R400D: Curious misfire

[John Vine]

An update....

So far, I've  checked loom connectors, TPS connector, TPS operation (via Easimap), ECU connector and earth, crank sensor connector and wiring, lambda wiring.  Nothing untoward.  #1 COP connector was a little dirty but at worst that would only cause an intermittent misfire.

This afternoon, I did a 50-min Easimap in-flight logging session (beautiful 7 weather), with a 10-min stop after 20 mins or so.  The car behaved impeccably both before and after the stop.  Unfortunately, I somehow managed to screw up the logging as there was practically nothing in the log, so I'll need to repeat it.

JV

[aerobod - near CYYC]

Hi John, the main thing with Easimap logging is not to move off the page with the panels you want to log before you have saved to a file, any page up or down will effectively purge the stored info and start the logging over on the current page.

Also check your "PC Logging Setup" is set to "Automatic - Log the parameters on the current page"

[John Vine]

Thanks for the tips, James.

I'd definitely set logging to "auto".  I think I must have moved from the initial page somehow. 

JV 

[John Vine]

Re #20:

As long as Easimap is left running on the laptop on the same page while the engine is off, it will resume logging when the ECU is powered up again, you will see the power cycling as events in the log.

Just a quick question, James.  Does that still apply if the laptop drops into sleep mode, with a Windows screensaver showing?  I ask as I noticed it had done so as I set off from my stop.  For info, I'm using an ancient Dell Inspiron under Vista for this.

JV

[aerobod - near CYYC]

Hi John, if the screen goes into screen saver, no problem, but if the PC suspends/sleeps, that would be a problem as the communication to the USB interface will also stop, it needs to stay active with the Easimap software. The ECU can be shutdown or fully powered off then powered on again without affecting Easimap or the interface, though.

I've used a standalone battery inverter in the passenger footwell in the past to power a laptop with a dodgy battery, but you want to ensure you don't use an inverter plugged into the Caterham 12V socket, as you could end up with a ground loop and potentially damage the Easimap interface, PC or ECU.

[John Vine]

Thanks, James. Helpful as always!

JV

[John Vine]

A further update...

I completely drained the tank (Intermarché E10) and refilled with V-Power E5 (£££ - ouch!).  I'd already fitted new plugs (NGK TR6AP-13, re-gapped to 1.0mm).  My 30-mile run went perfectly but, unfortunately, I contrived to lose the log again.

Realising that I'd foolishly changed two variables at the same time, I put the old plugs back (also NGK TR6AP-13, re-gapped to 1.0mm) and repeated the test.  This time, I managed to record a 40-min non-stop log.  I didn't stop midway in case I lost the entire log through some ham-fisted intervention.   (I plan to do a start-stop-start log next week.)

The car behaved perfectly - apart from a very brief hesitation as I accelerated (gently) out of a 30-limit.  The various Easimap traces seem fine, and I couldn't spot anything untoward.  However, should someone more knowledgeable be prepared to run an eye over it, I'll happily pass on the DAT file.

I've now put the new plugs back in.

So, at the moment I'm no closer to diagnosing (or even repeating) my original problem.  Maybe my upcoming start-stop-start log will reveal something?

More news later.

JV

[StevehS3]

Hi John, could it be a clogged fuel filter? Is it of sufficient mileage/age to need replacing? 

[John Vine]

Re #33:

Well, I don't think so.  When the complete cut-out occurred, I could still rev the engine very freely on dipping the clutch.

As it happens, I took the pump out (with integrated filter) in 2018 to adjust the float arm, and inspected the filter at the same time. It had almost nothing in it.

JV

[John Vine]

One thing I didn't mention earlier...

The old plugs looked more white than light tan (23K miles).  Is that relevant?  In the "old" days, that colour would have suggested a lean mixture to me.

JV 

[aerobod - near CYYC]

Hi John, as long as every plug is even and there are no black speckles or sign or erosion/melting of the plug electrodes, the lighter colour should be fine. Mine tend to be quite light tan, too, but my mixture is slightly rich under high loads (down to a Lambda of 0.84). If you put the plugs next to a sheet of white paper, they still are likely to be quite a bit away from white.

[John Vine]

Thanks, James.  That's encouraging.  I'll post up some photos of the plugs tomorrow.

