ducatitech.com

efi maintenance



As you may have read on the racing page, the 916 suffered a breakdown while competing at the May `01 wsmc race event. I then purchased the Marelli Diagnostic Software Tool to aid in troubleshooting. I've been meaning to look into this tool (pictured on the right) and this failure helped motivate the $399 purchase. Laptop computer not included. =)
    Symptoms:
  • sudden loss of power, as though the "kill swicth" was turned to stop
  • may be preceeded with backfiring
  • starter will run
  • fuel pump will cycle upon ignition switch turning to ON
  • all lights and electrical features work properly
  • bike may or may not "recover" after cool down time of unknown length
  • seems to be aggravated by high revs and/or sustained high revs

    Troubleshooting:
  • no indication of errors using Weber Marelli diagnostic interface
  • while removing each separate relay / box attached to the wiring harness not testable via the W-M diag tool, only one produced a similar fault
  • ECU swap did nothing for correcting the fault
  • When removing the pick up coil, the fault is replicated exactly
  • pick up coil Ohms out within spec
  • there is oil seepage around the pick up

I removed the pick up coil and replaced it with a spare on hand. The removed pick up coil was bulging slightly around its mid section. When comparing the two coils, the new one had white colored plastic and the old one was noticeably yellowed. Reinstalling the new one was a bit odd. I was a touch lazy and didn't want to drain the coolant and remove coolant hoses obstructing the mounting hole to get the best depth reading on the gear. So I put a piece of metal in and scribed it to note the depth to the hole. I then also compared the two pick up coil "heights" and noticed that my measurement of the depth was reasonably accurate, assuming the factory set the original coil to the right height using shims.

The workshop manual goes to good lengths to explain the efi system, but leaves something to be desired when it comes to mystery faults like this. For one thing, it does not say what the two black relay boxes are for under the seat. (You need to infer this from the wiring diagram and studying the wiring harness) They are daisy chained... the first feeds the ECU and the second relay. You'll notice two red wires going into this relay. When this relay fails (or is removed), the fuel pump continuously operates when the on/off switch is on, most electrical functions work, but the bike wil not start. When the second relay (the fuel pump relay) is inoperable, the fuel pump does not cycle when the ignition is on. There are three relays near the battery. One is the flasher relay, another (the smallest) is the fan relay, and the third is the "ignition relay". The third should be called the "ignition switch" relay, as when it is inop/removed, the bike appears dead, no lights, nada.

Due to the inconsistent nature of this fault, I'll have to "cane" the 916 a good bit to ensure that it is cured. The other liable part would be the wiring harness... good thing I have a spare.

Review of MDST tool
I went to install the software on a laptop, but it had no floppy drive, so I loaded it on a peecee and used the network to copy over the files. Install was a snap. Plugged in the hardware "key" into the parallel port (I don't care for "keys" but that's another issue) and the cable into the com port. Plugged the serial cable into the diag port on the bike's wiring harness and turned on the ignition switch. Note that it may be a good idea to pull the headlight fuse for long term testing to help save the battery. Really neat stuff... it uploads from the ecu everything... ambient temp, coolant temp, throttle position, revs, dc voltage, as well as a bunch of error checks that the ecu does. These error checks are both "now" and "historical". Oh, it also performs off line checks of the tach, injectors, fuel pump, coils, and can clear the ecu's error history.

When first hooked up, there were previous faults on one of the injectors, the tps, and the ecu rom. I thought there was the source for my track side fault. Then I tried out the other ecu with the fim 071 ultimap. Holy cow! The fim totally pisses off the weber-marelli ecu! All the scales and readings were blipping around like mad. It made me think of when you see "jamming" on tv. Then I tried the event technologies chip that was previously in my bike. It fired up a "current" "ecu rom error", so that is the source for that particular error. The tps error was likely caused when I did the adjustment, as you have to actually close the throttle bodies to do the adjustment. That, then leaves the #2 injector fault as the unknown quantity, but it does not explain the bike totally shutting off.

Flotsam
Apparently, some of this website's readers mistook this page for an overal troubleshooting guide for fuel injected Ducatis. In case you haven't figured it out, the above text describes what I consider to be a unique circumstance. If your bike is pretty new, you'll want to start off by checking that the hoses in the tank are secure. These have a tendency to split from what I hear. I've seen one pop of for no apparent reason. IMHO, poor running is caused by poor TPS setup.

Also, some have seen fit to decry my electrical trouble shooting. This I find most humourous of all the criticism received... Well, I think the electrical systems on Ducatis are very simple... a collection of wires and three handfuls of parts: inputs, outputs, and control devices. From my experience in things moto-electrical, technicians are quick to suspect the control box. I have always believed that the ecu is the most reliable component, though it may be the most complex and expensive.

Troubleshooting electrical problems involves determining the components involved then how they affect the afflicted system, then removing the fault. The above statement sounds simple enough, but there are whole fields of study devoted to this. One way, very useful in long linear systems, is to half step, where you go the midway point of a system and verify the condition. Here, I kinda stumbled around the periphery and found the fault quickly. =)