My DC Darling

[MrCurwen]

I want to ask for comments and improvement suggestions on my ongoing Darling project.

Schematic attached.

The voltage across the 1626 reads as 263V, with a current of about 36mA, so it's dissipating about 9.5 watts (specs say max 5W). No color on the plates, so I'm crossing my fingers and hoping they'll hold up.

6SN7's are running at about 9.7mA. I found the sound to improve much with more current - at first I ran them on 6mA.

I'm thinking the weakest link at this point is the Hammond OT. I'm looking into replacements. Also I think I should change the 1626 cathode bypass caps to 47µF Solen Fast caps, and add maybe another 4.7µF Solen cap to the PS (should help with transient reproduction, right?).

Both 1626's have a dedicated heater transformer, with center tap at 1626 cathode. The 6SN7's have common DC heater supply elevated with voltage divider from B+ to about +90V.

Duncan's PS designer says I have about 0.5mV ripple on the B+. No hum is audible even with headphones.

There is also a 1N4007 between cathode and grid of 1626 to prevent start up problems (forgot to put it in the schematic).

Any suggestions and comments on how to better my Darling would be much appreciated.

[DF96]

I would add a resistor (1M?) from the first triode grid to ground, so nasty things don't happen if ever the volume pot slider loses contact with the track. I would also add an input capacitor, so any LF which I can't hear won't cause intermodulation with sounds I can hear.

What is the purpose of the 14.3K resistor above the active load? Is it just to drop more voltage or is there a decoupler missed off the circuit? Does it have to 14.3k or would 15k be close enough?

I would not run an old (rare?) valve so far outside the stated limits. It might not go bang, but it will significantly shorten its life.

[kevinkr]

Originally Posted by DF96

I would add a resistor (1M?) from the first triode grid to ground, so nasty things don't happen if ever the volume pot slider loses contact with the track. I would also add an input capacitor, so any LF which I can't hear won't cause intermodulation with sounds I can hear.

What is the purpose of the 14.3K resistor above the active load? Is it just to drop more voltage or is there a decoupler missed off the circuit? Does it have to 14.3k or would 15k be close enough?

I would not run an old (rare?) valve so far outside the stated limits. It might not go bang, but it will significantly shorten its life.

Good suggestion on the 1M resistor..

In a conventional SRPP that resistor might help with PP balance, in this case it further raises the impedance of the active load, and drops about 130V which reduces the dissipation significantly and also assures that any version of the 6SN7 operates conservatively well within its published maximum plate voltage ratings.

The high operating current and dissipation probably will significantly shorten the life of the 1626, and one also has to wonder whether or not the OPT is running close to saturation.. (It's probably going to be replaced according to the OP - not a bad idea.) Personally I would not run the 1626 quite this hard, but I might run it somewhat beyond its ratings if no red plate is visible. The 1626 isn't terribly rare, but I imagine with all of the Darlings that have been built demand for this type is significant.

[MrCurwen]

Glad you mentioned it, I should add that 1M on the preamp grid.

About the 14.3k: it's actually four 82k / 1W resistors and one 47k / 1W resistor in parallel. I originally came up with the number after doing calculations - my goal was to get 150V on the preamp output. Of course the numbers didn't set exactly where I planned, but close enough. 15k would probably be just as good, though a bit more on the low side as per output voltage, but I find it useful to parallel resistors in this spot to get the power dissipation rating good (I've fried more than one resistor on the plate of 12AU7).

Without that 14.3k there the output voltage of the preamp would be around 222V (half of B+), and that would screw up the 1626 biasing. I don't have any more voltage headroom, so the preamp voltage must be dropped a bit.

Do you think it impairs the preamps performance? Would there be a more elegant way of doing this?

Hammond DSE is rated at 70mA DC. I've found that it sounds ok if you stay at less than 50% of that. There is quite a lot of hissing sybilants - it has become a serious annoyance, hence the search for a new OT. I'm leaning on James at the moment, they have a really good price.

What kind of distortion figures would one roughly expect from this kind of circuit? Any obvious misdoings?

[fdegrove]

Hi,

What kind of distortion figures would one roughly expect from this kind of circuit? Any obvious misdoings?

Unless I overlooked something the voltage dropping resistors in the B+ of the input stage aren't decoupled.
That creates an imbalance in the SRPP stage.

Cheers,

[DF96]

As output is taken from the lower anode, not the upper cathode, this is not an SRPP but a grounded cathode with an active load. The resistor will simply add to the anode load seen by the lower triode, so should aid linearity.

[hemgjord]

Why not try to put the 14,3 k resistor between the 6sn7's and cap couple to upper grid (mu-follower) should significantly increase the active load.

/Olof

[revintage]

grounded cathode with an active load. The resistor will simply add to the anode load seen by the lower triode, so should aid linearity.

You could as well run a resistor instead of the useless active load. The upper tube represents ca 12kohm and together with 14kohm you have ca 26kohm total. If wanting a active load go for a depletion fet CCS.