I just rebuilt my X's marelli starter solenoid this weekend, so ill try to be of some help. When it failed, it would make no noise or anything- but the voltage gauge would dip upon turning the key. At the time, the starter signal wire to the double relay was not connected, so there was nothing else to pull current when turning the key, indicating that the solenoid was getting power but not engaging.
I had such little confidence in the repair that I actually ordered a gear reduction starter before even reinstalling it, but that was pretty stupid considering that it now acts as if it's brand new. Before the repair, not only would it often fail to even engage (which was the issue that drove me to remove it), but it was also fairly sluggish. I assumed it was just an old and tired starter motor but I was wrong.
After pulling the starter, there's two bolts that hold the solenoid assembly onto the starter body. Once those are out and the wire from the starter motor itself to the solenoid is undone, the solenoid assembly should slide right out, leaving a metal cylinder in the starter body.
Once the solenoid is off, remove the two screws on the terminal side of the solenoid and it should loosen up once the seal is broken.
If you have a soldeirng iron on hand, this would be a good time to remove the dab of solder on the trace from the solenoid trigger connector. I did not have a soldering iron on hand, so I crossed my fingers and used force. The internal wire seperated right at the end, leaving a pointy copper wire and a nice hole for it to seat back into upon reassembly. Maybe I got lucky.
By now the cap should be quite loose, and the two copper wires on the other side of the solenoid should have enough slack to allow you to move the cap off of the end of the body, leaving room to service what's inside. These wires could likely be desoldered too, but, I didn't.
In my case, after testing before and after, two things were happening.
1- The plastic pieces which the contacts slide on when the solenoid is triggered were dry and dirty with what I presume is material splatter from the contacts.
2- The contacts themselves were in such bad shape that the resistance of the circuit, even when completed and the starter motor is engaged, must've been quite high. All of the current that the starter motor gets obviously travels through these two contacts in series, so if either of them is bad, it's reducing your cranking power. One contact will likely be in worse shape than the other as the side which contacts first (since lets be honest, nothing is perfect) still has an open circuit and will thus not have any arcing...the same cannot be said for the side which contacts last.
Both of these issues are easily fixed by removing the contact pin and contacts from the electromagnet section, cleaning the plastic sliding components, removing the tarnish on the copper contacts using a magic eraser ( a Scotch Brite would likely work well too), and lubricating the plastic sliding components with plastic on plastic lubricant. I used Berulub FR43 (500c cabrio top grease) as I know it will last longer than the assembly will and is well suited to the application. As a last restoration step, clean the other end of the contacts on the plastic cap itself- be careful not to apply too much force if you still have wires hanging on to it. Drop the contacts back in ( I did it 180 degrees offset from how it came out so they were mating with a new surface) and replace the cap. It'll take a few tries to get everything lined up, but eventually, it should go together without too much effort. The cap thankfully contains guides for the solenoid trigger wire, so it ends up back at its solder joint. I did not worry about resoldering it as the fit was so snug in the terminal.
When reassembling everything, apply a bit more plastic lube to the inside of the electromagnetic cylinder, as this is a plastic on metal interface. As an extra step, I applied a tiny amount of wheel-bearing grease to both the starter gear and the pin it slides on, and the hinge down inside where the solenoid sits.
All in all, it's likely less than 30 minutes of work, and in my case, yielded a starter that turns the engine over more than twice as fast as it did before (sounds new) , and requires much less current for the solenoid to engage. It has once not immediately engaged, when the cooling fan, headlights, brake lights, and blower were on... but this is an extreme case IMO- simply letting off the brake gave it enough voltage to fire right up without issue.
This was too damn long for such a simple DIY, sorry about that. Maybe I'm just a little cheerful still that the effort paid off when I had no expectation of it doing so.
