I don't know that is is take 25, but there has been a whole lot of them.

I've got a really good feeling about this one!

It's tough, the seal portion is self aligning, and I was able to get some pretty good tolerances (with iteration).

This one pulls on the mixer shaft with spring tension not transmitted through the impeller drive motor like some of the previous versions, so no more spending hours modifying motors and aligning them. It will be flexibly coupled to the quad motor with a pin/slot drive, and is tightly coupled to the impeller with a set screw.

I drove it 'dry' with my drill press at 3000 rpm with no lubricant for a couple minutes and it didn't self destruct or eat the seal. That's a lot better than some designs I've come up with. :)

The stainless steel impeller drive shaft size was reduced from 2mm to 1.15mm as I needed to increase the sealing surface area. On initial testing, the spring tension needed to self align was too high for the smaller surface area seal and instead of aligning, the seal ate itself. I do have some reservations about the smaller size shaft. An accidental cold startup of the mixer drive motor could result in a snapped shaft now. This is a real possibility as there is a bug in the software which causes the mixer to spin while homing after printing. I'm looking at decoupling the drive signal until at operating temperature the next time I'm in 'software developer' mode.

The bearing I scavenged from a large, high output, PC fan so it should withstand the axial loading for a good long time.

The seal in this version is glass fiber reinforced PTFE encased in brass, which was pre-compressed while at the upper temperature limit for the PTFE. (actually, I just cooked it with a lighter and pulled on the shaft) Why? The first time I heated each virgin seal, they would creep significantly and my mixer would end up being recessed by about 0.4mm. There it would stay for the most part, so to move it back to where it should be, I made it adjustable. The adjustment screw was in the way for this version, and you couldn't adjust it without taking the whole thing apart anyway, so it went by-by. The glass fiber reinforcement also makes the PTFE a lot stiffer, at the expense of increased wear which has yet to be quantified.

The new impeller for this version has a brass sealing ring which, despite my best efforts at fixing it on the shaft with extreme temperature differentials, popped off while I was trimming the impeller shaft to size. After a couple of attempts at machining an all stainless version, I gave up and put the brass one back on. It doesn't spin on the shaft without plastic in the chamber, but it might with...

The new ESC and a slightly beefier BLDC motor will be here soon. The previous motor had plenty of power, but I'm thinking that the larger motor should respond more linearly at low RPM. The old motor had a 3.1mm shaft, while the new one has a 3.0mm shaft so I haven't been able to do the final assembly or any real tests. I'm patiently waiting as my motor makes its way here from Arkansas.

One significant issue has presented itself. I did not have sufficient clearance between my PTFE feed tube passages and the bolt passages which retained the nozzle. While I was disassembling it after the last run, I allowed it to cool too much and some PLA which had made it's way onto the bolts took some of the AL with it. That's 6061 T6 aluminum which was just relocated by plastic. I cleaned the bolts and tapped out the ports again. I still have enough force to seal the nozzle, but it may leak now and the only fix will be to machine a whole new part. This was another time consuming side effect of being low on thermal compound, which also seems to work pretty well as 'anti-seize'. The bearings in my drill press are now making an intermittent 'replace me now' screeching sound, but as I've been abusing it with lateral loads for a decade, I can't complain really. I'm just going to run it like it is, until it leaks badly enough that I can't. Moral of the story? Don't cheap out on thermal compound.