Build thread all octal (6SJ7+6SK7) 5F2A with 6L6

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
You could drop bias and let voltage get higher. Now bias is bit left. Onthis attached loadline that green is about 30Vpp grid drive signal between -15V...-45v and when anode current "hit the wall" on left there is 36mA current to release on right and it would increase anode voltage sweep perhaps 50V...100V
Been reading tha Amato book all morning. So if I get the plate current down to around 70mA with plate voltage at about 420V it should be perfect?
 

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
Been doing some tweaking. First a total of 430R for cathode resistor but still to hot biased.
Two 270R in series seemed ok. Total 540R (quite high for a 6L6).
6L6
Plate 443V
Screen 448V
Cathode 35.7V
Plate current 62.3 mA
Ia 36.9 mA
Plate dissipation 84.7%

Load line seems ok with these 48 y/o eyes?

When digging in to this I’m getting curious on the 6SK7 triode. How should one find the optimal quiescent operating point? The cathode resistor value seems quite high.
 

Attachments

  • Universal loadline calculator for vacuum tubes - Vacuum Tube Amplifiers - DIY.pdf
    142.3 KB · Views: 8
  • CDFA4B1B-88D6-4B9E-B54D-15BBAAC6054A.jpeg
    CDFA4B1B-88D6-4B9E-B54D-15BBAAC6054A.jpeg
    268.1 KB · Views: 7
Last edited:

2L man

Tele-Afflicted
Joined
Nov 23, 2020
Posts
1,785
Age
62
Location
Finland
Been reading tha Amato book all morning. So if I get the plate current down to around 70mA with plate voltage at about 420V it should be perfect?
For HiFi there could be perfect Operating (bias) Point which produce highest output power lowest distortion. Even for clean quitar sound there is more opinions :) SE power stage output eventually clips but clipping stay warmer when other half tip clip first and distortion increase based to 2nd harmonic. Then when power tube is still driven harder second half tip clips and 3rd harmonics begin to come. So OP does not need to be centered. Usually SE clipping is not hard and wave tip(s) gradually flatten first.

OP/Bias is easy to change. Initially i install next up or two up higher value power resistor which alone make bias cold and then tack solder 1/2 W or 1W resistors parallel with it just flipping HV switch Off so it is fast. Often I use oscilloscope and signal generator because I have them but I also test playing and I have loudspeaker connected so ear plugs are needed. Loudspeaker because its impedance effect amp more than resistor load.

Then when I find Bias what I like I change the power resistor to value which is next up and then only one small 1,8k and up 1/2W resistor is needed and it look neater than what there came during testing.

Bias can be used quite easy as sound tuning adjustment just installing a switch which connect and disconnect one or two other resistors both side of the "cleanest of preferred bias". Potentiometer and resistor series with initial resistor work but potentiometer must handle the current and still series current limiter resistor is needed. Three different biases using resistor should be enough and it is simpler?
 
Last edited:

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
Even for clean quitar sound there is more opinions :)
So please tell me🙂.

The load line thing is quite informative once you get to know it a little. And the online calculator makes it a lot easier. The hard part for a new build is to find the theoretical bias that one should work around? In the tube manuals there are a lot of good info but not on all the different tubes there are.

Good tip about the larger Cr and a smaller 1/2 W parallel Resistor. Looks much nicer than my serial 10W resistors.

I think I’ll play around a bit with the 6SK7 to get it a bit “hotter”. now it is as loud as it can be and to be able to have more character I have to add a bit of gain from the previous stage?
Did you see my load line? One thing I didn’t get with it was the “Out. Headroom”. How would one do there?

The recordings are made with maste 100% and gain 90%. Volume decreased on the guitar.

Some sound samples. LED light shining on the open amp. Quiet when LED is off and guitar is away from the amp or the open side Is covered.


 
Last edited:

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
Sounds nice!
Both channels?
I think so to. Wish it was a bit louder though… an old speaker but also tried an 8 ohm P12N which also sounded nice. Will try a 12” 10-20W Goodman alnico speaker tomorrow but guess it wouldn’t increase output volume.

There is only one channel in this baby. I’ll need to update the schematic. I’m wondering if ill should decrease the 4.7K cathode resistor? There are no data that I can find that gives an hint on the values of a triode 6SK7. The 12au7 seems close but since I don’t have a specified grid leak resistor for v2 but the volume pot the values in the rca data sheets are not really useful. Maybe a 3K.
 
