Hiccup when blipping throttle from idle - 1.8 K series, MEMS 1.9

[Jon Bradbury]

I have a problem with my 1.8 K-Series Elise (noin VVC, it's the 118BHP model same as the MGF) . When the engine is hot and idling, it stumbles when you blip the throttle. Doesn't do this when cold. I have asked about this on SELOC and the MG-Rover.org forums. I've checked lambda sensor, IACV and TPS, checked for air leaks (none found).

Here's the MG thread to save me cut & pasting it here: https://www.mg-rover.org/threads/off-idle-stumble.872060

..and the SELOC thread, which contains some MEMS_diag log files: https://forums.seloc.org/viewthread.php?tid=478058

I am using Rover MEMS and MEMS_diag software on an Android tablet to access the ECU.

Current trim values are (as reported by the ROVER MEMS app):

• Ignition advance offset = 126 (range 116-140)
• Idle decay = 31 (range 10 - 60)
• Fuel trim 135 (range 0-254)
• d2-d3 2 (no range given)

..so I'm looking for a known good set of values for these parameters to use as a starting point. Any advice?

 

[rubenste...@gmail.com]

My experience is with MEMS 1.6 (Rover 100).    But I’ve looked at your graphs on the MG Rover forum and something doesn’t look right.       Your car is idling and you are occasionally blipping the throttle.    But the top graph shows the ECU going in and out of closed loop for no apparent reason.     I would expect it to stay in closed loop throughout, even when you blip the throttle.          (That top graph would seem more appropriate to very aggressive driving, where you are flooring the accelerator pedal frequently causing the car to go into open loop. )    At idle it should be closed loop all the time, with the 02 sensor switching back and forth around once a second - how do you pass the MOT emissions test with such erratic fuel control at idle, or has it?

So my main question is why is it frequently and erratically going into open loop on idle for no apparent reason?

What does it look like when you are driving at a steady speed?         Does the closed-loop graph stay at a steady high, or does it similarly keep dropping down to zero for no apparent reason?

[francisa...@gmail.com]

>>>...Your car is idling and you are occasionally blipping the throttle.    But the top graph shows the ECU going in and out of closed loop for no apparent reason....>>>

I think that is normal procedure.   As I understand it, the moment you take your foot off the throttle the ECU drops out of loop so it can stop the fuel flow while the engine is slowing down.  Once it reaches idle it goes back to close loop to provide the correct fuel flow.

[leopold]

You didn`t post your logfile here so I couldn`t have a look to it.But on your description and answers from Martin and Francis it could be the IACV. Normally this should compensate small changes in throttle movement.

[rubenste...@gmail.com]

“I think that is normal procedure.   As I understand it, the moment you take your foot off the throttle the ECU drops out of loop so it can stop the fuel flow while the engine is slowing down.  Once it reaches idle it goes back to close loop to provide the correct fuel flow.”

Thanks, but at the start of the graph and in a couple of places at the end, it seems to happen without the throttle having been blipped.     But if you say that’s normal, that’s good; it’s just something I’ve not seen on my MEMS 1.6, it’s certainly something you wouldn’t want to happen randomly during the emissions test!

[rubenste...@gmail.com]

I see now my mistake: I was concentrating on  the throttle percentages graph looking at the large blips and not noticing the very small changes in throttle percentage elsewhere.         That now makes me question why sometimes the throttle percentage shows a large opening (WOT?) and at other times (apparently) hardly a movement of the throttle, and yet both actions appear (on the rpm graph) to give the same results.       I wonder if this is an anomaly with the graphing software for throttle percentages rather than a reflection of something genuinely odd with the throttle position sensor?

[Jon Bradbury]

The TPS has been tested with a multimeter and it is reading progressively and smoothly as the throttle is opened / closed. No jumps steps or other roughness. I also checked that the ECU sees the same thing - it does. The throttle percentage errors you are seeing in the graph are down to the low sampling rate of the MEMS serial interface. Each data packet is acquired roughly every 500ms at best, so it isn't surprising that a quick throttle blip can be missed. Unfortunately the data packets are a fixed size, so that means I can't tell the monitoring app (MEMS_diag in this instance) to log just a few parameters in order to increase sampling resolution.

Regarding the other items, IACV is definitely OK, as it is maintaining a steady idle. it's also been pulled and cleaned. Lambda sensor has been tested OK with an oscilloscope. I have not had time to put a new log file on here yet, but the first one can be downloaded from here: https://drive.google.com/file/d/1GewpzyoWZnGijrLBQZP-N-EXzeHirtj9/view?usp=sharing.

Thanks

JonB

[leopold]

Please take an other diagnostic tool like memsfcr, allans memsanalyser, mems-scan from haro or mems-rosco. This logfile is missing some important information.

[rubenste...@gmail.com]

Sample rate!   Thanks, JonB.    I should have thought of that given that the effect of low sample rate is so obvious with lambda voltage!

On looking for vacuum leaks, there is a very effective method using water (see at 4:40 on https://youtu.be/0rQn9YBMZIg)

As you have an oscilloscope, perhaps you also have the HT pickup leads, in which case you might want to confirm this is a vacuum-leak problem by looking at the end of the spark line in your snap-throttle (blip) test for the classic signs of a lean mixture (around 3:30 in https://youtu.be/SUShO72Grq8).       Whilst your symptoms seem to be classic vacuum leak ones, perhaps it would be best to confirm it so you know where to concentrate your energies?

Is your scope a PicoScope?

[Andrew Jackson]

On your logfile it's showing coolant temp ~ 100C and intake temp ~ 57C. That seems pretty high to me.
What are you seeing in the real world?