I've had a chance to go over datasheets with Engauge Digitizer for over a dozen MOSFETs. Here is a quick top 5 list of best picks if you wish to switch yours out. Keep in mind, @ 147.5 Hz, switching losses aren't that high so Rds(os) dominates. This means that the stock MOSFETs dissipate ~0.75 W on full duty cycle.

Stock - AOD484

• Vds: 30V
• Rds(on) @ 3.2V: 74 mOhm
• Gate Charge @ 4.5V: 8.4 nC
• Ciss: 938 pF
• Coss: 142 pF

Top Choice - AOD240 (~45 mW dissipation)

• Vds: 40V
• Rds(on) @ 3.2V: 4.4 mOhm
• Gate Charge @ 4.5V: 22 nC
• Ciss: 3510 pF
• Coss: 1070 pF

2nd Place - IPD031N06L3 (~57 mW dissipation)

• Vds: 60V
• Rds(on) @ 3.2V: 5.6 mOhm
• Gate Charge @ 4.5V: 59 nC
• Ciss: 10000 pF
• Coss: 1700 pF

3rd Place - AOD510 (~67 mW dissipation)

• Vds: 30V
• Rds(on) @ 3.2V: 6.6 mOhm
• Gate Charge @ 4.5V: 21 nC
• Ciss: 2719 pF
• Coss: 1204 pF

4th Place - IPD031N03L (~70 mW dissipation)

• Vds: 30V
• Rds(on) @ 3.2V: 6.9 mOhm
• Gate Charge @ 4.5V: 21 nC
• Ciss: 4000 pF
• Coss: 1400 pF

5th Place - IPD036N04L (~86 mW dissipation)

• Vds: 40V
• Rds(on) @ 3.2V: 8.4 mOhm
• Gate Charge @ 4.5V: 28 nC
• Ciss: 4700 pF
• Coss: 1000 pF

One of the photos for the project is of the 1N4448TAP diodes I'm using to protect the MOSFETs, stock or otherwise. Here is a capture I forgot I had of the stock 'FETs and the voltage spike that occurs during the operation for both the hoetend and heated bed. As soon as both of these turn on, you will have a spike for both line up with each other from time to time. Even without that, this spike equates to about a 9.7 a draw over a duration of ~180 ns. I bet this is why the stock PSU likes to fail...