If you are just testing the board, solder in any diode on the board in place of the dead one and test away. The heat sink mounting is only needed on the amp after it is working and running hard.
A few years back, I tried a similar experiment. I bought a piece of aluminum tubing and cut it into 1/2 inch lengths. I then epoxied a few different diodes into the small pieces. The diodes that I tried this with were 1N914 and 1N4148, which both have small glass cases.
Then I clamped the test diodes and a real 1N3754 diode onto the surface of an old heatsink that I had laying around. I took meter readings of the diodes to get a starting point based on the room temperature at the time. Then I slowly heated the heatsink with a hair dryer and took readings from each of the diodes as the temperature increased to see how the three different parts reacted.
The two homemade ones reacted a little slower than the 1N3754, but they did react in a similar fashion. I wanted to try using different methods of gluing in the diodes, like different epoxies or silicons to see if that would make a difference, but I have not gotten back to this test yet.
You could also mount a two position terminal strip to the heat sink and then solder in a diode so that it's case is in contact with the heatsink. Then run two wires to the pc board to make the connection.