I was using a mystery TO3 transistor ( no markings on it, out of the used parts bin but was NPN, probably a 2N3055 or a 2N3773, but definitely not a BUX20, as those are not available in the standard TO3 package, but only in the massive high current package with heavy Kovarbase and leads) as a dummy load on a 5v rail, as I had to check it was capable of delivering 10A while still staying in spec.
Simple enough, transistor with 3 leads soldered to it, one heavy one to emitter that was bonded to chassis of the equipment, use any of the 84 M4x16 capscrews you had to remove to get the covers off, a thinner one to base leading to the decade box, and then collector was a thick wire soldered to the top hat of the device. In use take a nice clean cup, fill with cold water and drop the transistor in it, and measure current with the AVO8 on the 15A DC range, Dc voltage with the test rack DMM. As the test generally only would be 10 minutes or so, unless you had to adjust settings with another 2 decade boxes, as resistors were select on test and solder into the turret pins on the PCB, a lot more reliable than a variable resistor, but in general you only ran 10 minutes at a time. Water got hot, but generally not to the point of the transistor boiling it.
Same transistor survived, even being used to test power supplies, where you went well over 20A for a few seconds.
The BUX20 transistors were used in a lamp control, interesting in the power stage used a early optocoupler, and this drove a darlington stage made from a 2N2219A and the BUX20. Badly designed, the transistor was poorly heatsinked ( and not possible to change either without 1 ton of paperwork) so only thing was to select drivers for highest gain and thus lowest saturation voltage. Safety feature ( unintended) was the BUX20 unsoldering itself from the base and emitter wires when it got to 200C case temperature, otherwise lesser transistors tended to just go short circuit, blowing the 20A DC fuse, after blowing the main and spare lamp ( because pilot would switch in spare lamp of course when first went out, and this would blow the fuse on the now short circuit transistor) out. Interesting transistor BUX20, 50A collector current and usable gain at this as well, just really expensive, and only was beaten by power mosfets recently, just do not believe the power dissipation, hard to do without water cooling.
https://www.comsetsemi.com/Datasheet/bux20.pdf