I think your assessment is correct, though.

I wonder if whatever part of ABS (the A, the B, or the S) provides the inter-later bonding boils off or is destroyed from either heat cycling or granulation.

I thought of this as well, since the smell it gives off when heated has to be composed of something.

Here are some relevant snippets from the Wikipedia page for ABS.

• Heating up ABS during production (for molding or extrusion as well as in 3D Printing) can produce fumes of acrylonitrile.

• The nitrile groups from neighboring chains, being polar, attract each other and bind the chains together, making ABS stronger than pure polystyrene.

• The proportions can vary from 15 to 35% acrylonitrile, 5 to 30% butadiene and 40 to 60% styrene.

So, what I get out of this is that the "A" (being the nitrile) very well could be the thing that improves the layer bonding with its polar property, as well as possibly being a considerable part of the fumes from heating.  And if the original ABS happens to be on the low 15% side of "A" to start with, losing fumes from printing, extruding, and printing again would likely lower the % of "A" below the point of it being effective at bonding.

If this is the case, it might be best to find ABS with the highest % of "A" as possible if you are planning on recycling failed prints.

Just my novice opinion.

Printed parts are strong in the xy plane, but inter-layer adhesion is questionable

I had the same experience.

I wonder if whatever part of ABS (the A, the B, or the S) provides the inter-later bonding boils off or is destroyed from either heat cycling or granulation.

$1.5/lb shipped isn't bad!

Pellets go in nice and white, but come out slightly gray. Interestingly enough, the darkness of the filament is inconsistent despite the same temperature and extrusion rate. I also tried varying the extruson temperature from 170 to 190, and it made minimal difference in the color. Printed parts are strong in the xy plane, but inter-layer adhesion is questionable. On the other hand, the box was $25 for 16 lbs shipped, so... 7 reels of filament for the price of one ain't a bad deal

Here's some pictures of prints I've made with this filament: https://imgur.com/a/YuBIi
I'm printing the cross-sections for a foldable kayak. The prints came out great, aside from slightly weak adhesion. In this case, it doesn't really matter since the parts are nice and vertical. Also, I've acetone-treated almost everything, so I doubt it will be a problem.

It's tough right now because a lot of different people are using many different kinds of plastic with different shredding techniques with different claimed results.  On top of that, people are using different printers with different printing temps. More posts like this with specific information and results will really help figure out the best way to recycle our prints.

Anyways, I hope to get some plastic e-waste soon to send over to you to test.  Curious to see what everyone else thinks about the recycling process.

Think the poor print quality was more likely caused by the heat cycles of the extruder or the heat from the regrind process?  Maybe using a slower method to regrind would cause less thermal stress on the material.

The granulation process I use is what the professionals use. It is a polymer granulator specifically for granulation of plastics. Someone tell me if I am wrong, but I do not think a piece of plastic getting chopped up several times by a spinning metal blade creates a new heat cycle. Also, after and during granulation process, I never smell any burning or heated up plastic (trust me I know what burning ABS smells like), so this leads me to believe that the plastic has not been through another heat cycle.

Or maybe elmoret forgot to take out the silca gel pack, and ran that through the extruder.

I remembered to take it out. That would have been interesting, though.

do you have a contact with ultra high magnification digital microscope to compare a thin slice of each to maybe see what caused the grey discoloring? that would also be interesting for future development of this process. and maybe determine a viable life cycle.

Heat causes it, it looks exactly like when I leave filament sitting at 190C in the Filastruder for an extended period of time. Even under magnification, it just looks grey - no black specks or anything.

Some of it is caused by burnt (charred) plastic globs that were on a few failed prints. This gets chopped up by the granulator, then mixed in the melt zone of the Filastruder.

do you have a contact with ultra high magnification digital microscope to compare a thin slice of each to maybe see what caused the grey discoloring? that would also be interesting for future development of this process. and maybe determine a viable life cycle.

On a few of elmoret's prints, there were some questionable scorched ABS bits. I tried to avoid this, but those tiny bits may have discolored the filament. To fight the discoloration, I use black colorant to mask color imperfects.

Or maybe elmoret forgot to take out the silca gel pack, and ran that through the extruder.

Extruder type: Hybrid Experimental Filastruder with overdrilled (+0.2mm bore) 1.75mm melt filter, full barrrel insulation, horizontal, overvolted motor, roughly 1800 hours of use so far.
Printer type: SD3, E3D v6, Glass bed.
Extrusion temperature: 192°C
Extrusion rate without colorant: 24-30inches per minute.
Average diameter of filament: 1.69mm
Tolerances: +/- 0.05mm (same as observed with virgin MG94 ABS)

Photos:

Here's what it looked like when I got it back from Ggalisky. As a note, he was very quick to granulate and return - I on the other hand let it sit on a shelf for a week before finding time to run it.

It extruded just fine, nothing exciting to see. It fed without problems, I was actually very impressed with how uniform in size the granulate was - this is far better than what I've seen out of Filabot/Filamaker shredders.

Here's a comparison between filament made from 100% virgin natural MG94, and 100% regrind natural MG94. You can see the latter is darker and more grey, I believe this to be due to carbonization since it has gone through a few heat cycles.

Strength testing:
Sabic MG94 ABS has a UTS (tensile strength) of 35 MPa.

I hung weights from a piece of filament made from 100% virgin, and a piece of filament made from 100% recycled. In both cases, the piece of filament supported 7kg of weight, and failed at 8kg of weight. This works out to failing somewhere between 32.7MPa and 37.4MPa, with the predicted strength from Sabic falling right between those numbers. This means that at worst, the recycled filament has 90% of the strength of virgin ABS, but could be closer to 100%.

Other material properties:
100% recycled ABS seems to be just as flexible (not brittle) as virgin ABS. This is very promising.

Printing:

Unfortunately, i could not get the 100% recycled plastic to have the same inter-layer strength as virgin plastic, resulting in severe delamination. I would estimate less than a third of the layer bonding of virgin material. This makes sense, as everything I've read states that 100% recycled plastic is not the same mechanically - it turns out the tensile strength is similar, but the layer bonding is not even close.

I would recommend no more than 1/3rd recycled plastic at this time, at least with ABS.