Surface E In the diagram...
Good start.
Removing material from the locating pad is what you need to do to set the deflection. As you lower the height of the locating pad the more deflection there will be. Each time you remove material you need to test fit the extractor to see what the resulting deflection is. You need to bend the extractor a little each time you remove material from the locating pad so that when you shove the extractor back into the slide the locating pad will be in contact with the tunnel. If it's not in contact with the tunnel, you will not be able to get a true reading of the amount of deflection.
The most common way to screw this up is to remove too much material from the locating pad resulting in too much deflection. You don't want more than .010" of deflection if at all possible. I've been there, done that, and have the t-shirt to prove it. Learn from my mistake.
Then before I was done, I'd taken enough material away that there wasn't any tension.
This would be unusual. Even if you completely removed the locating pad, the extractor could be bent to contact the case rim and apply pressure. In fact, with no locating pad you'd have way, way too much deflection and no matter what the tension was on the extractor (within reason) the pistol would have functional problems.
Not long ago I tried to fit a couple of Wilson extractors to a couple of Caspian slides but the locating pads were so short that even right out of the box these extractors could not be adjusted for deflection. In other words, they had way, way too much deflection even with their unmolested locating pads. Those extractors could not be fit to those slides without doing some fancy welding to add height to their locating pads.
I had to bend it - Per Jerry Kuhnhausen's specs.
So now you know the extractor will always have to be bent so that the locating pad is in contact with the tunnel wall. Forget Jerry's specs. Set the deflection first and correctly. Then add tension. With minimal deflection you'll have a tough time bending the extractor enough to compromise feeding.
I didn't remove any material from the hook though.
Remember to check that the edge of the hook claw isn't in contact with the case as you go through the fitting process. It's not unusual to experience this contact as the height of the locating pad is reduced. The claw depth should be as long as possible without contacting the case. A range of .030" - .034" is considered good.
You already know that setting tension comes after setting deflection. Also, after setting deflection you can set the claw depth. Stop short of the range I noted and test fit.
Don't just file the claw down to .030" at the beginning. It's an iterative process. Step by step. Tiny bites at a time. Once the steel is gone, it's gone forever.
I taped some pieces of fine hacksaw blade together to make a tiny file and get into the groove.
You're talking about surface
C the tensioning wall, right? Be careful. You don't want a 90 degree angle where the tensioning wall and the hook meet. That will weaken the hook and can lead to the hook breaking off from the extractor.
But before, I was running it with clearance between the case rim and the groove of the extractor, as the extractor came out of the package.
Yeah, that's obviously not right. The case rim and the groove (tensioning wall) must make contact. They must be the only contact point between the case and the extractor.