• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

6DK6 amp schematics?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

Just wondering if anyone has had any experience with these tubes? What is the story about them?

I am planning a OTL cathode follower as per the schematic below.

Was wondering if anything special would be required for plate and filament power supplies?

I wonder what the total output would be for this amp...

Many thanks...

An externally hosted image should be here but it was not working when we last tested it.


Joined 2003
Paid Member
GE's 6DK6 maximum screen voltage is 150V, here it gets 198.5V. (TungSol gives 330V, though only 225V at this operating point.) Working the plate right to its power rating too.

> I wonder what the total output would be for this amp...

Output stage standing current seems to be 140mA. The peak output for symmetrical waveforms can be no more than the standing current. The RMS current is 0.707 times the peak current. So 100mA or 0.1A, squared, times 8 ohms, is.... 0.08 watts.

Distortion is likely to be over 10% even at half-power (0.04 Watts); I get ~33% THD (pure 2nd harmonic) when taken to just below the 0.08 watt clipping point.

The output cap fooled me. milliFarads? That's 1,000 microfarads. Don't be as dumb as me and try 1uFd, the bass cutoff will be 20KHz.

> these tubes? What is the story about them?

Better-quality TV IF tube. Low 3rd harmonic and IM distortion (high 2nd harmonic), high stage gain in capacitance-limited tuned-circuits, modest gain reduction possible. Good choice for wideband 10MHz-50MHz IF amplifiers. Darn odd choice for a single-ended audio amplifier.

> anything special would be required for plate and filament power supplies?

Heater supply can be anything that has the power. It won't be fussy.

This overloaded cathode follower has very little intrinsic ripple rejection from B+. It isn't that bad because gross inefficiency works to our favor. It has maybe 26dB ripple rejection. Taking 1mV ripple at the speaker as a criteria, you want 20mV ripple at B+. An economic rectifier cap gives about 2% ripple or 4V. So you need 4V/20mV or 46dB additional ripple filtering. One monster R-C, or two large R-C, or maybe a large L-C. The L-C case suggests a power choke much larger than this output choke.

Output choke DCR is critical. If forced to use something "too low", you can add resistance, but a choke with too much DCR will starve the already-thin standing current.


Joined 2003
Paid Member
> I think the output tubes should be 6KD6

It seems to use 15 output tubes and eats 150mA including the driver. That says <10mA per tube. The mighty 6KD6 would be total over-kill for 10mA each.

At the noted 150mA and 10 ohm bias resistance, grid bias is 1.5V. The grid voltage to hold 6KD6 down to 10mA each at ~200V on G2 is around -30 or -40V. With 10 ohms under 15 6KD6 the plate current is on the order of 150mA each, 2.2 Amps for the array of 15.

It also says heaters are 6V at 5 Amps. Fifteen 6DK6 plus a 12AU7 is 16*0.3A= 4.5 Amps. 6KD6 eats a whopping 6.3V 2.85 Amps each, so 15 of those and the 12AU7 is just over 43 Amps, somewhat more than the 5A noted on the plan.

I agree that the 6KD6 is a more interesting transformerless speaker driver. But it does not seem to be what Tim Williams had in mind.
I'm afraid it may not be worth the trouble. 40 free tubes = costly headaches. And if I'm successful, it would result in an amp that has too little power to be of use to me optimized for signals >10 MHz.

I think I will build some thing with the 6080 tubes that I also came across.

Thanks for the input!
chillysalsa said:
Maybe as a headphone amp, but then into a 32-300 ohms there would be even less power

Quite the contrary, in fact.

OTLs will give more power into a higher impedance load, since they are invariably current limited. The higher impedance allows them to swing more voltage before running into current limitation, resulting in more power.
Whoa! You guys jumped all over this one. :D

I would've posted sooner but besides sleep, it seems the OP also seperately e-mailed me for comments. Great minds think alike, my reply basically mirrored your reply PRR. :)

6KD6 would be a skookum OTL tube, particularly 15 in parallel, :D but I have 30 x 6DK6, only 4 x 30KD6. ;)

Oh, and yeah, I saved keystrokes by writing "1mF" instead of "1000µF". :clown:

This will cost you a low voltage supply and about a dollar of parts. Gets rid of that electrolytic. I'd use something like 100K/0.47uF for the servo time constant.


  • pcf_6dk6_otl.gif
    7.9 KB · Views: 833
SY: you've suggested modifying it with the servo, and putting the coil at a B- potential. Why is that? What would be the value of B- ?

I am still considering trying this out... the 6dk6 tubes are still quite available and cheap. Seems like it may make a nice amp for headphones, 300 ohm Sennheisers in particular.


Joined 2003
Paid Member
> But, a 1000uF cap... right in the signal path???

Don't look so shocked. Probably everything you listen to has been through something like that, probably many-many times.

It would help to go 5,000uFd-10,000uFd, to reduce the parasitic losses and hash that Tim and SY are arguing about.

I wonder how much it matters, with less speaker-power than a 1960 pocket transistor radio. There will also be some masking from the big-sweet 2nd harmonic.

As Jason says, and unlike a low-Z transistor amp, power is current-limited (for any rational number of 6DK6s) and will rise with impedance. At 600 ohms you can pull 6 watts. This would in fact be a killer hi-Z headphone amp, and would not suck in 32Ω (about 300mW). If your phones are 300+ ohms, you could finance a film-cap. You could have two jacks: one with 47uFd film for 300+ ohms, and the other with 1,000uFd for lower-Z uses. If you only do hi-Z phones, 3 or 4 6DK6s would be enough.

That is, if you like the curvature and 2nd harmonic. There are many-many other, more linear, tubes that would work. 12AU7/6SN7 comes to mind: same heater power per bottle, both sides in parallel make about as much current as one 6DK6. I have seen this plan with a few 6DJ8s. But 12AU7 never quite went away, and costs real money now, 6DJ8 has become faddish and good ones cost a lot, while anybody with a stockpile of 6DK7 would probably be willing to sell cheap.

Heck, one 6L6 will almost match fifteen 6DK6, be a little more linear, you can use the cheap $15 brands that burn-up in 470V guitar work. 6080 can work with this general plan but the biasing will be VERY different, and gain actually lower than with the bucket of 6DK6.

SY's servo will need a negative supply of about 140mA*Rch, where Rch is the total choke resistance. Using the 10Ω choke, -1.5VDC. I wonder how you make 1.5V 140mA without a cap on the order of 5,000uFd. True, the cap distortion is buffered from the signal path by the choke impedance.

> Whoa! You guys jumped all over this one.

Sorry. If I'd known that "Tim" was a regular here, I would have left it for you.
chillysalsa said:
SY: you've suggested modifying it with the servo, and putting the coil at a B- potential. Why is that? What would be the value of B- ?

If the cathode is held at 0VDC, the far end of the coil needs to be negative by the idle current times the DCR of the conductor- Ohm's Law forces that. The servo just keeps everything nice and tidy and drift-free.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.