A very simple single-ended parafeed design
post-13066248

#### Sodacose

Something wasn't sitting quite right with the above calcs and I realized that I wasn't taking into account the cathode resistor feedback with the load lines.

Working backwards from gain with Rk:

1Vrms is approximately 2.8V ptp.

2.8V ptp input times gain of 27 is 77V ptp.

77V ptp squared divided by (4200 ohms * 8) is 175mW.

That's where I got the 150mW (derated slightly) for 1Vrms input.

post-13067408

#### proid

Thanks Sodacose,
That explains a lot. And some other things i don't understand:
1. The Edcor WSM is 0.5W, does that mean it can only transfer 0.5W of power e.g given secondary tap drving 32ohm load then the transformer can only  pass as much as sqrt(0.5/32) = 125ma, is that right?
2. What do you think about quiescent operating point affect the sound quality beside power dissipation and power output? i.e increase/decrease plate voltage, anode voltage, bias current.

post-13068611

#### Sodacose

Hi Proid,

RE 1:  No DC current flows in the OPT because it is coupled through the capacitor.  Even in non-parafeed applications, the OPT still does not 'pass current' in a DC sense.  In transformers, voltage fluctuations in the primary are coupled to the secondary by the magnetic field created by the windings and conductive core.  There is no flow of current from the primary to the secondary; there are only AC voltages that are stepped up/down based on the relative turns ratios.  We must use impedance and AC concepts here rather than DC calculations.

The power rating in signal transformers is usually less a reflection of their maximum current than a reflection of their rated frequency response and distortion.  The inductive and capacitive aspects of transformers are frequency and impedance dependent.  To complicate matters, single-ended transformers do usually have a DC current rating, but this refers to the current in the primary only and reflects the gap that's needed to preserve inductance (bass) when the transformer is used single-ended and series-feed.  With parallel feed, the coupling cap blocks any DC current in the primary that might rob inductance through saturation.

Edcor does not get super detailed with their specifications on the WSM/XSM.  They do mention a 10Vrms (28V ptp) input on the primary of the WSM and a 50Vrms (142V ptp) input on the primary of the XSM.  Again, this is related to rated FR and distortion specifications and not a reflection of the strength of the insulation.  The Papa Rusa design exceeds the WSM primary voltage rating at full power but I didn't notice any detrimental effect on HF/LF or distortion (though listening at full power with regular sensitivity headphones is not a great idea for your hearing).  If that's a concern, go for the XSMs (barely more expensive).  The caveat (in theory) is that the larger a OPT gets, the more inter-winding capacitance there is to affect the HFs.

RE 2: The maximum anode voltage of the 6C45 is 150V according to the datasheet.  I've seen them run above this in more than one design.  They are very rugged little tubes.  YMMV.  Power dissipation is rated at 7.8W and with that I would be a little more careful.  At 150V on the anode, that equates to a maximum quiescent current of 50mA.  I like to design for 70-80% of maximum dissipation for longer tube life.  The other limiting factor here is the heatsink for the CCS though.  You need to dissipate the difference between the B+ and the anode voltage at the current through the tube.  If you'd like to run at higher currents, just be sure your heatsink will be able to keep up.  At 30mA, it's dissipating about 4W.

If keeping anode voltage constant, higher current will usually lower distortion but also decrease power (because the maximum swing of the grid is also decreased).  The opposite is true when lowering current (within reason).  Based just on plate curves, I don't think there is a ton of room for improvement with the operating point, but I did not experiment.  It sounded great at 30mA, so 30mA is where is stayed

post-13068614

#### Sodacose

Oops, doubled up response.

post-13183060