The Ultimate M49

[ln76d]

Nothing personal, it's just that "absolute" declarations along the lines of "this is the (only) right way (because i say so)" kinda tickle me the wrong way ;D

No offense or trolling intended though, i promise

Cool

I really try to not write posts looking like ""absolute" declarations along the lines of "this is the (only) right way (because i say so)"" - rather want to share my experience, and if something just didn't work for me - then how i should describe that?

 

[mhelin]

If you want to try to use silicon diode drop for biasing the better device are the BA159 and UF4007 diodes (also Cree silicon-carbide was mentioned), they are the quietest according to a diyaudioprojects.com member who tested different diodes for U47 kind of mic amp - diodes (two BA159 in series for 1.1V AFAIK) were also 1.5 dB quieter than resistor for the same job. With diodes you could switch the bias colder or hotter, like using the SiC for the lowest (0.8V), UF4007 for 1.0V, BA157 for 1.3V and red LED for the coldest bias. Small resistor in series for further adjustments. 

 

[bockaudio]

FWIW the cathode bypass cap corner frequency is below 10Hz so that's not the primary reason for a sonic difference with fixed bias (not that you can't hear the different VLF differences).
Years ago I instigated current sources for tube cathodes and in tube mics, found that the power supply makes a more significant difference. I don't know that you will really be hot rodding your M49 with a cathode diode.

 

[rmburrow]

I like the "B" version where the tube cathode and high side of heater are tied together. No cathode bypass cap required. The grid bias is fixed by a series divider off the heater voltage. If the heater voltage (H+) for a 5840 (in triode) or 5703 is 6.3 v DC, the divider values could be 100k to H+, 300k to common (ground), the junction of the 100k and 300 k to the grid resistor (typically 150 M to 1G ohm). It is mandatory the heater supply be well filtered! The divider values could be adjusted slightly to get -1.6 volts or so at the grid (remember this is respect to H+!) but 100 k and 300 k are standard film resistor values.

 

[metalockpick]

Cathode cap boosts the gain, and the cap value sets the frequency at which the gain boost happens.

Going tiny ~0.1 uF means boost at higher frequencies, and much less bass.
Increasing the value will increase the gain at progressively lower frequencies, though diminishing returns after some point, especially if you're trying to force frequencies through the transformer that are too high or low for it to pass.

Tantalum typically has lower ESR and lower internal loss, and "should" be better. Higher voltage parts typically have lower ESR and loss as well, though you don't need to get silly here.
PSU has a lot to do with the performance here, in filtering, voltage regulation, grounding, as well as the circuit and the transformer used (the saying "It's in the iron" has some truth to it).