6SK7 tube amp circuit

Hi guys,
Here is the schematic of my 6sk7 tube circuit, I am trying to nail the biasing value. So far the circuit works but just a bit noisy, I noticed that the noise is low on the CRO before I plug the output into the power amp.



Additional circuit info
- I have grounded the input and output cable together.
- the op amp to grid coupling cap has been poped twice when I used a 25V electro while tweaking the cathode resistor.
-the op amp driving the grid gets burning hot after I connect and disconnect the output to the power amp, and it dies after that.

Thoughts???

What are the power supplies for the op amps? +/- something I hope.

Why is the output tied to the grid of the tube?

If you disconnect that connection, the 100uF is still reverse-biased (unless the output is biased WAY positive) as shown---which would account for a s**tload of leakage and noise.

Why the huge value cathode resistors?

[quote author="bcarso"]What are the power supplies for the op amps? +/- something I hope.[/quote]

It's +-15V.

Why is the output tied to the grid of the tube?

Click to expand...

Without it, the output is incredibly "hairy" with big amount of crap embedded in the signal. By having that feedback it pretty much cleaned everything up, as soon as I increase the resistance of that FB it just start to get worse again...... :sad:

If you disconnect that connection, the 100uF is still reverse-biased (unless the output is biased WAY positive) as shown---which would account for a s**tload of leakage and noise.

Click to expand...

I will reverse the cap connection to see if that helps.

Why the huge value cathode resistors?

Click to expand...

The variable cathode resistor is used because I want to change the -ve grid bias voltage/bias point to see how it sounds, since this is a remote cutoff tube I want to be able to try biasing the grid down to -15V.

I have tried placing another 500ohm resistor in parallel but it didnt help in terms of noise.

Still got the windy vaccum noise tunnel output.....

With the pot turned down, the minimum resistance of the cathode is around 2K ohms which is still a bit high?? But even with around 500 ohm cathode resistance, it is not any better.... sigh....

:?

What's the ultimate goal of the circuit?

I see a lowpass to set the high frequency cutoff, and then some gain, in the two op amps, finally driving a tube with a fair amount of gain depending on bias (without the 100% feedback of the output-to-grid connection). With that feedback the op amp is basically shorting out the tube's output---the tube might as well not be there.

I agree you may need some cathode R but those values look way high.

Are you driving the op amp from a source with resistance to ground? When there is nothing connected the output will soar to about the positive rail---you might want to put a resistor to ground at the input.

Anyway, certainly some of your noise problems and probably your op amp heating are going to be from the capacitor leaking, and when it leaks it pulls up on the grid and eventually reverse-biases the 1uF grid cap as well---up up and away! But besides that the circuit has a lot of gain ( of order x30 from the tube at normal bias) so you would expect to get a fair amount of noise anyway.

I'd suggest a better way to play with bias is with another resistor (say a few hundred kohm) from the grid to the wiper of a pot across ground and -15V, rather than trying to use cathode current and its voltage drop.

[quote author="bcarso"]What's the ultimate goal of the circuit?[/quote]

I am looking to build various single tube circuits such as these in order to nail the biasing for each type of tube I have, also I can modularise these circuits by mix and match them for listening test and make it into a bigger circuit. They are intended as small building blocks I guess...

I see a lowpass to set the high frequency cutoff, and then some gain, in the two op amps, finally driving a tube with a fair amount of gain depending on bias (without the 100% feedback of the output-to-grid connection). With that feedback the op amp is basically shorting out the tube's output---the tube might as well not be there.

Click to expand...

Any suggestion for the circuit while including the feedback? I am goin to try pstamler's advise and put in a 47k resistor at the output of the second op amp.

I agree you may need some cathode R but those values look way high.

Click to expand...

Any suggestions?

Are you driving the op amp from a source with resistance to ground? When there is nothing connected the output will soar to about the positive rail---you might want to put a resistor to ground at the input.

Click to expand...

I will go ahead and try that.

I'd suggest a better way to play with bias is with another resistor (say a few hundred kohm) from the grid to the wiper of a pot across ground and -15V, rather than trying to use cathode current and its voltage drop.

Click to expand...

Thanks! Will try it now.....

Just off hand it looks like your plate resistor is pretty low in value, and the cathode high in value as mentioned. If you need to bring 100% feedback back to the grid to make it behave, something is quite wrong here. Consult some commercial circuits and tube data and see what they typically do for a single stage circuit like this. It should run just fine without any feedback except maybe a small (pF) cap to stabilize high frequency.

Update: I checked the tube data, and this tube is actually better suited for RF and IF amplifier circuits. You might have better success with a 6SN7, 12AU7 or something similar.

