12ax7 noise.. which side is best?

[Merlinb]

Gary - Thanks! So it's the shared cathode resistor, not the increased gain which reduces the noise level relative to the signal?

No, Gary is mistaken, but he continues to tell people that the bias resistor causes most noise. In fact it causes almost noise noise, and this was proved in another thread: http://music-electronics-forum.com/t...08/#post170708

In parallel amplifiers each tube amplifies both the incoming signal, and the noise generated inside the tube itself, equally. However, the two tubes add the audio signals add together at the output. But, the noise signals produced in each tube are random so when they are added they partly cancel out. Overall then, signal gets amplified more than noise.
Theoretically you get a 3dB S/N improvement every time you double the number of paralleled devices.

[Robert M. Martinelli]

Here we go again....
If memory serves me well we've been dealing with this in a recent thread, so a little search could be useful, anyway, the amount of noise reduction is equal to 1/Sqrroot(n) where "n" is the number of devices being paralleled, so, as Merlin correctly pointed out, paralleling two devices yields a -3dB noise reduction (0.707, or 1/sqrroot(2)), paralleling four devices halves the noise (-6dB, 0.5 or 1/sqrroot(4)) and so on.

Also, two things to remember, first, as we have already seen, the noise generated by a resistor is proportional to its value, and the cathode resistor is one of the least meaningful (because of its usually low value) in terms of noise, secondly, the cathode resistor is very often (almost always) bypassed by a large capacitor, so AC-wise its contribution to noise is non-existent, because whatever AC that might be present at the cathode (including noise) is shunted to GND by the bypass capacitor.

Hope this helps

Best regards

Bob

[mooreamps]

heh, heh, heh.....


You know what I find somewhat amusing.... I just finished do a set of gmm modes to a old Fender Champ a friend of mine brought in last week. It used to hiss like a snake. Now it does not... I don't know if either you two "actually" work on equipment ; or design and build equipment like I do. . But when I write comments like this on these issues, and am speaking to the guy asking the question ; Not You Two !! Secondly, they are comments based on measurements taken from "really live equipment". So, in the future, if you have some aching need to knee jerk a post ; just don't..... .......

-g

[Chuck H]

They're right Gary... Sorry. I do work on live gear. It's what I do. And I doubt any one with the nerve to post something contrary to any post would do so lightly without practical experience.

There are different types of noise. Hiss and hum being the most common. Since heater circuits are generally low impedance and love to inject hum on cathodes I can agree that lowering cathode resistor values will reduce hum. This effect would still be almost insignificant on fully bypassed cathodes. But hiss??? considering the relative impedance of the signal path compared to the cathode circuit and the fact that most cathode circuits are bypassed I can't see any technically explainable reason how hiss could be improved by modifying the cathode circuit. I'm one of the least technically saavy designers on this board and this seems like common sense to me.

JM2C

[mooreamps]

then they can duplicate the same thing, the same way I do. If they don't want to do that, then they have no leg to stand on.....

-g

[Robert M. Martinelli]

heh, heh, heh.....


You know what I find somewhat amusing.... I just finished do a set of gmm modes to a old Fender Champ a friend of mine brought in last week. It used to hiss like a snake. Now it does not... I don't know if either you two "actually" work on equipment ; or design and build equipment like I do. . But when I write comments like this on these issues, and am speaking to the guy asking the question ; Not You Two !! Secondly, they are comments based on measurements taken from "really live equipment". So, in the future, if you have some aching need to knee jerk a post ; just don't..... .......

-g

Sorry,
I'm neither taking the bait about this "technically", nor will I add lengthy explanations about the decades, not years, in which I've been working on equipment, designing, repairing, modding, building and so on...what I want to emphasize is the "exposed nerve" and quite offensive attitude you showed....

You have to remember that you're a member of a forum, so you have the right to "speak" and freely express your ideas, and, as Voltaire once said, "I don't agree with what you say but I'd die for your right to express your ideas".

Other members, OTOH, do have the same right to "speak" and express their opinion, as long as they don't insult anyone. If you thought that, since you answered the guy, you gained the EXCLUSIVE right to "talk" with him, well, let me tell you you completely misunderstood the reasons why forums like this exist, so you'd better reconsider the terms of your partnership. Many issues have been solved in the past by people "jumping in" and offering their help...It's in the forum nature!

Neither Merlin or I did insult you, or mean to do so...we just pointed out what decades of experience (not to mention countless textbooks) tell us...If you perceived that as knee-jerking, that's a problem of yours and your exposed nerves.

Maybe my English is not as good as I thought it was, but I hope I've made my point perfectly clear...If not, I encourage other members to "jump in" and correct me if I'm wrong, so that I can learn something new from this.

