6J1 tube buffer circuit diagram

[02GF74]

I've got two different 6J1 tube buffers off ebay that I am looking to fit inside CD players and would like to have circuit diagram. Googling seems they are based on Musical Fidelity X10-D

The first board amplifies the signal so that it distorts - the workaround is a resistor divider to reduce the input (which is what the volume control on board 2 does) but I would like to redesign it so it has unity gain.

Secondly I may want to redo for a new layout so that the circuit fits inside the CD player (currently I would have to make two holes for the valves in the top cover).

Both boards have a 3x voltage multiplies proving +/- 35 V ; I may want to put that on a separate board.

Anyway here are the two boards: (not sure if they are identical)


board 1

board 2

 

[adason]

I have x10-d schematics somewhere. None of those boards look like it. Have you considered the amount of heat inside the cd player it generates?

 

[02GF74]

Thanks for reply. I have seen the MF X10-D schematics. It has the voltage multiplier to give +/- 32 V DC supply from 12 VAC, like these two boards but the valve circuit differs. The MF has two valves whereas the ones I posted use one pentode.

I have traced the circuit diagram (for the first board, the rectangular one with 4 yellow capacitors) and it is as below.

Valves are a bit before my time so forgive my silly questions, hopefully someone can help.

1. voltage on grid (pin1) is very close to negative supply (1.4 V difference), is that ok? From memory the CD player output is 2 V pp so that gives little head room before the negative rail is hit.

2. info from web, gain of circuit is
-gm (Ra || Rp)
where Rp = plate resistance, Ra is R3 = 4.7 k;
is that correct?

3. I have seen similar circuit which has a bypass capacitor across R4 - what effect does this have on the gain?

4. how do I modify circuit to reduce gain?


re: heat - I plan to cut two holes in the top panel and have the valves poking through, may need to put louvres above the transformer.

 

[Koonw]

Hi, I think NFB will be the answer to all your questions, this is about unity gain buffer. However the gain can be changed if you want, by changing R6 value higher means more gain, have a look first.

Edit: it can take up to 10Vp input before clipping.

 

[02GF74]

Thanks for that. I see you changed value of R7 from 1.5 k to 100k. That alone would reduce gain as the input voltage is reduced by R7 and R2 divider.

And to confirm, you added negative feed back using C4 and R6.

What was the reason for changing value of R7? If I leave it at 1.5K and add the two feed back components, what will happen?

How do I calculate gain for the circuit you posted (I'm after gain of 1.0 or as close as is possible with minimal component changes).

 

[DF96]

What is the purpose of this circuit? You need to be clear on that in order to design it.

It certainly is not a buffer, as it has far too high an output impedance. You say you don't want gain. You also imply that you don't want distortion. So what is it for?

 

[02GF74]

I want to pass output from cd player through a valve without distortion or gain to pre amp. Just like musical fidelity do on their cd players and dm dac.

 

[Koonw]

Thanks for that. I see you changed value of R7 from 1.5 k to 100k. That alone would reduce gain as the input voltage is reduced by R7 and R2 divider.

And to confirm, you added negative feed back using C4 and R6.

What was the reason for changing value of R7? If I leave it at 1.5K and add the two feed back components, what will happen?

How do I calculate gain for the circuit you posted (I'm after gain of 1.0 or as close as is possible with minimal component changes).

R6 is 139K for unity gain, for calculation you can have a look in other post #24 But for this one is a little different I'll calculate later. R7 is part of NFB ratio. You can't not leave it as 1.5k is too low, but you can keep it as grid stopper resistor.

 

[jazbo8]

I want to pass output from cd player through a valve without distortion or gain to pre amp. Just like musical fidelity do on their cd players and dm dac.

If that's the case, why don't you just wire the tubes up as cathode followers? Or just replace the whole circuit with two pieces of wire and leave the filaments on for their decorative effects... Warm glow of the tubes without gain or distortion!

 

[adason]

just follow lampizator site...

 

[SY]

just follow lampizator site...

That's a collection of some of the most incompetent designs I've ever seen.