Try something like this:



The second opamp is now inverting, which means the overall circuit is noninverting; more to the point, it lets you bring the gain down as low as you want. There are resistors around the tube now, but for optimizing bias you should leave the switch open, so the feedback isn't operating. Make the tube as clean as possible without feedback, then switch in the feedback to see how much improvement you get. Using the OPA2604, you don't need a coupling cap to the grid -- at least, you shouldn't.

Oh, the lowpass filter is not maximally flat with your original values, but that may or may not matter in this application. If it does, we'll talk later.

Peace,
Paul

"Update: I checked the tube data, and this tube is actually better suited for RF and IF amplifier circuits. You might have better success with a 6SN7, 12AU7 or something similar."

Or at least a more common remote-cutoff pentode, if the circuit is going to be used for part of some variable gain app. This part was already relegated to the back tables in my 1958 RCA manual.

There are some tubes characterized for rf service that turn out to have nice properties for audio, although there are fewer circuits floating around for this. One guy I know is about to use a 6ER5 for a MM phono preamp. Another obsolete part but lots of NOS about evidently.

PS: Paul's suggestion is safe and sane, and playing with the feedback shold be instructive. If you could get your hands on bipolar or film caps of sufficient voltage rating for the output coupling or whatever we can all sleep better too.

> this tube is actually better suited for RF and IF amplifier circuits.

It fits the special needs of RF-IF work. It has low feedback, high output resistance, high Gm, and variable Gm.

Agree that my first-grab for power amp driving would be a low or medium Mu triode, not a pentode. But as long as it is in the socket, this tube can be triode-strapped and give Mu near 20.

> This part was already relegated to the back tables in my 1958 RCA manual.

Only because it is Octal and by 1958 all new-design had switched to minis. I think the 6BA6 (often 12BA6) is the mini, though it has higher GM than a 6SK7. I think supplies of 6SK7 are ample for DIY (anyway I gather that Learner has one in hand).

> I want to change the -ve grid bias voltage/bias point to see how it sounds, since this is a remote cutoff tube

In an ideal world: a remote-cutoff tube has more 2nd harmonic distortion than a remote-cutoff, possibly more gain near zero bias, and obviously less gain for very-negative bias. To some point, the sound will not "change" except in getting softer. Past that point, it will clip badly.

Trying to cut-off a remote-cutoff tube with cathode resistance is hard work. It seems to make more sense to semi-ground the cathode (50-200 ohms is traditional) and apply a negative voltage to the grid circuit. Especially since you have -15V laying around.

Trying to "nail bias" on a pentode is very difficult. You have two (or three) variables. I can hand-estimate a triode, but for pentodes I use the R-C tables. These do not cover low-impedance outputs suitable for driving a cable to a power amp.

For a number of reasons, audio gain control is rarely done with one remote-cutoff tube. Works for RF-IF, not audio. We need push-pull and also careful design of impedance and level.

It would be possible to build a gain=30 plus high-pass plus cable-driver in one 6SN7 (or two 6SK7) and no chips.

[quote author="BYacey"] Consult some commercial circuits and tube data and see what they typically do for a single stage circuit like this.
.[/quote]

I tried to look for one but haven't had much luck..... :sad:

[quote author="pstamler"]Try something like this:



There are resistors around the tube now, but for optimizing bias you should leave the switch open, so the feedback isn't operating.[/quote]

It's not very convenient to have turn on/off the switch every time you need to use the amp, eventhough I have no problem putting it in. That's the only draw back I am concerned about......... Thanks for the thoughts!!! :thumb:

Make the tube as clean as possible without feedback, then switch in the feedback to see how much improvement you get.
Peace,
Paul

Click to expand...

Man..... I think I have tried as many way as I can think of so far in order to achieve that but without any luck, no FB= hairy & noisy signal. Not only that, it has to be full FB. Even with full FB, it is still noisy......

[quote author="bcarso"] if the circuit is going to be used for part of some variable gain app. This part was already relegated to the back tables in my 1958 RCA manual.
[/quote]

I would try it for that application if I manage to bias the tube sucessfully as a class A amplifier, at the moment I am still struggling to achieve that. Well, kill the noise that is.....

I don't think Paul is proposing to leave the switch there all the time, or having to turn it on or off---it just means leave that connection open and get the amp as quiet as possible, then make the connection with the "switch" to find out how much the gain goes down and with it the noise.

After all this is a teaching experience, not something you are going to use directly for any application, right?

What does the noise etc look like on a scope at the output of the second op amp?

