First of all, I wanna reinstate the importance of understanding shrinkage, and shrinkage differences, for FDM.
Shrinkage exists and it’s very powerful, it’s a curse among all current FDM machines. And the difference of shrinkage between each layer, is often the culprit of most of our problems. Believe me or not, it’s also why there’s no point of bragging how fast a FDM printer can print, because shrinkage would ruin it all.
So remember shrinkage, and shrinkage differences, because they would haunt us a lot 
So starting from layer time view. You can see that the top surface layers are taking exceptionally long time to print.
The Internal Bridging Layer took a whooping 2800s to print, which is why it’s shrinking so much compared to other layers. The layers below the bridging layers only takes 190s to print. That’s more than 10 times of time difference alone. Then, it’s also blowing fan speed to 100%. The normal petg printing fan speed is 20%. So youre getting a huge ton of cooling from both high fan speed, and extra layer time, and extra layer time * extra fan speed. This alone causes massive shrinkage difference and can pull the corners upwards.
Now, another issue is that there’re too many straight corners.
FDM doesn’t like straight 90 degree corners. On the XY plane, if you can, always use a fillet.
Straight sharp corner = huge pulling force from shrinkage
Fillet round corner = much less pulling force
A huge pulling force can 100% lift the straight corners, or cause deformations, even if it doesn’t seem like it. Sometimes it can even break your model or printer. So avoid it at all cost.
Remember that you’re printing PETG. PETG is a high shrinkage material. And never underestimate the power of shrinkage, it can lift your corners, it can separate your model’s layers (yes it happens, literally breaking your model due to shrinkage alone while printing), it can lift up the build plate, it can break glass plates, it can destroy early printer beds…
Always take shrinkage into consideration when doing designs unless you’re exclusively using carbon fibre or glass fibre filaments. They have exceptionally low shrinkage rates.
Things that might help with shrinkage:
At the cost of overhang/bridging quality:
- raise chamber temp
- raise bed temp
- reduce cooling
At the cost of money and details allowed:
- Carbon / Glass fibre filaments
At the cost of part strength and printing time:
- Print slowly like old 3d printers.
At the cost of extra design:
- Fillets for all corners in XY plane, avoid all sharp corners. My personal guideline is at least 4mm radius of fillets if possible.
- Print in different angles with supports to prevent large flat surface issue.
E.g., a slight rotation spreads out the bottom surface into multiple layers. Each layer now has a much more uniform layer time and it won’t cause as much issues. Bottom surface layers are now using 350-600s each layer to print and that’s much better than from 190 to 2800
However I do wanna say that in the special case of your model, another way to reduce the issue is to just print a thick bottom surface.
This is your original settings’ layer time view:
The main culprit here is that bright green bridging layer. It’s taking 2800s compared to previous layers’ 180s. As long as you can smooth this out, you would get a much better result.
So a slow layer above a fast layer is big no no. But a fast layer above a slow layer is fine.
Using this logic,
Instead of using default settings, for this one particular model, if you print the entire bottom at 100% infill, by using modifiers, you can set it like this:
Then you would get a layer time view like this:
The rule of thumb is:
Slower layer on top of faster layer = bad results. E.g., 2800s on top of 180s = bad
Fast layer on top of slower layer = usually not very visible E.g., 180s on top of 2800s = nothing serious
Okay now I really need to go to sleep.
Chamfer 15mm has never given such a result anywhere else.
