U47EF Oliver Archut version

[Roman Beilharz]

That depends on your power supply unit. It may be that after switching on, when the heater of the tube is not yet warm, the 160V is exceeded, because the tube does not yet draw any current. Which PSU are you using? Schematic?

I don't have a PSU schematic, it was pre-built; but I am basically using this mic schematic with reduced H+ voltage and an added relay for true cardioid operation:


I am using a Mic&Mod standard PSU, where I have stripped pin 4 for relay switching via H+, changed the grounding scheme a bit and exchanged the XLR sockets for real ones. Looks like this now

 

[rock soderstrom]

It is difficult to make a reliable statement with this information. The power supply looks purely passive, the secondary voltage of the power transformer is not known.

I would play it safe and use a 250V coupling capacitor.

 

[Drosselmeier]

I am basically using this mic schematic

is in necessary to use c1 capacitor?

 

[oscarhuang]

 

[Khron]

View attachment 112168

..... But why???

If anything, if you go with @Drosselmeier 's idea, you can even get rid of R4 actually.

 

[Spencerleehorton]

The tube socket is isolated from the body, but I get what your saying, if the socket had some kind of shock mount attachment it would be better.
When I change those bits I’ll try a little mod so the socket has a shock mount.

 

[rock soderstrom]

But why???

second RC-filter for additional ripple filtering. I don't know if this is necessary but a very clean polarization voltage is a good thing.

 

[Roman Beilharz]

Thank you all for your suggestions; it's great to have so much expertise in one place! As I really love the sound of my mic so far - apart from the impractical LF sensitivity (and noise), I think I will stick with the circuit for now as is (unless someone convinces me of striking advantages of the suggestions above) , replace the 1µf transformer capacitor to 0,33µF (maybe even 0,22µF) and see what I end up with in another measurement and standardized audio recording round.

Love the internet

 

[MagnetoSound]

second RC-filter for additional ripple filtering. I don't know if this is necessary but a very clean polarization voltage is a good thing.

I agree. If it’s there already, why not use it. It can’t hurt.

 

[Khron]

second RC-filter for additional ripple filtering. I don't know if this is necessary but a very clean polarization voltage is a good thing.

I agree. If it’s there already, why not use it. It can’t hurt.

Sure, but as i pointed out in the FET847 thread (where some "blind" design decisions led to a 1gig / 100nF RC filter for the capsule bias), that would lead to a pretty humongous "warm-up time", for the capsule bias to reach full voltage, and the mic thus getting to its full sensitivity. In the FET847, that's a full FIVE MINUTES.

So the 100meg/10nF filter is somehow insufficient, with a corner frequency of 0.16Hz, and an RC time constant leading to a ~5sec "warm-up time"?

 

[rock soderstrom]

Sure, but as i pointed out in the FET847 thread (where some "blind" design decisions led to a 1gig / 100nF RC filter for the capsule bias), that would lead to a pretty humongous "warm-up time", for the capsule bias to reach full voltage, and the mic thus getting to its full sensitivity. In the FET847, that's a full FIVE MINUTES.

So the 100meg/10nF filter is somehow insufficient, with a corner frequency of 0.16Hz, and an RC time constant leading to a ~5sec "warm-up time"?

As always, there is also a counter-position. 5 min is a bit much, no doubt...
How do you calculate that, by the way? Short example, link? I only vaguely remember tau.
How big would the "warm-up time" be here in this specific case?

What also makes me think, none of the old guys did it that way as far as I know. There must be a reason for that. Probably the advantages disappear in the background noise of the other participants, so it probably doesn't matter and it is good enough.

 

[MagnetoSound]

Sure, but as i pointed out in the FET847 thread (where some "blind" design decisions led to a 1gig / 100nF RC filter for the capsule bias), that would lead to a pretty humongous "warm-up time", for the capsule bias to reach full voltage, and the mic thus getting to its full sensitivity. In the FET847, that's a full FIVE MINUTES.

So the 100meg/10nF filter is somehow insufficient, with a corner frequency of 0.16Hz, and an RC time constant leading to a ~5sec "warm-up time"?

Goddamn, rip it out then!

The tube should really warm up for at least 30 minutes anyway, though.

 

[rock soderstrom]

BTW.
https://en.wikipedia.org/wiki/RC_time_constant

 

[MagnetoSound]

11 seconds altogether, with 1100M in series.

 

[Khron]

How do you calculate that, by the way? Short example, link? I only vaguely remember tau.

BTW.
https://en.wikipedia.org/wiki/RC_time_constant

"It is the time required to charge the capacitor, through the resistor, from an initial charge voltage of zero to approximately 63.2% of the value of an applied DC voltage"

Getting to 100% takes about FIVE times that long.

11 seconds altogether, with 1100M in series.

That depends on the capacitor values; first one in the schematic is 10nF; you probably assumed the second one is the same? And 11 seconds to what, exactly?

I don't know if this is necessary but a very clean polarization voltage is a good thing.

How "clean" is "clean enough" then?

 

[rock soderstrom]

How "clean" is "clean enough" then?

I hope you can tell me. In numbers, please.

 

[Khron]

Let's go with the pre-existing 100meg / 10nF. That gives a corner (-3dB) frequency of 0.16Hz. It's a 1st order filter, so 6dB/octave, or 20dB/decade.

So 1.6Hz (one decade above) will be down by ~23dB, 16Hz (one more decade) will be down by ~43dB, and 32Hz (one octave above that), ~49dB down.

And that's still a frequency BELOW 50Hz mains, nevermind the 100Hz of rectified mains (in most of the world).

 

[Khron]

... And that's even WITHOUT counting the 10k/1uF filter that's already on the B+...

 

[rock soderstrom]

... And that's even WITHOUT counting the 10k/1uF filter that's already on the B+...

...and the filtering in the PSU. I get your point. It would be interesting to have a final number of rest ripple at the capsule that would be desirable. But I probably couldn't measure it anyway, so good enough is enough. Thanks!

 

[Khron]

When cost isn't that critical, i don't mind "guilding the lily" as much as the next person, but on the other hand, i also prefer to keep things within reason. And/or to have good (or at least sane?) reasoning for whatever design choices i make

Throwing parts into a circuit willy-nilly, "just because", does not count as "good", nor "sane". Not in my book, anyway...