Last edited:

2L man

Tele-Afflicted
Joined
Nov 23, 2020
Posts
1,785
Age
62
Location
Finland
I think I’ll play around a bit with the 6SK7 to get it a bit “hotter”. now it is as loud as it can be and to be able to have more character I have to add a bit of gain from the previous stage?
Did you see my load line? One thing I didn’t get with it was the “Out. Headroom”. How would one do there?
Now OP is to the right of center.

In practice when small power triode drive pentode the headroom (clean power) stop increase where loadline meet zero volt grid line, where pentode begin to pull grid current, which triode can't deliver much, and what it can deliver get mostly lost to grid block resistor. HiFi builders use an interstage transformer or MosFet drivers. We players can use more advanced power tubes which 0V grid line is higher on anode current scale.

You can point cursor where the loadline meet 0V grid line and input what anode voltage sweep is from OP/bias point. Green sweep to the right of OP comes same g1 voltage sweep but negative phase. If bias is at -20V it go to -40V. In practice grid signal can sweep 0V to -40V when bias is -20V. When green is same length both side of the OP mean low 2nd harmonic distortion.
 
Last edited:

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
Now OP is to the right of center.
Do you mean with the 6SK7 or the 6L6? Probably with the 6L6.
So as I said I have been reading a lot about load lines but one can’t learn it all in a couple of days.

So what’s your opinion regarding tweaking on this one?
I’ll include the one I got with the actual values (posted) above with V+ 444 and the one you did make during the planning of this amp project. I think I’ve got a nice sounding amp that is very clean sounding.
I’ve got the grid bias voltage at -40V and in your starting example you got -20🤯. I guess still more reading to do…
 

Attachments

  • Load line Ebb 400.PDF
    142.5 KB · Views: 6
  • Load line Ebb 440.pdf
    142.3 KB · Views: 5

2L man

Tele-Afflicted
Joined
Nov 23, 2020
Posts
1,785
Age
62
Location
Finland
I don't pay attention to grid voltage much. Bias current is important. Naturally Vgc (Voltage between grid and cathode) effect the Bias. Screen voltage has significant effect to Bias and how Vgc settle when cathode biasing. Tuning SE power stage sometimes there is possible to get bit more output power or more pleasing sound.

Yes, possibly pre amp limit your amp output power? Eventually an oscilloscope is needed. Multimeter AC voltage measures RMS quite well but quitar string and pickup and how string is picked is too high variable. I have read that connecting cellphone as signal source is done and possibly then tuning pre amp is possible using multimeter measures.

I would trust loadline tuning because it work and using ULLC it is fast. Basically use same voltage sweep between grid lines and then look anode current which turn voltage sweep on anode resistance. Same anode current sweep make higher voltage sweep over higher anode resistance and this where voltage amplification comes. But then output impedance increase and then next stage input impedance has effect. Lower anode resistance make output impedance lower but lower the gain too. That is all I can say of that in practice.

I have tuned only 12AX7 pre amp stages and first I input the data of famous amps to ULLC and understanding how they effect signal came. Next I made same changes to an amp and oscilloscope verified they work just like the resistive loadline but actual tubes can have significant effect because even same type tubes have differences.

Power amp does not work on loadline because loudspeaker impedance depends of frequency and it "spread" the loadline.
 
Last edited:

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
I don't pay attention to grid voltage much. Bias current is important. Naturally Vgc (Voltage between grid and cathode) effect the Bias. Screen voltage has significant effect to Bias and how Vgc settle when cathode biasing. Tuning SE power stage sometimes there is possible to get bit more output power or more pleasing sound.

Yes, possibly pre amp limit your amp output power? Eventually an oscilloscope is needed. Multimeter AC voltage measures RMS quite well but quitar string and pickup and how string is picked is too high variable. I have read that connecting cellphone as signal source is done and possibly then tuning pre amp is possible using multimeter measures.

I would trust loadline tuning because it work and using ULLC it is fast. Basically use same voltage sweep between grid lines and then look anode current which turn voltage sweep on anode resistance. Same anode current sweep make higher voltage sweep over higher anode resistance. But then output impedance increase and then next stage input impedance has effect. That is all I can say of that in practice.

I have tuned only 12AX7 pre amp stages and first I input the data of famous amps to ULLC and understanding how they effect signal came. Next I made same changes to an amp and oscilloscope verified they work just like the resistive loadline but actual tubes can have significant effect because even same type tubes have differences.

