There will be major improvements in the next release of Marlin, particularly in heater control. The difference is astounding, and ADC errors that have plagued Marlin for years have finally been fixed. No. Really. Even better is a completely new heater control paradigm called MPC (Model Prediction Control), which makes PID obsolete, IMHO.
At least on my Predator, the factory part cooling blowers are 24 volt, 0.2 amp Boosfan JD4010M, and they truly suck. They barely move air.
I replaced them with Delta BFB0412HHA-A units, wired in series to just one FAN0 connector. These are 12 volt, 0.08 amp units, and are physically a drop in replacement. I use two in series electrically so that each drops 12 volts, and the pair can run off of the same 24 volt FAN0 connector with the same PWM.
Although it is not considered "best practice" to wire BLDC motors in series, these are locked rotor protected, so that unless one blower fails shorted phase to phase (very unlikely), their low current draw and high impedance will prevent catastrophic results.
Why do this? The original units would barely start at 75% PWM, and barely move air at 100%. The new units actually start moving at about 5%, run smoothly and reliably at 10%, and will lift the printer off the bench at 100%. Just kidding, but the airflow is astounding, and actually having real air speed control over a large range is a huge benefit. And no monkeying with PWM settings in Marlin, either.
Plus, the new Delta blowers have dual ball bearings, so they should last for years. I don't know what the old blowers had for bearings, nor do I care, since they barely worked anyway. I may dissect one to find out.
The Delta blowers are commonly available from many sources, starting at about $7 each. I got mine in 2 days from Mouser for $14 each.
I use the Bakak duct from here on Thingiverse, and highly recommend it. The factory original duct basically just blew on the nozzle with the anemic air flow from the factory blowers, and I got rid of it long ago.
EDIT: I opened one of the old Boosfan JD4010M. It doesn't really have ANY bearings, per se. Just a plastic shaft rotating directly in a hole in the plastic housing. It's a miracle that these things spin at all.
I was hoping that Marlin would leap to 2.1, but no. Anyway, the latest full release works well on the stock Predator, and you can even use the MarlinUI, rather than the ClassicUI. MarlinUI is a full color touch screen interface. Temperature control seems more stable, as it should, since they finally realized that the Trigorilla Pro uses 12 bit, rather than 10 bit ADC's (thanks to me). In my opinion, it's still not good enough for platinum RTD sensors, but I'm working on that.
Motion control (steppers) seems to have been tightened up a bit, although I have no proof of that, other than M48 reports more realistic and repeatably believable values.
If you do use MarlinUI in configuration.h, you will have to "de-comment" Mesh_Edit_GFX_Overlay in configuration.h to get it to compile. Under ClassicUI, it compiles fine, and seems to work well. Bug report submitted.
I have a large STL that will go right to the extents of the build plate.
Everything lays down perfect and sticks well for 85-90% from center out, but that last outer bit close to the edge never gets close enough to stick down.
I have leveled and the probe hits all the points, is there anything else I can do?
The printer is really set up well and dialed in, I really want to use it for this STL
Attached is a description of a very different approach to piezo probing. I have a prototype on my Predator, and it works....frighteningly well.
Just throwing it out there. I don't make them, sell them, or have parts for them. The basic idea is the take away.
Here's a video of the probe in action: