Ah ha! You see, that is a common misconception. A temp tower provides visual confirmation of whether the filament and the recommended settings agree. In my experience, they often don’t. Here’s an example from yesterday:
I opened a spool of Anycubic PLA I got for $12 during an Amazon promotion to test its performance. The spool had just come out of the dryer. Its weight dropped from 1132g before drying to 1126g after drying, but I suspect most of that loss was due to the cardboard spool.
I performed extensive calibration with this spool. Here’s what the temp tower looked like. When evaluating the temp tower, I always start by inspecting the sides without temperature markings to remain unbiased. I then note how many markings align with the best results. Sometimes, it becomes a trade-off, as it did in this case.
Three edges showed the best results:
Next, I flip it over and inspect the remaining edges.
Only one leading edge was worth noting:
The best result was, arguably, at 215°C—well outside the range stated on the box.
BTW: Anycubic is now off my list for future purchases. I have better options in the $10–$11 range, and they’ve since raised the price back to $14 for what I consider an underperformer. While it wasn’t bad enough to warrant using Amazon’s return policy, I won’t be buying from them again.
____________________________________________
Suggestions:
You still haven’t specified which type of PLA you’re using. I’ve personally experienced numerous issues with both silk and matte filaments, especially regarding layer adhesion after pauses. While this might not apply to your situation, my workaround has been to temporarily increase the nozzle temperature by 5–10°C after pausing, then lower it back down after the layer change.
Additionally, if you have a chamber fan running, you might want to disable it momentarily. As @h3li0 suggested, speed could also be a factor. Switching to quiet mode for one layer might help as well.
Let us know if you find a remedy so the community can benefit and good luck. 