Hello Xav,
euhellmachin a écrit:
I have a little question concerning output transformer impedance,
I first investigated the "best" (giving most Po) impedance ratio empirically w/o gNFB (means, open loop) using some transformer and changing resistors at the secondary, while computing the actual reflected impedance along. It turned out an impedance ratio of about 220 would be suitable for 4-8 ohms loads. Since impedance ratio is voltage (or winding) ratio squared, taking the root of the empirically found impedance ratio will give the voltage ratio. Since a mains/power toroid was asked for, dividing mains voltage by the computed voltage ratio will give the needed secondary voltage of the PT-toroid to look for, which was about 15V secondary.
Sadly, I don´t have this first load matching table anymore, but see below ...
Then, knowing the needed voltage ratio, I bought a 230V:15V toroid (actually 115V+115V : 15V+15V, primaries in series, secondaries in parallel) and double checked with and without gNFB applied the same way, making a second load matching table (the first one is quite similar, except the first used xfrmr had a different voltage ratio):
I think this is quite self-explanatory
Citation:
and heating supply, do we need separate heating supplies, or polrised one, or floating?
You inevitable need separate heater supplies for upper and lower tubes, or you would exceed permitted heater-cathode ratings (and get noise and all).
Usually I power P-type 300mA series heaters using a DC CCS, but I made some preliminary experiments, finding out what the could heaters initial current surge of two PCL82 in series is, and it turned out that it only was about 0,5A, for very few seconds. So I decided just to use two 32VAC secondaries for upper and lower tubes, heaters of upper and lower tubes connected in series.
Futher experiments, using a simple, bypassed voltage divider to "lift" the secondaires showed that there was not the least hum induced, but upper tubes (remember - full swing at the cathodes - swinging around the lifted, fixed heater potential!) sometimes gave a slight crackling noise. Maybe this was a tube with bad h-k isolation, but I didn´t really check. Instead, since some room still was left inside the chassis, I took the opportunity to try out a circuit from the Morgan Jones book, called "THINGY", which essentially is using low-Z output emitter followers also acting as RC filters to elevate the heaters to iddle cathode potential + 40V (RCA recommendation):
Built, it looks like this:
It performs admirably, no crackling anymore. Highly recommended instead of the usual method of using just a bypassed voltage divider to elevate heaters! Computations for other elevation voltages are very simple - if you don´t have the MJ book at hand, have a look at this thread, which gives step-by-step instructions to compute parts values:
http://geek.scorpiorising.ca/GeeK_ZonE/index.php?topic=2695.msg35434#msg35434
Hope this answers your questions,
Tom
Sorry for my bad camera 
)