Power amp does not work on loadline because loudspeaker impedance depends of frequency and it "spread" the loadline.
I was just about to write in som new investigation results before leaving for lunch.
Played with Cr values of the 6SK7

Cr 2.3k gave Cv 8.4 and Pv 112V. Louder output volume but started to break up too soon.

Cr 3.2K gave Cv 10.5V and Pv 126V.
Like this the best if one thinks on before. Clean but possible to make it break up.

Cr 4.7K gave Cv 14.3V Pv 149V. Clean almost all the way up and a bit tower in dB. More “sterile”. Need to learn on how to bias them preamp tubes for a tube without previous schematics and load lines.

@2L man will read you reply.
 
Last edited:

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
I don't pay attention to grid voltage much. Bias current is important. Naturally Vgc (Voltage between grid and cathode) effect the Bias. Screen voltage has significant effect to Bias and how Vgc settle when cathode biasing. Tuning SE power stage sometimes there is possible to get bit more output power or more pleasing sound.

Yes, possibly pre amp limit your amp output power? Eventually an oscilloscope is needed. Multimeter AC voltage measures RMS quite well but quitar string and pickup and how string is picked is too high variable. I have read that connecting cellphone as signal source is done and possibly then tuning pre amp is possible using multimeter measures.

I would trust loadline tuning because it work and using ULLC it is fast. Basically use same voltage sweep between grid lines and then look anode current which turn voltage sweep on anode resistance. Same anode current sweep make higher voltage sweep over higher anode resistance and this where voltage amplification comes. But then output impedance increase and then next stage input impedance has effect. Lower anode resistance make output impedance lower but lower the gain too. That is all I can say of that in practice.

I have tuned only 12AX7 pre amp stages and first I input the data of famous amps to ULLC and understanding how they effect signal came. Next I made same changes to an amp and oscilloscope verified they work just like the resistive loadline but actual tubes can have significant effect because even same type tubes have differences.

Power amp does not work on loadline because loudspeaker impedance depends of frequency and it "spread" the loadline.
Thanks for all input and help. One thing about the load line that I’m thinking of. When you did the first calculation with a +400v on the plate your bias was around -20V (I did include the pdf above) when I ran the actual numbers now with a plate voltage at 444V I got it at around-40V (also included). How can it differ so much? The cathode resistor value was chosen to give it a plate dissipation value <90% and OP in the “green”.

Went with a Cr of 2.9K since that’s what I got at home which was closest to 3.2K.

I do also thinking of adding a resistor in series with the choke between plate and screen node. Now the screen voltage is 446V and a bit higher than the plate voltage.

Edit. Might not need to lower the screen voltage. https://el34world.com/Forum/index.php?topic=12281.0
 
Last edited:

2L man

Tele-Afflicted
Joined
Nov 23, 2020
Posts
1,785
Age
62
Location
Finland
On the first loadline I attached the plate dissipation was few Watts above the 30W line but there I did use the voltages you posted. But bias was too hot and clean output was not as high as operative voltages would allow. Loadline did pass g1=0V below the knee which would increase Screen current or because grid block resistance would clip drive signal. If there is screen resistor after B+2 then increased screen current drop screen voltage and it lower grid effect to anode current and it would clip high power output signal tips. If Screen is fed direct from B+2 filter capacitor the screen voltage stay higher longer.

Screen voltage has significant effect how control grid effect anode current. Test changing screen voltage on ULLC and you see how grid lines height change. Screen current is quite low but increase rapidly when anode voltage approach zero and it happen when power tube drive signal is high enough. When bias is -20V only 40Vpp is needed but if bias is -40V then 80Vpp drive is needed for full output. It is nit exactly like that but this is how problem can be approached.

Driver tube and gain stages before might not be able to make 80Vpp signal and if they can it might be already distorted by 2nd harmonic because operation is not linear.

Possibly if you change EL34 power tube the amp immediately produce more output power because its needed drive signal is lower.

Using unconvential driver tubes is interesting and when successful very revarding but often they make proportionally much more noise to the amplification they make.

HiFi builders can use heavier OTs which can stand higher current without saturating because they do not road their amps much and they are also willing to spend more money to more expensive transformers so they can set steeper loadline where operation is more linear and loadline pass g1=0V above the knee and Screen current does not cause distortion. Sometimes they spend still mire money to Interstage transformer and drive power tubes to positive A2 or AB2 operation. Cheaper way is to "throw in sand" which mean silicon MosFet drivers :)

I have not finished a real A-class PP amp yet but they do have potential because OT does not need to be as huge as a SE OT. Also there comes more mains hum cancellation which does not happen on SE OTs. Obviously PP OT should be bit bigger for A-class than AB-class because bias current will heat it more.
 

