• 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.

6SN7 Linestage with 6EM7 Regulation

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm using LTSpice and PSUD2 to model a 6SN7 linestage preamplifier inspired by George Wright's WLA12A line preamp. The WLA12A is the only tube preamp I've ever heard, so I'm trying to match experience with book learning here to figure out how such a circuit might work and possibly to build it.

Project goals:
* No ICs
* 6X5 rectifier
* 6EM7 series pass high voltage regulator
* 2 x 6SN7s set up as voltage amplifier and cathode follower
* B+ supply of 267V (measured by an owner)

I read up on tube series pass regulators, almost all of which used pentodes as examples. My only experience thus far is with triodes, so I found this 6F8G preamp with a 6EM7 regulator: http://www.diyaudio.com/forums/tubes-valves/71300-photo-gallery-179.html#post1334490/

After looking around for examples of 6SN7 voltage amp/cathode follower line stages, the Vacuum Tube Valley Octal Line Preamp project from VTV magazine (issue 11, Spring 1999) seemed close to descriptions of the Wright preamp. Another builder modified the power supply from an RCRC to a CLCRC filter (6SN7 VTV Line Stage), which was interesting.

In order to get a handle on all of this, I decided to mash together in LTSpice the 6EM7 regulator stage of the 6F8G preamp with the signal section of the VTV linestage, just to see what the simulation looked like (attached screenshot). Using a DC supply of 365V in front of the regulator stage -- low and behold, it happens to result in a B+ of 267V on the second 6SN7 plate.

With a 1V AC 1kHz input signal, a 50K resistor between input and first tube grid (i.e. at half the specified 100K volume pot), and with a load of 4.7K on the output (i.e. the input impedance of the power amplifier which follows the line preamp), we get a nicely shaped sine wave that swings between +15V to -15V and a current of +3.1mA to -3.1mA (attached screenshot).

I've also attached two power supply sims done in PSUD2, a version of the original VTV PS (RCRC) and the modified VTV PS (CLCRC), both with a 6X5 rectifier (attached screenshots). They result in about 361V and 8mA with a very tiny Diff measurement on the load resistor, which is set to 45K, which I guestimate to be roughly the load resistance presented by the 6EM7 when run at that voltage. The values can be tweaked to get the 365V, and I'll fiddle with that later.

So my questions are, Does this seem like a viable start? Does either of the power supplies seem better applicable to the project? Any suggestions on how/what to change?

Thanks,

Jeff
 

Attachments

  • gilberte-linestage-regulator-signal-beta-01.jpg
    gilberte-linestage-regulator-signal-beta-01.jpg
    90 KB · Views: 864
  • gilberte-linestage-beta-01-ltspice-graph.jpg
    gilberte-linestage-beta-01-ltspice-graph.jpg
    76.9 KB · Views: 854
  • gilberte-linestage-rcrc-beta-01.jpg
    gilberte-linestage-rcrc-beta-01.jpg
    133.3 KB · Views: 746
  • gilberte-linestage-clc-beta-01.jpg
    gilberte-linestage-clc-beta-01.jpg
    124.4 KB · Views: 812
Last edited:
So, I'm starting to think about operating points for the 6SN7 and am wondering what advice we have about that.

I simmed the Jan Veiset 6F8G circuit listed above, swapping in 6SN7s, and got results close to those listed on the schematic, which we can get closer to by manipulating the DC power voltage (attached screenshot).

The three circuits so far give the following sim measurements:

Jan Veiset 6F8G --> 6SN7 Conversion (6EM7 regulated)

B+ = 290V

6SN7 (1st half)
* Vp | Ip = 104V | 3.9mA
* Vk = 2.7V
* Vpk = 101.3V

6SN7 (2nd half)
* Vp | Ip = 290V | 6.1mA
* Vk = 105V
* Vpk = 185V


Wright-Inspired 6SN7 (6EM7 Regulated)

B+ = 267V

6SN7 (1st half)
* Vp | Ip = 100V | 5mA
* Vk = 2.7V
* Vpk = 97.3V

6SN7 (2nd half)
* Vp | Ip = 267V | 8mA
* Vk = 105V
* Vpk = 162V


Vacuum Tube Valley Octal Line Stage (6SN7, from VTV issue 11 Spring 1999, p. 10) (whole archive here)

B+ = 370V

6SN7 (1st half)
* Vp | Ip = 136 V | 5mA
* Vk = 3.7V
* Vpk = 132.3V

6SN7 (2nd half)
* Vp | Ip = 370V | 8mA
* Vk = 143.5V
* Vpk = 226.5V

I started looking at load line graphs in the 6SN7 datasheets I've collected, and I'm a little overwhelmed. The easiest one to read is the GE version, which covers the GTA and GTB variants. (attached screenshot)

So now my questions are:

To go about choosing operating points for the two halves, should I start by picking the most linear (i.e. evenly spaced graph intersections) across a line from the X to Y axes?

If so, does "Plate Voltage in Volts" on the bottom of the graph mean Vpk or simply Vp?

Does any of the sims above seem preferable, given your experience?
 

