Bill, right, the 170 is 220 volts; both are built on the same platform.
You say that the power supply gets hot and the welding power drops off. I am surprised that the overheat sensor did not trip. Maybe what is getting hot is not in contact with the overheat sensor. Someone mentioned that some transformers are wound with aluminum wire, that could be a problem in it self.
I take it that the fan is in good working order and that the air vents are all wide open, and air is not leaking from gaps in the cover?
When welding long seams the ground cable became uncomfortably hot. The fix was to replace the # 6 cable with a # 4 gauge welding cable.
Then I noted that the arc and bead improved, the energy lost in heating the cable is now going into the arc. Poking around under the covers I found other hot spots. The three wires coming from the transformer were connecting to the rectifier assembly by a loop around a bolt.
Now the current carrying capacity is related to the cross sectional area of the conductor, the ticker the wire the less it will heat for a given current. The wires from the transformer are adequate and are not the problem. Remember it is the area that the current flows through that matters. The surface area that the current flows through that is in contact with the wire, stud and nut is very small and represents a high resistance/ voltage drop/ energy loss.
I removed the wire loop from the stud, closed the loop back on it self, slipped a heavy copper terminal eyelet over the doubled wire and lightly crimped it in place, than I filled the void in the ferrule with solder. When assembled the transfer area the current passes through wire to solder to eyelet to stud is now larger= less loss.
Now this part is very, very tricky. I removed the rectifier assembly from the standoffs. You will notice that there are 4 spiders holding the button diodes in place on both sides of the heat sink plate. I very carefully marked the diodes with White-Out so that the orientation, which side of the button is in contact with the heat sink and which side is in the legs of the spider. It is imperative that the diodes be replaced in the same orientation, the side of the button that was in contact with the heat sink plate must be the same. If any of the diodes are reversed that would be a short circuit.
With the diodes removed from the heat sink, I polished the plate and the side of the button that contacts the heat sink plate.
Reassembling the diodes in the correct orientation I put a small drop of Heat Sink Thermal Grease on the side of the button that is in contact with the heat sink plate. The reason for the thermal grease is to improve the transfer of heat from the diode to the heat sink. The heat sink is hotter and the diodes run cooler. Reassembled everything being sure that all connections are tight and in the correct place.
The next mod may not fit your needs or may not be possible.
I added a second capacitor of the same values in parallel with the first, that increased the output current 50%. It also added to the heat stress.
Make notes and take photos to identify how you removed things and not trust your memory. If you have a meter, make note of the diode polarity, some meters have a special stop to test diodes. Make note before you disassemble the spider.