"Airflow must be applied before or simultaneously with the application of power, including the tube heater, and should normally be maintained for several minutes after power is removed for tube cool-down."
The above is good enough for me.
On one of my hb 2 x 4-400 amps, with bottom lid off...and rf deck pushed back on the desk, with just enough overhang, such that I could take some V measurements at the sockets, I put my finger on pin #1..... and just about took my finger prints off. It was fn hot...and then some. Those pins run stupid hot.
On my hb 3CX-3000A7 amp, initially I was using a SK-306 chimney, made for a YC-156 / 4CX-5000 tube...with their 4.94" OD anode cooler. The 3CX-3000A7 tube anode is only 4.125" OD, so used layers of 1/8" thick silicone rubber and a SS hose clamp, to plug the gap. We had the blower + heater on ( 7.5 vac @ exactly 50 amps, using a clamp on ac ammeter around one leg of the fil cable going to the next shelf below).
Ran it for like 10 mins, then shut down the fil and blower...and removed tube.... so I could tweak the silicone rubber down a bit on the anode, which required 1st loosening up the SS hose clamp.
I was sitting in my office chair, with tube vertical, and base of tube between my crotch, and a nut driver, tweaking the hose clamp etc. Tube base inner + outer collets were good and warm. But after 1 min, it was getting hotter and hotter + HOTTER. It was ready to burn my gonads, so lifted tube up...and gotta outa the chair asap. Then felt the inner / outer collets..and indeed they are hot..... but cooled off fairly soon.
And I had loads of air, with a pair of dayton 4C4006B blowers running, and also a magnehelic gauge installed, which read 1.15" pressure. 70 cfm airflow...and a helluva lot of pressure.
On all this hb stuff, with expensive tubes, when done for the night, amp in standby, B+ shut off, fil shut off.... but blower stays on for 5-10 mins. Depending on what u were doing beforehand, that anode can get stupid hot...with the requisite airflow.
Tubes cost too much, even to rebuild, so I leave the blower on for a while. I also use the small Variac to bring up the fil V slowly to 7.5 vac. I have a fil V meter...and also a fil ammeter. Fil current comes up from 0-5-10-15-20.....50 amps, over a 20 second period...by hand. Same deal when done for the night..and reducing fil V back to zero. That's my..'new way' of doing 'step start' on directly heated metal tubes. Cold resistance of these tubes is like only 1/10 of the hot resistance, and I was using a surplus fil xfmr, rated for 8.25 vac @ 79 amps. It has a 208 vac input, so the variac has to be used, to make it work. Just about all of the fil xmrs, ( except henry radio 8k ultra) are capable of one helluva lot of peak current, when u lay essentially a dead short on em...like > triple rated current. I'm 90% convinced that's where a lot of grid to fil shorts come from.
On the L4B 2 x 3-500Z amps, I just let the amp sit on standby for a few mins, then shut it down..which shuts everything off, no blower over run used, but that may well change on my next round of mods.... more of an experiment.