When you drive an iron core hard the permability goes down. Stick to Ui of 6 to 10 or you will have problems. Don't snort burning teflon as it will kill you! 4X 4-125s will be a low Z output. You will not need much inductance anyway so I would stick to 1/4 inch copper tubing.
Thanks for your reply. The cores have a μ=8 so three of them should be ok at the relatively modest power output. (I expect to run 3000v @ 250ma, maybe 250W or so carrier output, which isn't a particularly low plate Z). I had thought about running the HV through the tank coil to eliminate the plate choke, but the dc bias might indeed lead to saturation of the cores, and also that opens up a new can of worms with regard to resonances.
I calculated the required tank inductance to be around 13 μH which takes 17 turns or so on the toroid stack - 1/4" tubing won't fit, and even #10 would be getting close on the inside of the turns. I suppose I could go back to a single-layer "air" coil but I was hoping to minimize the strong RF fields atop the chassis, as well as the size of the coil.
Since I don't plan to use the amp above 7.3 MHz, and will only be generating 250W, or 1KW PEP for AM (maybe a little more for SSB), is a conductor that large really necessary at 40m? Most of the designs I've seen have really heavy wire/strap for use on 10m or above.
What do you think about the link coupling, whether to a toroid or an air-wound tank coil?
thanks
Charles