 

[02GF74]

LOL. It be like early transistor radios that were advertised ad having 10 transistors, 5 of which were glued to the case as they werent actually used.

Ok, so maybe no distortion is not the right term, i want the valves to add their "warmth" to the signal, early dacs were quite harsh on ths ears.

 

[Koonw]

....
Ok, so maybe no distortion is not the right term, i want the valves to add their "warmth" to the signal, early dacs were quite harsh on ths ears.

Do you have mute transistors in DAC output? I recently removed one faulty one, output is clipped if > 1V. Try to remove one to see any difference, only a slight pop when switch on.

 

[Koonw]

Here are the unity gain calculation other gain also included in case you need it as line amp.
Resistance 470k, reactance of C1 and C2 are ignored
They've little effect on feedback @ the values shown.

Supposed that Aol is 7.21, to give 1.5V output, input level required is:
1.5/7.21=0.208 or 208mV at J (grid)
The input level is reduced while the output is held constant at 1.5V
Gain db V Input(1.5/X) X R6(K)
0 1.5 1 132.2
6 0.75 2 315.1
9.5 0.5 3 584.9
12 0.375 4
14 0.3 5
15.6 0.25 6 4M
For 0db(1X), voltage between Vin,Vj = 1.5-0.208=1.292,
Vout,Vj=1.5+0.208=1.708V
I=1.292/100=0.01292, R17=1.708/0.01292=132.2K
For 6db(2X), Vin-Vj=0.75-0.208=0.542, Vout-Vj=1.5+0.208=1.708V
I=0.542/100=0.00542, R6=1.708/0.00542=315.1K
For 9.5db(3X), Vin-Vj=0.5-0.208=0.292, Vout-Vj=1.5+0.208=1.708V
I=0.292/100=0.00292, R6=1.708/0.00292=584.9K
.....
For 15.6db(6X), Vin-Vj=0.25-0.208=0.042, Vout-Vj=1.5+0.208=1.708V
I=0.042/100=0.00042, R6=1.708/0.00042=4.067M

 

[DF96]

Ok, so maybe no distortion is not the right term, i want the valves to add their "warmth" to the signal, early dacs were quite harsh on ths ears.

OK, so you do want distortion but not too much. The best way to do this is to make a cathode follower using an inappropriate valve (such as a small VHF/UHF pentode) and run it from too low a supply rail voltage (30-50V should be OK). Hey presto! You have a "tube buffer".

 

[Koonw]

Ya pentode is pretty good as cathode out buffer, at least better than a triode as the gain desired can approach unity. Sonic wise some dislikes due too high 3rd or odd harmonics.

Here is a sch for cathode follower.

If you're looking for tube sound it should be harmonics rich with even and odd evenly tapered off (see TubeCad harmonic restoration articles), the distortions are purposely introduced as "overload" will create rich harmonics. I'm not sure if 6j1 fits into that, see for yourself. Maybe I'll plot other FFT designs for comparison.

 

[Koonw]

Do you see any difference? Attached cathode buffer.

 

[Koonw]

Here are FFT plots with noise floor clean up. Left is anode, right is cathode.

.tran 0 {simtime} {dlytime} {timestep}

.options plotwinsize=0
.options method=gear
.options numdgt=7
.param numcyc=40
.param dlycyc=1
.param FFT=16
.param simtime=numcyc/Freq+dlytime
.param dlytime=dlycyc/Freq
.param timestep=(simtime-dlytime)/2**FFT

.param Sine=1
.param Vsine=.1
.param Freq=1k
.param Burst=1
.param Burstrate=2

 

[adason]

this is what I found about chinese version of x10-d

 

[02GF74]

Thank you, i have that one. From my understanding it uses two triodes per channel, something like 2N3 and is copy of musical fidelity x10-d..... yeah, i notice youve said that.

The ones im investigating have the same voltage muliplier, not sure how heater supply is done but both heaters are in series across presumably rectified 12 vac supply, but one triode (pentode actually) per channel.

Been doing a bit of reading and it seems the valve works better at higher voltage, would replacing diodes in the muliplier by schottky to gain approx 2v be worthwhile?

Also what is 7.0v to heater too high for 6j1?