6SJ7 Pentode Operating Conditions

General discussion area for tube amps.

Moderators: pompeiisneaks, Colossal

User avatar
rp
Posts: 2528
Joined: Sun Apr 25, 2010 4:21 am
Location: Italy

Re: 6SJ7 Pentode Operating Conditions

Post by rp »

didit wrote: Fri Mar 10, 2017 8:20 pmWill be putting trimmers in for Ra & Rg2, and perhaps Rk
Sounds like a recipe for a straight jacket and padded cell. I thought I looked through all the Gibsons but missed the GA50, two 6J5s instead of a 6SN7, that's a crusty old circuit. So this was a Jim Hall amp? That's a cool fact.

I just figured out Vbb/Ebb is the supply voltage not the Va! Anyone wanna clue me in what bb means?

So, that helps, but I don't think the table from the data sheet is very useful, I looked at several 6SJ7 and dozens of pentode circuits and they are all over the damn place and none just 'drop-in' to the table anywhere, got to use the charts, I guess. Anyways, many of these 6SJ7 circuits have pretty much the same voltages and resistor values as I do so If I'm not at a proper operating point with regards to current they shouldn't be either. Beats me. I want to try Martin's operating points and see what I hear before I throw this whole pentode thing out the window and put it back stock. I hate being stirred from my paint-by-numbers torpor.
User avatar
didit
Posts: 976
Joined: Thu Sep 25, 2008 3:37 pm
Location: Canada

Re: 6SJ7 Pentode Operating Conditions

Post by didit »

6SN7 is simply a pair of 6J5 in single bottle. Simpler and cheaper, and also still in production. Design borrows too from GA50T.

Trimmers will not be permanent. Just way easier (at least for me) ranging values without mess of clipping or tack-soldering fixed value parts. Can be tweaked while playing, lending more hope hearing nuances of relative changes.

Vbb is base voltage. Perhaps more common usage for transistors and the like. I might have scribbled B+ when not striving for consistency with Martin's notation.

Best .. Ian
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: 6SJ7 Pentode Operating Conditions

Post by martin manning »

Old tube data sheets will use Vb or Eb for plate voltage, Vbb or Ebb for the supply voltage. The b comes from the days when tubes were battery powered and the sources were labeled a, b, and c. See: https://en.wikipedia.org/wiki/Battery_(vacuum_tube)

The double letter practice is still used for transistors, where Vc is collector voltage and the supply voltage is Vcc.
davent
Posts: 236
Joined: Fri Apr 07, 2006 1:45 am
Location: Southern ON

Re: 6SJ7 Pentode Operating Conditions

Post by davent »

Dave Jones, around 2005 put out a schematic for the Octal One which was based on the 5c1 but with the 6sj7 cathode biased. Colossal recently built one and i built one a number of years ago using a 6aq5 for the power tube.

Mine i played around with the 6sj7 values and settled on this.

Image

Here's Dave's thread for his Octal One.

http://www.sewatt.com/node/21612

Schematic for Octal One

http://www.sewatt.com/files/sewatt/JonesOctalOne.pdf

dave
brewdude
Posts: 651
Joined: Mon Mar 31, 2008 6:26 am
Location: Napa, CA

Re: 6SJ7 Pentode Operating Conditions

Post by brewdude »

My SE amp w/ 5693 has a 12ax7 gain stage and TB tone stack preceding the pentode, but I looked up the values I recorded in my notes:

B+ = 312v
Ra = 100k, Va = 206v
Rg2 = 1M, Vg2 = 40v, Cg2= .05uF
Rk = 820R, Vk = 1.1v, Ck = 10uF

I should also disclose that I used a choke input power supply.
User avatar
Colossal
Posts: 5038
Joined: Sat Oct 20, 2007 9:04 pm
Location: Moving through Kashmir

Re: 6SJ7 Pentode Operating Conditions

Post by Colossal »

davent wrote:Dave Jones, around 2005 put out a schematic for the Octal One which was based on the 5c1 but with the 6sj7 cathode biased. Colossal recently built one and i built one a number of years ago using a 6aq5 for the power tube.
For my Octal One, I started with the stock Jones values for the 6SJ7. The plate resistor was 220k and the screen 2M2. This gave significant compression and the screen voltage was about 27V. After giving it some time, however, I just found it to be too flat and a bit dull/dark and lacking presence. I added a bright cap, still no joy. So I drew up a load line and ended up with:

Vbb = 292VDC
Va = 135V
Ra = 47k
Rg2 = 220k (47n to ground)
Vg2 = 77V
Rg1 = 47k
Rk = 600R
Ck = 10uF
Vk = 2.6V

That livened things up a lot and it was much more balanced and chimey. I used a C-L-C input.
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: 6SJ7 Pentode Operating Conditions

Post by martin manning »

Very close to my 47k Ra operating point!
User avatar
Colossal
Posts: 5038
Joined: Sat Oct 20, 2007 9:04 pm
Location: Moving through Kashmir

Re: 6SJ7 Pentode Operating Conditions

Post by Colossal »

martin manning wrote: Sat Mar 11, 2017 2:16 am Very close to my 47k Ra operating point!
Very cool! :P
User avatar
rp
Posts: 2528
Joined: Sun Apr 25, 2010 4:21 am
Location: Italy

Re: 6SJ7 Pentode Operating Conditions

Post by rp »

Now I'm all excited again!
User avatar
rp
Posts: 2528
Joined: Sun Apr 25, 2010 4:21 am
Location: Italy

Re: 6SJ7 Pentode Operating Conditions

Post by rp »

Colossal wrote: Sat Mar 11, 2017 1:40 amAfter giving it some time, however, I just found it to be too flat and a bit dull/dark and lacking presence.

That livened things up a lot and it was much more balanced and chimey.
OK, My experience exactly. I had pretty much the Jones set up, sounded fine but a bit generic and cheap. This is more dynamic and bright and clear, less compression and more snap, much more sustain. locked into focus like happens with a nice SLR camera. Not 110% sure which is better, but this has more vim and life and that's always better. I used metal films, vintage '80s Holcos, as I didn't have the values in cc - must be that and not Martin's math :lol:

These are my values - Ebb 253V (got pulled down a lot) Ra 47K / 137V, Rg2 220K / 91.5V, Rk 820 2.6V

If I want to measure Ra & Rg2 current do I do it across the resistors or to ground? I know, ohm's law, but I want to see it.

So, same experience setting up my 5C8, tubes suck, they work fine even when they are off, even way off, hard to figure out to a dilettante. Reminds me of a high-end hifi mag long ago that had a cover showing two state of the art multi kilo dollar power amps, something like a solid state Krell beast the size of a steamer trunk versus some a little pure silver wound OT SE 2A3 amp like a Wavelength or Audio Note with the headline, "If one of these is right, the other must be wrong."

Again, beats me, glad people here are around to work it out for me. But can I ask, why does this sound better? Is it that my first set up was on far edge of the curve? Below the knee? Not centred, not enough swing, not enough current? Were all the others set up wrong? Leaving asides ears and tastes, and quirks like cold/starved conditions, doesn't everyone read the curves the same way - as Martin and Colossal just did? How about the original Fender grid leak set up, Fender had great engineers, did the grid leak bias change Ra and Rg2 conditons so much that just matching the voltages w/ cathode bias didn't work as well?

Lesson: people that know what they are doing build better amps then does that don't.
brewdude
Posts: 651
Joined: Mon Mar 31, 2008 6:26 am
Location: Napa, CA

Re: 6SJ7 Pentode Operating Conditions

Post by brewdude »

I can't offer answers to most of your questions, but I can describe my observations.

What I've found is that the lower screen voltages squish the load lines down together and higher grid voltages spread them apart.

I built an amp that employs a VVR to adjust the screen voltage of the 6SJ7. I have it biased such that the load line passes through the knee somewhere around the middle of the VVR pot's travel.

With the screen voltage low it is very compressed and gets fuzzy quickly with stronger signals which works good at lower volumes/gain levels (There is a 12ax7 gain stage that proceeds the pentode).

Also, there is a VVR that controls the entire amp's voltages. If the global VVR is turned down much the amp sounds better with the screen voltage set higher.

As the screen voltages are raised, the tone opens up and becomes more dynamic (much more to my taste). If you continue to raise the voltage of the screen the becomes much more glassy (less fuzzy) and cuts through the mix much better.

I find that by balancing the gain, screen voltage VVR and global VVR I can tune the amp just right for the
texture/volume/room/mix.

I must admit that I am not adjusting the plate or cathode conditions.

