Bypass issues with R2 clone

[Pr1me]

Hey all, long time lurker and big fan of all of the build logs posted here, the experiments everyone's done, and the research that's been put into Phil's fantastic design. (You're the man, Phil!)

I've done a ton of reading over the course of several months, so I'm familiar with the general principles and whatnot that makes these things tick.
Armed with that, I set out to build a design pretty close to Retired2's.
Unfortunately, I'm experiencing some disappointing bypass issues.

Specs:
-4" PVC lines with doubled 45s, wyes, etc.
The closest tool is my miter saw, where dust only needs to travel through about 14LF of pipe to get to the DC.
-23"-ish diameter, fits over a 32 Gal trash can
-6" chamber height
-1-1/8" drop slot
-4" round to 3-7/8" x 5-7/8" rectangular transition
-HF DC inlet modified to fit 6" inlet
-DC exhausts to free air (waiting for air filter to arrive)
-6" outlet to DC
-Despite intending to build it to match the rotation of my DC, I accidentally built it the opposite, so I also installed fluorescent bulb cover tubing in the outlet piping like R2 did as an air straightener.
-Flange of the bellmouth is 3" above the baffle (which makes it 3" from the top)

I attempted to fabricate an integral bellmouth on the pipe itself by heating a section of 6" ABS and forcing it over a turned MDF plug, but didn't realize until afterwards that I'd caused the pipe immediately behind the bellmouth to bottleneck inwards.
Clearly I should've made the plug deeper to support the pipe wall during forming.
I'm not sure if it's enough to negate the purpose/benefits of the bellmouth or if it impacts airflow enough to matter, but figured it's worth mentioning.
I don't believe I've achieved a large enough flange for the bellmouth, but since R2's experiments show that almost anything is better than a squared end, I was happy with my meager attempt.

I fully expect some fine dust to be ejected so I have no illusions there, but I figured it would be able to handle basic sawdust from a tablesaw, miter saw, etc. with a little more success than I've had so far.
It appears that the dust does several revolutions and half is deposited while the other half is ejected.

Are there any obvious errors in my build that would cause this? My first thought is to lower the outlet pipe, but I haven't given that a shot yet.
Could anyone give some suggestions on what I could change to improve it? Maybe increasing the drop slot width?

[Pr1me]

Update:

I discovered that the seal between the trash can and baffle was severely misaligned, allowing air to be sucked up around the trash can edge instead of being contained within it.
Once I fixed that, it's working perfectly now.
Posting in case anyone else has issues and hasn't thought to check that minor detail.