I am attempting to reduce the hash over the filaments. For the TSL-100 there is a DC filament recto which runs off the AC heater tap, supplying DC to only TWO of the 8 valves in that amp. Conversing with a friend of mine, he told me as long as the noisier tubes (high gain circuit) are treated there is not much of a difference to treat the remainder. Be that as it may, there might have been cost analysis involved for the manufacturer to leave the others on AC. If this be the case, very well, but the plain simple fact is this is a 100W amp, and my speakers are high-efficiency types, which doesn't help, not to mention this house has no ground. So, I measured the BR62 (6A 200V) bridge-rectifier off the PCB. 15mm by 15mm. It's crammed in there. Not only so, it gets pretty hot, and the traces are burnt believe it or not. I had to beef up the traces with wire last major overhaul and also ended up drilling holes in the tube mount plate in effort to circulate some air across it. I have decided to replace the factory BR62 with, and hoping it will fit, a BR104 (10A 400V). I will raise it above the PCB so that the traces remain intact. The aim is to reduce thermal heat and improve the order of harmonic noise associated with the B-R. The best one the parts supplier had performs as expected up to 75 deg. C. then declines in performance. I could not locate a data sheet on the factory component, but chances are it's chocking the current and exhibiting thermal run-away. When the amp has ran for a while the hash becomes a bit noticeable. So if there is any improvement between the upgrade and the old, it should be heard first. I am replacing the (4) 3300u electrolytics with 4700u Nichicons. Again, space is tight back there so keeping the footprint small is necessary. If any improvement occurs, I might experiment with reassigning the filaments off V3 to the DC rail. This will be to verify whether the reverb tank/FX loop might provide less hash when engaged. The Axiom MAR-100 PT is a beef cake, not worried about consumption. One thing I've noticed in experimenting with the Blackheart BH15H, removing the DC filament BR altogether and going AC, is that the thermal heat is no more in that spot. Also, the filaments are glowing brighter, meaning that they are receiving the necessary amperage to operate properly, which translates to cathodes operating at proper temps, which translates to HIGHER HEADROOM and a fuller audio frequency pass-through. Aside from the 60 cycle hum (old house), the amp thumps thunder on the transient lows, tames the spikey highs, oozes with pleasant mids, and just SINGS the way a tube amp should. The new current draw (3A per tube?) doesn't seem to upset the PT any. At any rate, my experimenting is not finished with this amp. I am going to install a 25A 1kV B-R, 20000uF Rubycon, and .1u disks across two 100R, to see if that clamps down on the hash AND provide enough current to keep 'em tubes happy. If anyone has experimented with their amps and found a viable solution, I'd love to hear what you did!