Willsenton R8 power supply, 6SN7

[StepheK]

I'm working on one of these amps for a friend and seeing what I can do to improve it. It has an "extra" 6SN7 that is being used as a voltage regulator of sorts. Attached is how this is wired in the amp I am working on, with the measured voltages, and am curious why they would add this tube rather than just using a resistor/cap to drop the voltage to the input/driver stage of the amp where they want it to be. Any insight would be appreciated, I've never seen this before.

 

[gadget73]

Usually with active regulation you get better noise rejection than just RC filtering can provide. Thats my best guess.

either that, or for some reason they wanted to add a delay to the supply to whatever this feeds.

 

[kyle]

Here's another R8 schematic showing a different arrangement for this regulator

 

[TriodeLuvr]

they wanted to add a delay to the supply

I think that's exactly right. According to the schematic, B+ ahead of the first dropping resistor is +445V. That voltage will be present everywhere until the tubes conduct and create a voltage drop across the various resistors. The 6SN7 solves that problem without adding a resistive divider or other shunt method that would consume additional B+ current.

You know, those 450V caps are being operated way too close to their ratings. I wonder if the voltages printed on the schematic are even accurate when the unit is operated on a 125VAC mains supply like I have at my house. The manufacturer really should have used 500V caps (or two 250V in series).

Jack

 

[gadget73]

its pretty late in the chain though, the only place it would delay anything is the voltage amp. Not sure why it would need to delay just at that spot but *shrug*

 

[kward]

And it doesn't help if you hot cycle the amp. Seems silly to me as the schematic stands. On the other hand if you add a zener in place of the fixed grid resistor it might be a better regulator. Not that you need one in that position.

What does the top of the amp look like? Meaning does the extra 6SN7 provide some sort of symmetry visually? That may have been the true reason rather than just adding a 6SN7 that does essentially nothing.

 

[TriodeLuvr]

its pretty late in the chain though, the only place it would delay anything is the voltage amp. Not sure why it would need to delay just at that spot but *shrug*

Because the grid of the second 6SL7 will otherwise see full B+ voltage until the tubes conduct. They could have resolved that with a NE-2 bulb instead of wasting a 6SN7, but it's Chinese.

Jack

 

[kward]

Found this:



And this

 

[FlaCharlie]

I'm curious about this too as I came across the schematic for different amp that uses a 6V6 in a similar manner. In both cases the voltage supply to the input tube is the only part affected.

How would either a 6SN7 or a 6V6 provide a delay? And how much of a delay would it provide?

In an amp that uses SS rectification I've always seen thermistors used to reduce inrush current. What is the tube doing that a thermistor (or multiple thermistors) wouldn't do?

 

[gadget73]

Basically its set up like a cathode follower. Near-full voltage to the plate, some amount of voltage to the grid, output through the cathode. In this case instead of all of the current flowing through a cathode resistor, its using the voltage amp stage but as far as DC current goes, a powered tube is really just a resistor.

Reason it works as a slow start, no current will flow until the 6SN7 (or whatever is being used as the pass tube) lights. You could just as easily have gotten this done with a diode tube, the only real advantage to the triode vs a diode as a simple delay is that you can control the output voltage by setting the grid voltage. A diode tube would just drop what it drops for a given current flow.

better designs use a reference voltage source. In the olden times that was a gas regulator tube, now you'd use a zener diode. They also tend to have feedback from output back to input so it actually regulates. In a power supply its called an error amplifier, but its still negative feedback like you'd use in an audio amplifier.

Because the grid of the second 6SL7 will otherwise see full B+ voltage until the tubes conduct. They could have resolved that with a NE-2 bulb instead of wasting a 6SN7, but it's Chinese.

Jack

true, didn't consider that. I like simple things like neons or clamp diodes. Less wasted heat too. Is that truly a problem though? I have a couple things with solid state power supplies and a direct coupled section. Would be easy enough to stuff a neon in there.

 

[TriodeLuvr]

Is that truly a problem though? I have a couple things with solid state power supplies and a direct coupled section. Would be easy enough to stuff a neon in there.

That's what I did in my last build with a SS doubler supply. It's probably more of an issue with 9-pin miniatures, but I still wouldn't want to see +450V on the grid of a 6SL7 every time the amplifier is powered on. The NE-2 in the schematic below ignites for a few seconds at turn-on, then fades out.



Jack

 

[FlaCharlie]

its pretty late in the chain though, the only place it would delay anything is the voltage amp. Not sure why it would need to delay just at that spot but *shrug*

That's what I did in my last build with a SS doubler supply. It's probably more of an issue with 9-pin miniatures, but I still wouldn't want to see +450V on the grid of a 6SL7 every time the amplifier is powered on. The NE-2 in the schematic below ignites for a few seconds at turn-on, then fades out.

Wouldn't all the tubes be exposed to this momentary high voltage too? Is it more important to protect the input tube from the high voltage than the others?

