Starter: grind, bind noise

[AMC of Houston]

Just to document for posterity, I cross-referenced the 2 starter bolt part numbers to the AMC Standard Parts Catalog.  Here's the poop:

J4001884: BOLT, Hex Head, with Conical Spring Washer, 3/8x16, 1.190" long, Grade 5, plain finish

J4006684: BOLT, Hex Head, with Conical Spring Washer, 3/8x16, 1.375" long, Grade 5, plain finish

These are not shoulder bolts.  I recently had my original starter off my '85; and there was no play between the bolt holes and the bolts.  No "adjustment" at all.   I still wonder if your starter nose is off of something else (Ford application?).  Did you see any part number stamped on the nose anywhere?

Probably doesn't help; but there it is.

[Canoe]

Which year(s) are those part numbers from? I know the (some?) '88 uses a different starter.
I have a vague recollection of conical spring washers on my '81 wagon's starter.
The length of threads in the housing for the lower bolt is longer than that of the threads in the upper.

For this starter, its nose is a longer nose and longer Bendix than the standard Ford ones, but I've no assurance it's the AMC setup. There is significant play between the holes and the 3/8-16 bolts, thereby relying solely upon the locator plate for registration of the starter to the centre of the flexplate, with the starter being able to move, hence exert force, hence wear, on that opening in the locator plate. 

Update on noise.

• 1/15 starts, starts with noticeable noise
• 1/15 starts, starts with a pause, then resumes turning over and starts (going to check electrical cables to eliminate that)
• the rest of the starts are quick and low/no noise - wonderful

The "shim" is provided by two fender bolts, not a proper full shim gasket. That may be a factor, but at the proper torque of 18 ft.lbs., it shouldn't be?

It sat for six or seven years with a full tank of gas. Finally ran that down to 1/4 and filled it up again. Running smoother now with a little more power, so I redid the vac gauge dizzy setting. Doesn't seem any different. We'll see how it drives tomorrow.

[Mitch]

Not sure if this well help but...

https://www.hemmings.com/stories/article/ford-amc-motorcraft-starters

[AMC of Houston]

Those 2 bolts I mentioned are indeed only from 1979 thru 1987.  Parts book says '79 thru '83 take 2 of the J4001884 bolts.   Is your car an '88??   I'd be happy to look up the '88 bolt info just for grins.

[Canoe]

No, mine's an '86 and I'd already looked up the parts when I found the discussions about the mythical shoulder bolts.

[Canoe]

Not sure if this well help but...

https://www.hemmings.com/stories/article/ford-amc-motorcraft-starters

A little. That's what I quoted from in my first post.

[Canoe]

What noise was left was decreasing.

Until today. The noise is back and with a vengeance. Worse. So bad I drove home without stopping anywhere so I wouldn't have to start it again. Starter is firm in place. Will have to investigate.

[Canoe]

... Will have to investigate.

Startup, it did not want to rotate after engaging. Third try it did. Drove to my garage. Top shield of the cat fell off and got crunched along the way.

Pulled starter:

• 2/3 of the Bendix teeth are ripped off, exposing the shaft all the way back to where the teeth start.
• One of the washers that is on the shaft at the nose fell out, ripped open and mangled. 
• I'll have to pull the plugs and rotate the engine to inspect the flexplate in situ. 
• Will need to get all the pieces out. Really hoping a magnet helps enough with that.

Not a happy camper.

Torn between a new/reman starter vs. getting a Bendix and rebuilding this one.

I'm thinking the stress it was put under ripping those gear teeth off means putting a new Bendix in would be a suspect fix?

Still don't know the original problem; only that going to lower stud & 'shoulder' eliminated most of the noise most of the time. Also, the flexplate may have had a fault.

p.s.
Got a air cleaner assembly in 5 1/8" to put on top of the 2150. There were only two in the scrap yard to choose from. Neither were offset, so I'll have to do a little fab work. If I had nothing else to do, that would be fun.

[Canoe]

photo of destroyed starter Bendix gear
the gouges in the opening aren't encouraging - probably some teeth damaged by the Bendix gear bits flying around...

[Canoe]

Fishing for Bendix gear bits at the bottom of the bell housing using a magnet. The idea is to drag them to the starter hole opening to get them out of there. I don't want anything left in there to bounce around and interfere with the flexplate and new starter. And if there's a flexplate tooth found, I'll need a flexplate for sure.

Two 3/8" rare earth magnets didn't work. I can feel them grab steel bits, but they loose their hold on the bits when I try and drag them upwards towards the hole. The Al housing is too thick, like an air gap, to maintain a hold on the bits. I'll try a pair of 1".

