .. get the synchrometer showing as close to even as you can, with the total overall airflow giving a tickover around 950, and then just accept that the adaptive trims will come out how they will. In cases like this close enough is close enough. Don't go chasing the last 1 or 2 percent, you'll never win!

  • i know your sag has different injectors, but as a guide for everyone else the lucas injectors are generally not well matched as standard.

s3dave

Original Poster:

208 posts

165 months

Sunday 7th November 2021
quotequote all
spitfire4v8 said:
Nick above had pretty much nailed it ..

Remember that after every change of airflow you MUST reset the throttle through the software on the sp6 cars.
My routine is, make slight airflow change, reset throttle, reset adaptive maps, wait and check what effect that has had.
joolz, thanks for the advice. I will work on the throttle balancing next weekend. I did not reset throttle last time, so that was probably part of my problem. The tick over has always been high, so will also adjust this. Thanks again - will report back with a new trace after next weekend.

s3dave

Original Poster:

208 posts

165 months

Friday 12th November 2021
quotequote all
Update.

Throttles balanced and all very close now (bank 2 were all a little low). Throttles and adaptives reset many times during process. Engine sounds no different, in fact now I have lowered the tickover, it is hunting and less stable. In fact it’s all got worse! Obviously I now have an issue, the Bank 2 adaptives are high and hence the AFR 456 is red. I noted the Lambda 2, took much longer to heat up than lambda 1 on startup. Plot below. No other issues with the car, it is looking more and more like Lambda 2 is an issue, unless someone has other suggestion?



spitfire4v8

4,017 posts

188 months

Friday 12th November 2021
quotequote all
Have you done the lambda heater earth mod ? your ecu doesn't appear to be resetting throttle 1 to 15percent .. throttle 2 is fine. You may have a lambda fault, or you may have a wiring fault, or both. more investigation needed.

s3dave

Original Poster:

208 posts

165 months

Friday 12th November 2021
quotequote all
Thanks. Lambda sensors are 4 pin and the plugs look clean and dry. I have no idea about the lambda earth mod. I will check the wiring tomorrow and probably order a new Lambda sensor

s3dave

Original Poster:

208 posts

165 months

Friday 19th November 2021
quotequote all
Update:
Lambda 2 changed. Easier than I thought, once battery removed. Adaptives look better (but still not the same). I think car sounds better - however the "dThrottle" on the software is now showing inactive. Is this a problem? Apart from changing lambda 2 and disconnecting the battery I have done nothing.


hardtailer

76 posts

140 months

Saturday 4th December 2021
quotequote all
The lambda mod consists of connecting the ground wire of each lambda sensor's heating circuit to either the negative terminal of the battery or a nearby chassis ground.
Make sure you pick the correct wire as with universal sensors often both the heating wires are white. The other one carries +12V when ignition on.

To see whether you can expect an improvement from this mod observe the tps signal from ignition on and engine off, through starting (don't apply throttle pedal!) to engine idling.  You could see it jumping around to values representing between 20% and 40% throttle opening and experience very hard to no starting.
This should be done with a cold engine in order to have the sensor's heating circuit drawing the highest current thereby maximising its negative impact on the reference voltage to the throttle position sensor (tps).*


Turn off engine, remove the fuse for the lambda sensor (heater) and repeat the above.
You should now observe a rock steady signal that remains at the value it showed with ignition on, engine off.

If you don't see a wildly varying tps signal in the first test, then the mod might have already been done by a previous owner. If not done, it's still very advisable to do so in order to prevent future headaches.

*)The high heater current causes the high throttle opening value readings although in reality the throttle plates are still closed.
Based on the false tps readings the ecu injects far too much fuel for the amount of air actually entering the cilinders thus flooding the engine. It won't start unless you (almost) floor the throttle pedal.
Whereby you're ridding the cilinders of too rich a mixture as well as replacing it with a combustible mixture (at higher throttle openings the false tps value either deviates less from the actual value or the mixture is not so far off as to not ignite, perhaps even both. I can'tremember exactly what I observed on my 2003 t350).

Referring to your question in your first post: sensor on bank 2 varying quickly over its entire range proves it is in proper working and the reason for the mixture varying at a higher rate than taht of bank one lies elsewhere.

Plot 2 after balancing shows bank 2 has an idle mixture that is so rich that lambda sensor 2 is almost at the end of it's range whereas it's at the lean end when giving it some gas.
This can be caused by an erroneous signal for bank 2 or a vacuum leak at cil.4 to 6.
To confirm the first possibility See if unplugging tps of bank 2 changes anything first with the fuse for sensor heating in place and then repeat with fuse removed.
For which of the 4 combinations of tps2 and lambda sensor fuse connected/disconnected is the bank 2 lambda signal comparable to that of bank 1 in shape (range and rate of varying)?
Please report back what you find.

More info on the lambda ground mod
https://www.pistonheads.com/gassing/topic.asp?h=0&...

And wildly varying tps signals when starting:
https://www.pistonheads.com/gassing/topic.asp?h=0&...

Edited by hardtailer on Saturday 4th December 13:27

s3dave

Original Poster:

208 posts

165 months

Monday 6th December 2021
quotequote all
Hardtailer, thank you. I have just seen your suggestions. As you can see from plot 3 (after I changed the Lambda sensor) the Lambda readings were more consistent than plot2. I took the car out yesterday and it seemed to drive well. I will go through your tests on the Lambda earth and tps2 next weekend and report back.

s3dave

Original Poster:

208 posts

165 months

Sunday 12th December 2021
quotequote all
hardtailer said:
Please report back what you find.
Hardtrailer. Thank you for your advice

TPS rock steady on start-up. I can only assume the earth mod has been done. As it is CliveF’s old car it comes as no surprise as he did a lot of good things.

2nd test. As you would expect disconnecting TPS 2, the lambda sensor drifted badly. Maybe because I gave it a run last week and the adaptives settled down, but with it all connected both lambda’s behaved similarly, but Lambda 2 was slightly more volatile with a larger range and not maxing out at the end of its range. Note it was Lambda 2 that I changed. I also noted now both lambda’s seem to heat up at about the same rate whereas before Lambda 2 was slow; so I guess there was some issues with Lambda 2.

Unless someone tells me otherwise, I feel that I can leave things as they are, I will check occasionally, but I wonder if anyone has a both lambda signals exactly the same in range and rate of change. I would be interested in seeing their plots.


hardtailer

76 posts

140 months

Sunday 12th December 2021
quotequote all
Thanks for posting your findings.
I agree with you that you should be fine now.
A quick(?) test could be to swap lambda sensor one and two to see if the behaviour of Quick response and not-maxing-out stays with the new sensor when fitted to bank 1.
Otoh you for the price of a generic 4 wire bosch sensor it might be a better option to just fit a new one to bank1 exhaust manifold and then compare to a datalog with the current bank 1 sensor.