Attachments

  • gilberte-linestage-beta-03-ltspice.jpg
    gilberte-linestage-beta-03-ltspice.jpg
    100 KB · Views: 476
  • 6F8G-6EM7-schematic.gif
    6F8G-6EM7-schematic.gif
    28.1 KB · Views: 736
  • 6SN7-load-lines-ge.jpg
    6SN7-load-lines-ge.jpg
    249.8 KB · Views: 319
In the above LTSpice circuit is a mistake. R9 and R10 are connected together, but in the original circuit they are not.

To go about choosing operating points for the two halves, should I start by picking the most linear (i.e. evenly spaced graph intersections) across a line from the X to Y axes?

One simple (and maybe most effective) way to find the most linear operating point is to connect the distortion meter at the output and adjust the bias to minimum THD.
 
Last edited:
Good eyes! Thank you, artosalo. I fixed the model.

Since I started working with that schematic, I've been wondering where the lower outbound connection (diamond shape below the 39k resistor) in the power supply attaches in the signal section. I assumed it was a mistake in the schematic, but is there something more experienced eyes can see?
 
Several years ago I made a linear regulated based in the 6EM7, and it worked pretty fine, but if you want "all tube" like me, then I may suggest to replace the zener with a neon bulb (Without its resistance), it also works very well. Also, a version of the "see saw" amplifier can be used as regulator, with good (or better) results for me. Like this one post #107.
 
Thanks for the suggestion on replacing the Zener diode, Osvaldo. I will certainly look into that.

This morning I manually drew load lines using the formulae at this page as well as the values resulting in the LTSpice sim. They are rather different.

Following the instructions on the page linked above, using Ohm's law we calculate the plate current by dividing B+ by the value of the plate resistor.

Taking just the first half of the 6SN7 on the VTV circuit, that would be 370V/47k = 7.8mA.

However, the transient analysis graph in LTSpice gives the plate current as 5mA.

Which one do we think is more accurate?
 
Simulations are always this, simulations. True circuits have very different behavior, I checked it several times. In any case, an error of 50% is well in the reasonable point of view. Once I simulate the horizontal oscillator (Cathode Coupled Astable Multibrartor) with ECC82/12AU7, and give me a much more different results that in reality. Later I understand that models don't accept positive grid swings.

In respect to the neon lamp, take into account that most of them are around 60-80V and needs a higher voltage to strike, and remove any cap in parallel with them, they will oscillate (Relaxation oscillator made), and with a current of about 100µA is sufficient to keep them lighting.
 
Last edited:
Following the instructions on the page linked above, using Ohm's law we calculate the plate current by dividing B+ by the value of the plate resistor.
Taking just the first half of the 6SN7 on the VTV circuit, that would be 370V/47k = 7.8mA. However, the transient analysis graph in LTSpice gives the plate current as 5mA.

Which one do we think is more accurate?

If I understood right, the calculation you made (370V/47k) is for determining the left point of the loadline at the anode current scale. This calculation alone does not give any anode current value of the tube.

I have wondered what you are going to achieve with the LTSpice, because the only results you have shown are dc volts and milliamps.
With LTSpice I am used to analyze frequecy response, THD at different output voltages, etc.
Are these features of the amplifier meaningless to you ?
 
Osvaldo: That's a cool a way to get a pilot light, rather than simply adding one at the on/off switch.

Artosalo: Gah! Of course. This is all new to me, so the search for the B+ current vs. the actual anode current didn't register with me at the time I was figuring that stuff out.

Frequency response and THD are important to me, but for now I'm just trying to learn the basics of tube operation. Once I have the operating voltages and currents set up, then I'll sim for FR and THD.

Since last night, I've been reverse-engineering the Wright WLA12 from photographs and a resistor color code chart, and it does not conform to what I understand to be a voltage amplifier and cathode follower (which is what he called it on his website). On the 6SN7s, the cathode of the first triode is connected to the plate of the second by a resistor (pin 6 to pin 2), which looks more like the topology of an SRPP or mu follower.

The pictures I have are from internet sales, and some of the connections are obscured by components. If I can't figure it out over the next several days I might have to abandon this and just build one of the three schematics with which I started this thread.
 
OK, I've decided to build the original VTV Octal Line Stage and then modify it later with a regulator, etc.

Now I'm confronted with a basic question in hardware selection.

I want a rotary selector switch to choose between 4 input sources. There are all kinds of rotary switches with numbers of poles and decks, etc. What kind of switch do I need?



I'm confused by the
 
I'm looking at the 6EM7 regulator in a Wright WLA12, and there appears to be a possibly 200K resistor from pin 3 to 8 (cathode 2 to heater) (pic attached). What would be the purpose of this? To elevate the heater supply (6.3V) so that it operates within the 200V Vhk maximum?
 

Attachments

  • Screen Shot 2018-06-05 at 11.26.12 AM.png
    Screen Shot 2018-06-05 at 11.26.12 AM.png
    187.5 KB · Views: 192
Several years ago I made a linear regulated based in the 6EM7, and it worked pretty fine, but if you want "all tube" like me, then I may suggest to replace the zener with a neon bulb (Without its resistance), it also works very well. Also, a version of the "see saw" amplifier can be used as regulator, with good (or better) results for me. Like this one post #107.

Or, this might be a better option, replace simple resistor R9 with a ccs to bias Zener, should get much better performance.
 

Attachments

  • 53659C70-0283-4566-86B3-BAF96EB83295.png
    53659C70-0283-4566-86B3-BAF96EB83295.png
    35.5 KB · Views: 173
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.