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
If there is screen resistor after B+2 then increased screen current drop screen voltage and it lower grid effect to anode current and it would clip high power output signal tips. If Screen is fed direct from B+2 filter capacitor the screen voltage stay higher longer.
This I didn’t really get what you meant.

Screen is fed from B+2 node and I have a 4H, 167R Hammond 194A choke between B+1 and B+2 node and passes a 1K, 5W screen resistor before screen pin on socket.
I was thinking to reduce the screen voltage by adding a resistor in series with the choke to reduce the screen current and also B+3 which now is a bit to high also. When screen voltage is lowered to 350 the bias drops to -27 which seems beneficial. So according to what you wrote I should then lower the screen voltage and tweak a bit more to aim for bias at around -20V and to try to not let the Loadline pass g1=0V below the knee?

Will also add a bleeder
Using unconvential driver tubes is interesting and when successful very revarding but often they make proportionally much more noise to the amplification they make.
This is actually ok as far as I can tell. Should have gone the easy way of using an 6J5 but this triode wired 6SK7 should be ok if my reading on the tube specs is correct. Very low noise.

That hifi world is an interesting place. Got lost in those forums fast🤯.

Got hold of this cabinet for 20€ with these 15” monsters in it. Need to find out more about them.

Now time for morning coffe in my favourite muumin mug.

Edit: schematic a bit wrong. B+3 is 425! Need to lower that

BD458A1D-A545-4BD3-A8B8-D17CD357DDBA.jpeg
AB8F4AA6-B2A5-4899-A21E-D552E355BDE4.jpeg
763D3837-9118-4DC2-974D-36AEDF5F2F05.jpeg
 
Last edited:

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
So maybe done🤔.
Did some modifications and some helpful stuff (1K 4W pot and a 25K pot that I put leads on with alligator clips) to be able to tune the values without solder and desolder.
V1, 6SJ7
Plate 84V
Screen 36V
Cathode 1.5V

V2, 6SK7 (wired as triode)
Plate 92V
Cathode 7.1V

V3, 6L6
Plate 446V
Screen 380V
Cathode 29.5V

Calculated values
Plate current 59.5V
Bias -32.8V
Plate dissipation 90%
HD% 2nd:15.92 3rd:1.73 4th:1.17 THD:16.06

Still not really pleased with the output volume. Uncertain that the 6SK7 is able to drive the 6L6 to it full extent. Will try to disconnect the tone stack and the “driver” tomorrow and connect the 6SJ7 directly to the power tube like in a 5C1 and try different speakers and alter between this amp and a 5F2a and 5C1 to the same speaker to see if it is all in my head. And maybe use a 6J5 as a driver to see if it improves.

What is the characteristic of a good driver tube?

Should I lower the plate voltage to around 410-430V so it can handle a EL34? Having a 5K load.
I do also have some problems with the filament voltages. 6.8V. 1 Ohm resistors on each heater secondaries lowered way too much and 0.5R too little. Having a CT on the filaments so need to have a resistor on each “legs”. Need to order new resistors.

Also just finished a 5F2a out of old radio parts and an old Philips 8” alnico speaker and I putted it all in an old whine bottle wooden box. Rally need to learn how to do PTP.

When this is done the Ampeg B15-NC build will start. Should be nice to just follow a known schematic.


54FDC098-4872-4A01-972D-435468B19865.jpeg
907E0B7F-3204-4181-A2C9-C941EF88546A.jpeg
E03CA9A9-4C81-4BFE-B0A3-6D5B903DB515.jpeg
E34D39F3-E693-4178-81DE-6173F8A9D9D1.jpeg
 
Last edited:

mountainhick

Tele-Holic
Joined
May 2, 2021
Posts
527
Age
65
Location
Rocky Mountains
I do also have some problems with the filament voltages. 6.8V. 1 Ohm resistors on each heater secondaries lowered way too much and 0.5R too little. Having a CT on the filaments so need to have a resistor on each “legs”. Need to order new resistors.
Not most elegant, but I have used 1.X ohm resistors in parallel for this job. My DMM is not sensitive that range, so it is trial and error to get it right. Might be pretty tight in your layout?
 