[Canoe]

Two stacked 1" disc rare earth magnets with a steel backing disk were able to attract and drag the broken Bendix gear teeth to the starter opening. Lots of little steel spinters. Magnets were very effective at pulling those out of my fingers. I retreived the four missing gear teeth. All nine accounted for: five still on the starter and four pulled out.

Pulled the plugs and had my nephew rotate the crank while I inspected the teeth on the flexplate. All its teeth are there. No obvious defects. This was not a high-quality inspection; I need to do better, but can't see how I can do that without pulling the flexplate out. I also need to check the locator plate to tooth distance for all teeth, to check for warping that wasn't obvious to the eye.

Checked that the locator plate is flat by using a steel rule.

• Possibility that the starter was turning before the Bendix was adequately extended and engaged.
• Possibility that the DIY threaded sleeve added to the lower stud (replaced bolt) was not adequate - may need the same on the upper bolt.
• Possibility that the two fender washers as shims are inadequate to hold the torqued starter from bending away from the flexplate teeth. Will remeasure the locator plate surface to flexplate teeth distance, and bench test the new starter for extension-before-turn and measure its extension, to ensure those are in spec.

[Canoe]

Have starter in hand.
Have the small plate off and will do a close inspection of the flexplate.
Will try and get some paper in place to test for space inside the starter mounting hole, to see if I can get a piece of steel in there with a 4.5" 3.25" hole cut in it to provide the correct hole size and location (the hole in mine is worn elongated lower). I don't know if there's enough original hole to allow for correction location. If there is, I figure I can JB-Weld and pin it in place and easily clear the flexplate. P.S. then fill the original plate up to match the inserted plate with the correct hole size.

[Canoe]

Finally have time to get back to this.

I'm not liking using the partially chewed flexplate. Seems my choices for a replacement for the J3232138 are:
- ATP Z-101
- Pioneer FRA-108

Anyone have any advice on the quality of the ATP vs. the Pioneer?

[Canoe]

Finally got back to this today.

Barely any room for my intended 1/8" 'gasket' mounted to the inside of the location plate, with a correct sized starter hole to fixed the enlarged hole. But it can be done, and extended towards the flexplate centre outside of the starter flange so it can have two bolts securing it in place in precise registration while the JB-Weld dries and staying captive.

My starter lower bolt to stud with threaded-on 'shoulder' mod won't work with the new starter.

• The old starter's lower bolt hole in the starter flange is 11.0 mm. Mod would mount this starter closer to where it should be, in spite of the enlarged hole in the location plate. Don't know if this was/wasn't good enough, or if the starter blew its bendix gears due to binding from damage already done to them or to the flexplate teeth.
• That hole on the new starter is 9.6 mm. Mounting bolt fits in with minimal play, and even less once the bolt is in far enough that the unthreaded shank (like a shoulder) is in the flange. I'd expect this means the new starter will bolt in much closer to the correct location re. the flexplate. It also means I get to remove the stud I made and installed in the lower hole in the bell housing...

I think it's time to measure the upper hole in the bell housing... If it's tight enough to the bolt shank, there's no point in the fixing 'gasket', as the difference in mounting position would be under 1 mm, possibly under 0.5 mm.

[Canoe]

Seems the whole starter issue may have been self-inflicted. The Bendix gear rested below the flexplate and it would  manually extend to fully engage the flexplate teeth. But I never bench tested the extension using power.

Now I see that the old starter's Bendix rest position is well below that of the replacement starter. Measuring the new starter, its below the flexplate teeth, but just. I may need a gasket to ensure it clears. It will easily have full engagement.

It's plausible that the old starter simply didn't extend far enough for a full engagement under voltage. But in its damaged state, I don't think it's valid to measure that now...

On the mounting of the new starter,
I got two stainless 3/8" 1 1/2" bolts, which have an un-threaded shank next to the head. The shank is under 3/8" and tapers somewhat narrower towards the start of the threads. With just a little play, they would act fairly well like a shoulder bolt for the flange hole & bell-housing hole sizes. I considered wrapping the shank with some thin wire for a tighter fit.

Encouraged as a shoulder bolt substitute to provide starter registration relative to the bell-housing, to see if there was better, I went to a different store. I got two more of the same stainless bolts: same size, even same manufacturer and same stamps on the head, but they have a slightly different colour tone. These shanks start barely under 3/8" (so close I can't measure the difference with the plastic calipers) and tapers but less slope and to less narrow than the first pair - nearly square to the axis and off by less than can be measured with these plastic calipers (although they allow the taper to be detected). These second two have distinctly less play in their respective holes than the first pair. These will very closely register the new starter to the bell-housing.

Now for the detailed inspection of the flexplate...