Designers of SS rectified vintage amps didn't seem to be concerned with such things and I'm not aware of high tube failure rates. I like to add thermistors but perhaps they address a different issue?

In my current breadboard project I'm using a 'cockeyed bridge' with a 6AX4GTB damper diode to provide a slow start but, of course, it requires heater voltage. Much cheaper than a 5AR4, though.

 

[TriodeLuvr]

Is it more important to protect the input tube from the high voltage than the others?

The second section of the 6SL7 is the only tube that will experience B+ on its grid.

Jack

 

[gadget73]

Wouldn't all the tubes be exposed to this momentary high voltage too?

the plates would be yes, but none of the other grids would be. Only the second grid ends up at full B+ before the first triode starts to draw current and pulls down the plate. In a cap-coupled setup the grid resistor will keep it at ground, its only a grid directly tied to a plate that would have this situation.

how much an issue, no idea. I do actually have something that sees this condition. Its basically a Dynaco Mark III front end with a 6AN8, pentode plate to triode grid. I could easily stick a neon from triode grid to cathode like Jack's schematic to force the grid to not get over 80-odd volts positive vs the cathode. Maybe next time I have those on the bench I'll do it. I did have one of the 6AN8 start to get noisy and I don't really know why. Its fine for H-K leakage, but it was giving me random popcorn noises. Swapping the tube to the other amp moved it, swapping them for another set shut it up.

 

[TriodeLuvr]

Unfortunately, tube manufacturers don't specify this parameter. It's possible to make an educated guess, based on amplification factor. The smaller the number, the larger the cathode-grid spacing is likely to be. IOW, a 12AU7 should be able to handle much more voltage than a 12AX7. I wouldn't expect any of this to be a problem with the Mark III due to the delay in the 5AR4.

Jack

 

[gadget73]

Its not a Mark III, its something with solid state rectification that uses the same 6AN8 driver setup as the Mark III.

 

[TriodeLuvr]

There are a couple notes about this at the Valve Wizard site. One suggested alternative to the NE-2 is to use a diode with its anode connected to the grid. This works on the principle that the grid will never be more positive than the cathode, except during startup. Thus, the diode won't conduct in normal operation. I'm not entirely comfortable with this though, because diodes exhibit capacitance that varies non-linearly with the applied reverse voltage. The NE-2, on the other hand, is completely inert after it extinguishes.

Jack

 

[FlaCharlie]

The second section of the 6SL7 is the only tube that will experience B+ on its grid.

the plates would be yes, but none of the other grids would be. Only the second grid ends up at full B+ before the first triode starts to draw current and pulls down the plate. In a cap-coupled setup the grid resistor will keep it at ground, its only a grid directly tied to a plate that would have this situation.

So the only reason they're doing this is because the second section of the 6SL7 input tube is configured as a cathode follower, which drives the grid of the 6SN7 driver / phase inverter.

What's the advantage of driving the 6SN7 with a cathode follower?

Why not just cap couple one section of a 6SL7 (or whatever tube) to the grid of the 6SN7? Or even direct couple one section, as the Eico HF-87 does, and eliminate the cathode follower?

 

[TriodeLuvr]

Why not just cap couple one section of a 6SL7 (or whatever tube) to the grid of the 6SN7? Or even direct couple one section, as the Eico HF-87 does, and eliminate the cathode follower?

One more time... Where is this amp made?

 

[gadget73]

The NE-2, on the other hand, is completely inert after it extinguishes.

I suppose you could argue its got a very miniscule amount of capacitence but it can't be much. A couple pf at best? probably less than the lead dress around the socket.

So the only reason they're doing this is because the second section of the 6SL7 input tube is configured as a cathode follower, which drives the grid of the 6SN7 driver / phase inverter.

not so much because its a cathode follower, but because its direct coupled. Would have the same problem if the output was from the plate of the second triode. The first triode plate ties straight to the second triode grid, and with no current flow you get the full whack of voltage at the grid. Cathode is at ground through the resistor, so you've got 400-odd volts between the grid and cathode. Would be the case even if it was a typical anode follower with a small K resistor.

What's the advantage of driving the 6SN7 with a cathode follower?

only reason I can see for it, this allows the inverter to be direct-coupled to the voltage amp stage at more reasonable voltage. You can do without it, but just to take a guess, the plate is probably 100v or so higher than the cathode. That would put the grid of the inverter higher, which forces a much longer "tail" resistor so the grid to cathode voltage is reasonable. That would force a low current situation at the inverter, and/or too-small plate resistors at the inverter which would limit the voltage swing, possibly to the point where it can't drive the output stage to full power.

now whether or not the operating points and all make any sense at all is unknown.

its honestly not a terrible thing if its executed properly, makes it so the whole amp only has two coupling caps, so one possible phase shift point. Or if they did it wrong, its just one more place to screw the signal up.