Edit: I started a new thread in fear of distracting from this one http://ampgarage.com/forum/viewtopic.php?f=6&t=30027
Last edited by brewdude on Sat Mar 11, 2017 5:18 pm, edited 1 time in total.
User avatar
jjman
Posts: 753
Joined: Sat Oct 06, 2007 2:33 pm
Location: Central NJ USA

Re: 6SJ7 Pentode Operating Conditions

Post by jjman »

With the original goal of allowing for better pedal response I wonder if addressing the load on the pedals, while preserving the grid-leak bias, should be considered? My Silvertone 1451 has a single stage grid-leak-biased pentode preamp into the power tube. And it presents very close to a 1meg load to the pedal/guitar with the key being the placement of the volume pot on the input. I measured the gain (the signal-increase kind) and it is ~100. Although having the volume after the preamp would tame preamp noise when turned down below “10,” mine has almost no preamp noise. Works fine with pedals and sounds great. I added the main power isolation tranny and converted to proper grounding for safety, since it was a widow maker when made.
You do not have the required permissions to view the files attached to this post.
If it says "Vintage" on it, -it isn't.
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: 6SJ7 Pentode Operating Conditions

Post by martin manning »

rp wrote: Sat Mar 11, 2017 2:49 pm...I didn't have the values in cc - must be that and not Martin's math :lol: ...These are my values - Ebb 253V (got pulled down a lot) Ra 47K / 137V, Rg2 220K / 91.5V, Rk 820 2.6V

...If I want to measure Ra & Rg2 current do I do it across the resistors or to ground? I know, ohm's law, but I want to see it.
I picked the 47k because it's a standard value, after that I calculated values and left it to you to find something close and would have chosen the same ones as you did.

You can probably measure voltage across the resistors or to ground to get current. Assuming your meter is 10M input impedance it shouldn't make much difference.
then, sounding a little frazzled, he wrote: Sat Mar 11, 2017 2:49 pm...But can I ask, why does this sound better? Is it that my first set up was on far edge of the curve? Below the knee? Not centred, not enough swing, not enough current? Were all the others set up wrong? Leaving asides ears and tastes, and quirks like cold/starved conditions, doesn't everyone read the curves the same way - as Martin and Colossal just did? How about the original Fender grid leak set up, Fender had great engineers, did the grid leak bias change Ra and Rg2 conditons so much that just matching the voltages w/ cathode bias didn't work as well?
My reading of the curves and calculations are shown below. Hoping to make it as clear as possible, I used nomenclature matching the GE chart. The math is nothing more than the simple but extremely powerful Ohm's law. As you can see I used the recommended plate current to get started, and arrived at the same Ec1 too. Note the load line is well below the Ec1 = 0 knee, and Blencowe recommends this for guitar. Gain is about 75, which is only a little higher than the typical 12AX7 using 100k Ra. The other examples cited seem to be using much higher resistances, resulting in much higher gain. The grid leak bias probably results in less grid bias (maybe ~1V), and that might work better with higher Ra.

This data sheet suggests for Class-A1 amplifier operation:

Eb (V) 100 250
Ec2 (V) 100 100
Ec1 (V) -3.0 -3.0
Ib (mA) 2.9 3.0
Ic2 (mA) 0.9 0.8

From this, assume Ec2 = 100V, and a DC operating point at Ib = 3 mA. For center bias on the load line, Eb will be Ebb/2 and Ra will be Eb/Ib.
1) For Ebb = 290V, center bias Ra will be 290/2/0.003 = 48.3k
2) Choose Ra = 47k. Load Line Io = Ebb/Ra = 6.2 mA
3) Choose Ec1 = -3.0 crossing LL at approximately Ebb/2, Eb = 150, Ib = 3.0 mA
4) Read Ic2 at Eb & Ec1 = 0.82 mA
5) Rc2 = (Ebb - Ec2)/Ic2 = 232k
6) Ik = !b + Ic2 = 3.82 mA
7) Rk = Ec1/Ik = 785Ω
You do not have the required permissions to view the files attached to this post.
Last edited by martin manning on Mon Mar 13, 2017 1:16 am, edited 2 times in total.
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: 6SJ7 Pentode Operating Conditions

Post by martin manning »

BTW, using the 100k Ra point I proposed above using the same Ebb would result in gain of around 100. Using the logic above and holding the op point from the data sheet with EB from 100 to 250 gives a range of gain from 53 to 133.
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: 6SJ7 Pentode Operating Conditions

Post by martin manning »

...
Post Reply