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
Not most elegant, but I have used 1.X ohm resistors in parallel for this job. My DMM is not sensitive that range, so it is trial and error to get it right. Might be pretty tight in your layout?
I think I can squeeze it in. Next on my list.
My main problem has been lack of output volume and quite a dull sound. Been a bit obsessed with the power tube section and started with the wrong way of doing things. I did solder it all together from the beginning and that was not the best way. Everybody learns if one bangs the head on the wall enough times and really want to understand.
So now, since this is not the last amp I will build, I made a couple of different value and wattage pots with wires and alligator clips so I can do a bit of breadboarding in the chassis before everything is soldered in place.

I bypassed the cathode resistor of the 6L6 with a 220uf cap and the amp came alive. A bit to gainy but since I have a master volume and a parametric EQ I could get a good tone out of the amp. Thea layout must have been ok since no hum except a tiny one when everything is on 100%. Another fun stuff is that with this amp I can really hear a difference between grid leak and cathode bias mode for the 6SJ7. @printer2 switch really adds a lot in this amp. To make it function with a foot switch could really be a good thing in a gigging situation.
The bypass cap feels like the easy way and my gut feeling tells me that it’s a cheat.
I really think the 6SK7 can’t deliver the Vpp needed as it is running now. The problem is that there is no data on the tube connected as a triode. Some loadlines and the plate resistance would have been nice. I think I need to bias it hotter and increase the plate load resistor and the plate voltage so it can deliver the Vpp needed. More reading and trying stuff. Now I got to the point where I’m slowly learning and not just doing and that’s a route I’d like to continue. But the tone is what’s important…

There is a lot on the subject of a mismatched OT impedance and compensate with a higher plate voltage and the consequences it gets for driving the tube, THD and screen voltages that I don’t fully understand. @2L man explained a lot and helped me but I want to know as much as he does. Probably need to quit this medical doctor stuff to further explore this vacuum tube amp stuff.

Over all this is a keeper and hopefully even better soon. Have started on the head build. Might skip the Ampro cab for this one and use another 12” Goodman alnico speaker that is only rated for 15w and build a separate speaker cab for it so I can run all the other 1-10w amps in it if they need a “bigger” speaker🤔.

Already planning my Ampeg B15-NC and hopefully @andrewRneumann will build the 6N7 amp that I can copy🙂.
 
Last edited:

printer2

Poster Extraordinaire
Joined
May 24, 2010
Posts
8,781
Location
Canada




May find some useful information in these. I just did a quick enough peek to see there might be a nugget or two in them.
 

NTC

Tele-Afflicted
Joined
Mar 17, 2003
Posts
1,376
The bypass cap feels like the easy way and my gut feeling tells me that it’s a cheat.

Definitely not a cheat, especially if it gets you where you want to go. Most cathode biased output tubes use a cap to bypass the resistor. It increases the gain. Apparently an unbypassed cathide resistor has lower gain due to negative feedback. With the cap in place, the cathode is virtually grounded - max gain for a given plate load. With no cap, the signal appears at the cathode in phase with the signal. This reduces the effective grid to cathode signal, therefore gain is reduced.
 

Jerry garrcia

Tele-Holic
Joined
Nov 14, 2021
Posts
563
Age
48
Location
Sweden
Definitely not a cheat, especially if it gets you where you want to go. Most cathode biased output tubes use a cap to bypass the resistor. It increases the gain. Apparently an unbypassed cathide resistor has lower gain due to negative feedback. With the cap in place, the cathode is virtually grounded - max gain for a given plate load. With no cap, the signal appears at the cathode in phase with the signal. This reduces the effective grid to cathode signal, therefore gain is reduced.
Hi and thanks. I understand that but when I usually add a bypass cap on the Rc of the power tube I’ll get a small but noticeable increase in gain and output. This time there was like a different amp. The best way I can describe it is that it felt like the 6L6 finally could be driven to its full potential. Like when swapped to an “easier” driven tube, i.e. the EL34. That’s why I thought that it lacked sufficient Vpp in the previous stage. Makes sense?
 

2L man

Tele-Afflicted
Joined
Nov 23, 2020
Posts
1,785
Age
62
Location
Finland
Higher value cathode bypass capacitor increase gain because it decrease local feedback and cathode does not follow drive signal too much which would decrease actual drive which there is needed between cathode and control grid.

Without drawing loadline I am quessing that perhaps you have biased that pentode driver tube too hot when its anode voltage is less than 100V and it limit its gain? Sometimes tube datasheets have tables for several voltages and anode and cathode resistors and what the gain and output impedance then come.
 




Top