MBE 967 communication help

[Paul Deslandes]

You need to ensure that when setting up the idle speed, the ECU is looking at site 0, which is defined by the TPS voltage being less than about 400mV.  To do this:

1. back off the TPS so that the voltage is lower than 360mV, say 300mV

2. start engine and warm up to operating temperature

3. adjust idle using throttle setting screw to correct idle speed

4. stop engine, turn ignition back on with engine not running.

5. adjust TPS voltage to 360mV by adjusting the TPS body rotation.  Do not touch the throttle idle screw.  Keep checking TPS voltage as you tighten the clamping screws - it may take a few goes to get this right so that the TPS is locked and the voltage correct.  Don't overtighten.

Start engine and check that idle revs and TPS voltage are correct.

Paul

[Smithy77]

Just thought I'd provide a quick update for anyone interested...

My visit to see the magnificent Paul D (thanks for all your help and hospitality, Paul!) was a success, although it wasn't without problems. We (well Paul) had no problems splicing the serial connector into the loom, obviously, but we immediately hit a wall when we tried to talk to the ECU, just as someone in the thread previously alluded to. Neither of Paul's laptops (An XP laptop with USB+adaptor, or Win7 laptop with serial) were able to communicate, although the computers did recognise it was connected to a 967 MBE ECU. Luckily, I had brought with me an old Dell laptop with XP and serial port, so we installed Easimap 5 and we got in first time. We then hit the 2nd wall which was the Caterham encrypted PIN code that was required. After some extensive Googling we manage to find a method of cracking the code which involved copying the map file to hard drive, reading the file with a hexadecimal application, and then converting 2 bits of coding from hexadecimal to decimal to give us the 4 digit PIN - Bob's your uncle!

We connected up a wide band lambda into primary 4 to get an idea of fueling and it was running alarmingly lean. This is contrary to Track n Road's advice after doing the same thing when I was there. Not their fault mind you as I had swapped injectors 1 and 4 over in between measurements; When I first got the car home, and after being at Track n Road, I checked the spark plugs and plug 4 looked rich while the others looked lean. Suspected faulty injector 4, did the swap and indeed the colour variation of the plugs appeared to follow. This week I sent my injectors off for ASNU cleaning and got them back 2 days later. The result - the suspect faulty injector was actually in full health while the other 3 were 5% down on flow rate! All are within 1% of each other now. Things are starting to make more sense.

So whilst at Paul's we decided to add 20% more fuel to the map right across the range and left everything else alone. The result was quite amazing! Virtually no miss fire to speak of now and I would almost go as far to say it is driving as good, if not better, than when I first test drove it.

All that is left now (surely) is just a proper mapping session to get it all sorted, which is booked for next Friday. I've done a few bits ahead of the remap: injector cleaning; new fuel filter; new spark plugs; fit lambda boss in collector. I still can't quite work out how it went from driving nearly perfect on the test drive to being totally un-drivable in the space of a few weeks and not many miles. All I can put it down to is fueling and the injectors very quickly loosing their flow rates.

So much for a 'quick' update

[Jonathan Kay]

Congratulations to you and the magnificent Paul.

This week I sent my injectors off for ASNU cleaning and got them back 2 days later. The result - the suspect faulty injector was actually in full health while the other 3 were 5% down on flow rate! All are within 1% of each other now.

I don't understand that bit... why do they now match?

Jonathan

PS: 

After some extensive Googling we manage to find a method of cracking the code which involved copying the map file to hard drive, reading the file with a hexadecimal application, and then converting 2 bits of coding from hexadecimal to decimal to give us the 4 digit PIN

Are the FBI recruiting... :-)

[Smithy77]

Jonathan, the injectors now match because they were all cleaned!?? Or am I missing something...?

 

PS. It was a team effort ;-)

[Paul Deslandes]

Splicing in the serial connector was relatively easy, not least because of the beautifully tidy and powdercoated engine bay - what a pleasure to work on.  All but two of the required connections just needed pins/tails added to the connector, but the +5V and 0V needed to be cut into and new tails soldered in.

For anyone contemplating doing this themselves, or indeed any other loom mods, don't use cutters or a knife to open up the insulation, just very carefully cut away the sleeve with the hot tip of a small soldering iron (don't inhale the fumes!!).  This avoids any risk of cutting any of the wire strands; there are few enough of these without breaking any. 

Expose about 4mm of wire and tin with fresh solder.  Strip and tin a similar length of the end of the wire to be attached and solder as a 'laid on joint'.  Laid on joints are seriously frowned upon in industry as they don't have any mechanical strength, but they are really the only option in these circumstances and are ok so long as the joint isn't under any mechanical strain and the solder flows sufficiently to join the two wires. 

If working on a connection where the loom wire end can be pulled from the connector, slide a small piece of heatshrink sleeve over the new joint and heat with an airgun to shrink, being very careful not to melt the insulation on any of the cables.  Otherwise wrap with self amalgamating tape.

It's definitely best to attach the tails to the loom first, trim them to the correct length, and then attach the serial connector, having first sleeved the bunch with shrink tubing,  This ensures a neat job.

The issue of serial ports, XP etc was interesting and wasn't one I'd encountered before.  Fortunately Pete's old laptop did the trick and we were able to talk to the ECU.  What was odd was that when using the USB/serial adapter, the PC was 'seeing' the ECU, knew what it was but couldn't talk to it.  Apart from having to enter the correct PIN, we had to try a few configuration files so that the maps presented correctly.  With the wrong file the revs axis didn't render correctly but we found one that worked.

Glad it's working out Pete and look forward to hearing the result of the mapping session.

Paul