JV

[John Vine]

Plug photos:

They all look even in colour, albeit a bit "white".  #3 shows a little black "staining", but I don't know if that's significant:

JV

[aerobod - near CYYC]

They are quite similar to mine, John. Sometimes I see a shadow of the electrode on the insulator depending on the plug orientation due to the thread. "Very light grey" is how I would describe the colour.

[John Vine]

Thanks, James.  Reassuring.

JV

[John Vine]

I'm beginning to wonder whether the misfire could be due to an internal problem in the MBE992 ECU.  If it happens again, I'm toying with the idea of asking SBD to take a look.  However, if the 992 is now obsolete, no doubt they'll suggest I replace it with a 9A4!

In the meantime, I think I'll check that the ECU pins are clean and correctly located.  Does anyone happen to have an MBE992 pinout list (I can't find one on the SBD website)?

JV

[Jonathan Kay]

Would that be in any of the Caterham factory wiring diagrams?

Jonathan

[John Vine]

Well, sort of, Jonathan (and a good suggestion).  My own Assembly Guide includes a wiring diagram for the ECU with wire references (such as NW58 on Pin 36 ending up on coil #1).  But it's a royal pain to trace them all, made worse by the poor print quality of the diagram.

What I'd really like is the 992 equivalent of this 9A4 list. Of course, it could be that the 992 pinouts are the same anyway!

JV

[Jonathan Kay]

"What I'd really like is the 992 equivalent of this 9A4 list(link is external). Of course, it could be that the 992 pinouts are the same anyway!"

Is James suggesting here that they are the same, or only the pins used for CAN?

https://www.caterhamlotus7.club/comment/2112434#comment-2112434

Jonathan

[aerobod - near CYYC]

Hi John,

Pinouts for 9A4 and 992 are the same, but function on pins 16,17,19,25,27,33 & 34 can vary as those pins are all programmable using Easimap.

When I swapped my 992 for a 9A4 I didn't do any repinning, but I did add an additional wire in Pin 28 for Baro sensor function initially with the 992 which then showed the right value in Easimap, it just didn't do anything with it due to the locked Caterham map.

[John Vine]

Thanks, James.

JV

[John Vine]

Time for an update...

I'm pleased to say that, with Easimap's help, the problem is now 100% solved.  It was the lambda O2 sensor, aided and abetted (maybe) by a corroded connector.

So, the prize goes to...

Olly Amos (Amos91) in post #3.

 

A couple of days ago, I recorded an in-flight Easimap log of around 45 mins.  This is a snapshot:

The blue trace is Lambda voltage (left axis), and the brown trace is engine speed.

Note the long spells where no sensor activity is apparent.  During these spells, the engine cut out completely.  The spikes in engine speed are when I dipped the clutch and revved briefly after one of these cuts (the peaks are 5770 and 5990).  The car was becoming all but undriveable.

Back home, I removed the sensor and inspected the wiring and connector.  The wiring was sound but the connector terminals were badly corroded:

I cut off all the terminals (male and female) and crimped on new ones.

I then recorded another log (this time, idling from a cold start).  It was clear that the corrosion wasn't the cause of the problem, as lambda activity should register once the coolant reaches 60C:

Purple: coolant temp
Blue: lambda voltage
Brown: engine speed

This morning, I replaced the sensor and recorded another warm-up:

Lots of activity at 60C this time!

A 20-mile road test confirmed that everything was back to normal.

But I'm still puzzled that a failing (or failed) O2 sensor could cut ignition 100% at times, rather than simply cause a lot of rough running and misfiring.  Can anyone shed light on that?

JV

 

[Jonathan Kay]

Thanks for adding the answer. 

Jonathan

[aerobod - near CYYC]

Hi John,

I think the voltage of the duff sensor dropping to a low value (that it seems to do when you have the "cuts") could cause a significant increase in the amount of fuel due to the ECU detecting a lean mixture due to the low Lambda voltage, the excess of fuel in that case would cause the plugs to foul temporarily. Using full throttle would overide closed loop control and default back to the open loop fuel value.

Your final injection time plotted on top of the Lambda voltage and engine speed  together with Lambda control state would show if that was the problem.

[John Vine]

Thanks for that, James.  So, during the cuts, more a question of too much fuel rather than too little!

Are "final injection time" and "Lambda control state" specific Easimap panels for a 992?  I can't see them on my recording.

JV