Buddy you have you a strange configuration, thus you are getting a strange result. This seems more like an experiment than a workable solution. I see wrong resistor values and wrong types of capacitors. Try not to use electrolytic caps in your feedback path, they leak voltage and thus change the circuit specs. Let the simulator guide you, not take the place of rational design criteria. SWC is better.

Analag

[quote author="PRR"] I think supplies of 6SK7 are ample for DIY (anyway I gather that Learner has one in hand). [/quote]

Make that a 100+ :grin:

That's why I am so desperately trying to make this thing to work "well", my problem now is the noise...... :roll:

In an ideal world: a remote-cutoff tube has more 2nd harmonic distortion than a remote-cutoff, possibly more gain near zero bias, and obviously less gain for very-negative bias. To some point, the sound will not "change" except in getting softer. Past that point, it will clip badly.

Click to expand...

Right, I was hoping to experiement the audible effect on the non linear transconductance characteristic as shown in the 6sk7 tube data sheet and the effect of single phase clipping by shifting the grid bias point along the load line.

Trying to cut-off a remote-cutoff tube with cathode resistance is hard work. It seems to make more sense to semi-ground the cathode (50-200 ohms is traditional) and apply a negative voltage to the grid circuit. Especially since you have -15V laying around.

Click to expand...

I was just looking for the shortest route in terms of wiring and complete the circuit but I will definitely try that now

Trying to "nail bias" on a pentode is very difficult. You have two (or three) variables.

Click to expand...

Yea, thats why I usually work out a reference point of biasing and then use a variable resistor to tweak around that point. Instead of soldering and desoldering like what I have done previously, that was a real nightmare!!!

I can hand-estimate a triode, but for pentodes I use the R-C tables. These do not cover low-impedance outputs suitable for driving a cable to a power amp.

Click to expand...

How does these RC tables work? Don't think I have seen one...

For a number of reasons, audio gain control is rarely done with one remote-cutoff tube. Works for RF-IF, not audio. We need push-pull and also careful design of impedance and level.

It would be possible to build a gain=30 plus high-pass plus cable-driver in one 6SN7 (or two 6SK7) and no chips.

Click to expand...

That sounds like a good circuit, cause I have those tubes! :green:

However, it might need output transformer?

I am going to rebuild this tube circuit again without any chips just to get the biasing right, hopefully I can solve the noise problem......

[quote author="bcarso"]I don't think Paul is proposing to leave the switch there all the time, or having to turn it on or off---it just means leave that connection open and get the amp as quiet as possible, then make the connection with the "switch" to find out how much the gain goes down and with it the noise.[/quote]

Right, I see...

After all this is a teaching experience, not something you are going to use directly for any application, right?

Click to expand...

Yea, trying to apply the theory and comapre the result to see if they match. If something is not right then it is a good indication of where I would need to revise and gain a better thorough understanding of it.

On top of that, I was hoping to use every single circuit as a single amplifer module as building blocks for a bigger circuit.

What does the noise etc look like on a scope at the output of the second op amp?

Click to expand...

I'll post some pics soon but first I have to rebuilt the circuit AGAIN, this time without the chips or anything just the tube itself.

[quote author="analag"]Buddy you have you a strange configuration, thus you are getting a strange result. [/quote]

Strange config???

I thought that was as simple and basic as it can get, all the parts in there are the min requirements I have just made them variable for tweaking.

This seems more like an experiment than a workable solution. I see wrong resistor values and wrong types of capacitors. Try not to use electrolytic caps in your feedback path, they leak voltage and thus change the circuit specs. Let the simulator guide you, not take the place of rational design criteria. SWC is better.

Analag

Click to expand...

I want to find a workable solution but I need to experiment in order to find it....

Care to suggest some resistor values? I will substitute the caps. Man...I just want to make this thing work, I'll try anything..... :green:

The 6/12sk7 beast shall be tamed :twisted:

First question is...what is that thing. I see the output shorting the control grid, so that's 100 percent feedback. Verify if the tube is good. Make a simple triode circuit and see if it works. To do it, connect the screen and suppressor grid to the plate. If it works, you can simply pull off the screen and suppressor off the plate and start tweaking from there.

Analag


Try this

[quote author="analag"]First question is...what is that thing. I see the output shorting the control grid, so that's 100 percent feedback. Verify if the tube is good. Make a simple triode circuit and see if it works. To do it, connect the screen and suppressor grid to the plate. If it works, you can simply pull off the screen and suppressor off the plate and start tweaking from there.

Analag


Try this[/quote]

OK, I'll try that.

Already built the circuit but will try your biasing value and see what I'll get, don't have access to the camera until Sunday so I'll post some pics then.

Thanks Analag! :grin: :thumb:

This better be good! :green: