The all DHT SET Headphone Amp

[TheGimp]

Probably so. I didn't look at the PS. I was just changing the amplitude of the input.

[iko]

And if you can't find a HV shunt reg that can do 260V output here's what I use. It'll be very quiet and outstanding output impedance. You do need to know what you're doing to get a good pcb. More info about it you can find in this following old thread HV shunt regulator

[regal]

No the SSHV2 is good to 450V, when you DC couple it just makes since to use a tube for the first stage that operates linearly at a lower plate voltage, when you DC couple B+ is the addition of the first tube's plate, the second tubes plate, and the second tube's cathode. So 260V isn't the issue, over 450V is the issue.

I guess I wouldn't give up on DC coupling, especially since I showed how to filament bias the output tube. This gives the minimum caps in the signal path (first and second tube filament biased and no coupling cap.)

[andyjevans]

Originally Posted by regal

The signal is in the filament bias resistor loop, it doesn't dip to the 4.3K resistor which is just elevating the ground, just like biasing a DC heater with an indirect heater tube.
So the big cap is not in the signal path, just there as noise filter. This design is filament bias (no cathode bypass cap) with an elevated filament. It is means to have DC coupling without a cathode bypass cap.

No - I still don't get it. Think of the filament+filament resistor 32 ohms as simply the filament supply. The signal goes to ground through the 4.3K resistor. It can't go to ground through the filament supply because that's floating. So the cathode bypass does have the same effect as it usually does. I can't see there's any free lunch here.

[regal]

Originally Posted by andyjevans

No - I still don't get it. Think of the filament+filament resistor 32 ohms as simply the filament supply. The signal goes to ground through the 4.3K resistor. It can't go to ground through the filament supply because that's floating. So the cathode bypass does have the same effect as it usually does. I can't see there's any free lunch here.

OK then leave off the cap that is parallel to the 4.3k resistor, the tube is biased, there is little to no degeneration since the "battery" bias exponentially greater than the tube's bias. I just put the cap there for noise reduction, I think it could be .1uf. See what I am getting at? The signal can't can't go below the "battery".

[iko]

For DC coupling why not stack two SSHV?

[regal]

I guess I don't see the need, the input tube will would not be powered by the SSHV because it would have a CCS or gyrator. If we use a negative supply for the input tube to enable a dc couple then we need to add a coupling cap at the input, so no benefit.

I think what I posted a few pages back makes the most sense and has no practically no caps in the signal path. I'll add detail to it and attach the asc file this weekend.

[andyjevans]

Originally Posted by regal

OK then leave off the cap that is parallel to the 4.3k resistor, the tube is biased, there is little to no degeneration since the "battery" bias exponentially greater than the tube's bias. I just put the cap there for noise reduction, I think it could be .1uf. See what I am getting at? The signal can't can't go below the "battery".

I still don't get it - the filament supply + 32 ohm resistor constitutes just a kind of "augmented" filament supply. It has no effect on the the signal circuit until it's grounded, which it is through the 4.3k resistor. The 32 ohm resistor is fairly insignificant in proportion to the 4.3K resistor. So effectively this looks to me no different from a floating filament supply grounded through a 4.3K resistor, which would then usually be cathode bypassed.

Can anyone else cast some light on this?

[iko]

You can't get rid of the cap across the 4k3 resistor if you want the tube to properly amplify the signal. If you want to get rid of that cap you need a stacked supply. Then the 4k3 resistor also disappears. Which also makes the filament bias useless. So, if you want DC coupling then you either to the resistor plus cap, and the filament bias is useless IMHO, or you have a stacked supply, and the filament bias again brings nothing to the table.

Myself, not sure how much of a fan of DC coupling I am... is the effort and complications worth it?

[45]

Originally Posted by regal

If we use a negative supply for the input tube to enable a dc couple then we need to add a coupling cap at the input, so no benefit.

I wouldn't be so sure about this statement. In real world things tend to be not so simplistic. Firstly the input cap can be as good as it gets, secondly it would be a small size cap and last, but not least, it works at rather modest signal level.

45