Bogen DB 20 - Whats going on with that 12AT7 filament

[LooseChange]

Check out the schematic. To the right of the rectifier tube 5U4. Can someone explain the 12AT7 filament circuit ?  I think it's some sort of dc circuit but looks odd.
Thanks!

[John]

They might be using the 12at7 as a diode.

[LooseChange]

Don't think so... I still think it's some kind of DC heater. I seems to be used to add a bit of fixed bias to the power tubes (-7v).
Here is another schematic drawn a bit different.

[RicharD]

The filament is just grabbing free DC for V1 to keep thangs quiet.  I'm more interested in the -7V on the grids in a cathode biased circuit, with or without the feedback winding.  Yet another cat skinned a little differently.

[sluckey]

Notice that the PT HT centertap is not connected to ground? CT connects thru a 560 ohm to ground. And there is a 100uF cap from the CT to ground also (the cap is drawn backwards in the schematic with the red box). Anywho, this was a common way to develop a negative voltage source back in the day. This simple negative source is used for output tube bias and it is used to power the preamp tube heater. Also notice the preamp tube is wired for 12v operation? The schematic with the red box shows -10vdc bias but the other schematic shows -7vdc bias. I bet the actual amp is closer to 10 volts than 7 volts, since that preamp tube is lit up with the same voltage. What do you actually measure?

So, the 12AT7 filament is not preforming some black magic. It's simply being lit by the negative dc voltage.

[sluckey]

I'm more interested in the -7V on the grids in a cathode biased circuit, with or without the feedback winding.  Yet another cat skinned a little differently.

It's just a mix of cathode bias and fixed bias. I don't know if there's some advantage, or if it's some hifi trick. My Magnatone M10A has a similar bias scheme.

[Willabe]

The schemo (2'nd) says PT is 70-0-70 and they list 435dcv after 5U4 rec. The CT is in between 2 (stacked) filter caps, like a doubler, but they have added the 470R in parallel with the bottom filter cap. Can that be right?      

[jjasilli]

So, the 12AT7 filament is not preforming some black magic. It's simply being lit by the negative dc voltage. But this filament also seems to form another path to ground for the PT CT, with its resistance in parallel to the 560R.  There's yet another ground reference, after the inductor, through the 20K resistor!  What happens to B+ voltages if the 12AT7 filament fails by: a) going open? b) shorting internally?

And what happens to the power tube grid bias voltage under such conditions? 

[sluckey]

The circuit is really a lot simpler than y'all are making it. Look at the attached pic. I eliminated all distractions but left the B+ rail and bias supply intact. Hopefully it will be clearer.

But this filament also seems to form another path to ground for the PT CT, with its resistance in parallel to the 560R.

Yes, that's true.

There's yet another ground reference, after the inductor, through the 20K resistor!

That 3K and 20K form a simple voltage divider to knock the screen voltage down to 325V. That's all it does. This type voltage divider is common in old hifi amps.

What happens to B+ voltages if the 12AT7 filament fails by: a) going open? b) shorting internally?

a) B+ will decrease slightly due to the increase in the negative supply.
b) B+ will increase by about 10V. The PT CT will now be connected directly to ground.

And what happens to the power tube grid bias voltage under such conditions?

Well, if the 12AT7 filament shorts, the bias voltage goes to zero. So? The output tubes are also cathode biased so they will not be damaged. If the 12AT7 filament opens the negative bias voltage will increase and the output tubes will produce less power. Either way the 12AT7 filament fails, you will lose the magnetic phono preamp.

A little background on this bias circuit... It was very common back in the days when 6L6s looked like coke bottles. You've seen the more modern trick of putting a zener in the CT to reduce B+. That trick works because the zener will have a certain voltage drop across it. That voltage will be negative in respect to the chassis ground. So now, instead of referencing the B+ to ground, you're referencing it to some negative voltage. The negative voltage will subtract from the B+.

Putting a resistor in the CT lead does the same thing that a zener does, except the negative voltage dropped across the resistor depends on the load current of the B+ rail. Higher load current means more negative voltage dropped by the resistor. Since this negative voltage is applied to the grids of the output tubes, you have a kind of AGC (automatic gain control) for the power tubes. If the tubes draw more current, the negative bias voltage increases and tends to cut back on the power tubes conduction. And vise versus.

Any of this make sense yet?

[jjasilli]

Yes, thanks, much more clear.  It often amazes me how those old amps have circuit designs which seem quirky by today's standards!  The 1st filter cap is a bit lifted from ground, but what the heck!

Come to think of it, that may be a nifty way to power relays @ either 5 or 15VDC

[HotBluePlates]

Well, if the 12AT7 filament shorts, the bias voltage goes to zero. So? The output tubes are also cathode biased so they will not be damaged. If the 12AT7 filament opens the negative bias voltage will increase and the output tubes will produce less power. Either way the 12AT7 filament fails, you will lose the magnetic phono preamp.

Notice also the output tube cathodes have a load winding in the OT. Given the distributed load, where the power output of the cathode is also part of the output of the amp, maybe pure cathode biasing would require too large and costly a bypass cap to effectively bypass the output to the cathode winding. There may also be considerations of how that capacitance would interact (resonate) with the cathode winding.

You could due simple fixed bias like McIntosh did, but then you don't get the free d.c. heat for the input 12AT7, plus you add a couple extra parts.

[sluckey]

Notice also the output tube cathodes have a load winding in the OT.

For some reason my mind has totally blanked out that winding until right now! I got so involved in trying to explain the negative supply that I just couldn't see it. 

Lotta stuff happening in that amp that you don't get to see in a lo-fi guitar amp.

[PRR]

It's just plain self cathode bias. Done as complicated as possible. (I think there were previous designs and it got complicated in simple steps.)

I don't see -7V at grids, I see -10V?

The 560 resistor is moot. 12AT7 12.6V 0.15A is 84 ohms; another 560 is a fine trim. If you figure this plus the 220 resistors there is 183 ohms totally resistive self-bias. Well, 73 of those ohms are common with small-tube plate currents, but that's maybe 6mA and obviously the 6L6es are sucking something like 150mA, so ferget the 6mA.

If phono heater opens the 6L6es will run cold, near 60mA/pair. It will play with reduced power (as was said).

The question is: needing ~~25V bias, why not take the second 12AT7 aboard the DC train? Fisher did that with a 4*6V6 amp. Very low hum. This would avoid the 220+50u cathode networks, which are not needed (but may have worked great in a previous design).

The amount of NFB to cathodes can't be huge (like McIntosh), or the 12AU7 driver would not be enough. Probably 6dB-10dB here.

Like that Fisher, this has variable current NFB from V388 to adjust damping.

Seven phono-EQ settings!