Period (and I really mean it).

Best regards

Bob

[Merlinb]

I don't know if either you two "actually" work on equipment ; or design and build equipment like I do. . But when I write comments like this on these issues, and am speaking to the guy asking the question ; Not You Two !!

Yes, but when you're answering his questions with nonsense, we are to be forgiven for contradicting it.

And it's not just us two. You remember the thread. Everyone with an ounce of real circuit knowledge showed up your theory as pure fantasy.

[Steve Conner]

One of my homebuilt amps has gain control by a variable cathode resistor in the first stage. (The variable resistance is in series with the cathode bypass cap, so it doesn't change the DC conditions of the stage, it remains biased at the regulation 1.5mA or whatever a 12AX7 is supposed to run.)

As I decrease the resistor to increase the gain, what do you predict happens to the noise? I say that it goes up, for two reasons:

1: The smaller the cathode resistor, the less negative feedback to cancel the tube's internal noise.
2: The smaller the cathode resistor, the higher the stage gain and the more the thermal noise of the guitar and first stage grid resistor is amplified.

I described the absolute noise level, but what do you predict happens to the signal-to-noise ratio? After all, decreasing the cathode resistor also amplifies the signal more.

And, would this trend change if the variable resistor was not AC coupled, and also affected the DC conditions?

How would the SNR be affected once the output from the stage became high enough to overdrive following stages?

Here is the schematic of the first stages. http://scopeboy.com/baxeq2.gif

Remember, this equipment is not a fantasy, it actually exists and I can drag it into the lab to check your answers.

[R.G.]

then they can duplicate the same thing, the same way I do. If they don't want to do that, then they have no leg to stand on...

Hi, Gary. Odd that (once again) in your view other people have to prove what they say and you don't. What a surprise!

I have this theory that for any given situation, there is a quotation from a C&W song somewhere that applies. In this case, the appropriate line is from Gary P. Nunn's "London Homesick Blues":

"They say 'You're from down South,
And when you open your mouth,
You always seem to put your foot there."

I did. I wrote a mathematical proof that showed...

Which, of course, cannot be shown for any kind of external review, for national security reasons. 8-) We don't want Al Qaeda getting hold of something like that, do we?

Didn't you also have a mathematical proof that fuses exist primarily in an extra dimension of N-space, or something?

[TheTinMan]

No one has answered on of Mike's questions: should the cathode, anode and grid of the unused triode be grounded?

Or, based on text and photo in post #16, Mike do you have parallel triodes for the first stage now?

Last, is the noise floor for a 12AX7 triode constant or does it increase with gain? Somewhere else recently there was a discussion about this. The consensus (AX84.com maybe?) seemed to be that bypassing the cathode with a cap on the first stage was beneficial because it raised the signal level relative to the noise floor. Does that make sense?

Mike, that last question may seem like a hijacking but it relates to you increasing the value of the plate resistor on the first stage.

Cheers,

Chip

[guitarmike2107]

Thanks chip, I have moved on from this now, I ended up grounding all three unused pins, not sure that it made that much of a difference though. I did compare the hiss to some of my other amps and it’s actually pretty decent. Its actually one of my favourite amps now…. for the time being of course

I always had parallel triodes for the first gain stage, which is not that noisy actually, my problem was the second gain stage after the volume control. I might just parallel that triode up to prove a point, but I don’t think I need that extra bite added to this amp.

[TheTinMan]

Mike - Who could blame you for moving on after this brew-ha-ha?

One last thought based on the location of the tube with the unused triode. Have you tried substituting another volume pot? Maybe the hiss is coming from that relatively large carbon comp resistor?

Cheers,

Chip

[Merlinb]

I always had parallel triodes for the first gain stage, which is not that noisy actually, my problem was the second gain stage after the volume control. I might just parallel that triode up to prove a point, but I don’t think I need that extra bite added to this amp.

In guitar amps practically all hiss comes from resistance in series with grids, not from the tubes themselves. This can only be reduced by using smaller resistances, less attenuation between stages, bulk metal foil/wirewound resistors or a radical re-design.

[randyc]

Late To the Discussion

This is an old thread, encountered serendipitously while searching a related topic.

Thanks for the mention of my book, written during an unpleasant period of recovery from cancer surgery. (My wife purchased a laptop computer, suggesting that I consolidate comments I had posted to various internet forums. The exercise provided me with many hours of constructive entertainment.)

I noticed comments regarding spreadsheets that I developed to support some of my approximate design procedures. I can provide these if a file-sharing site (not the flakey one that I selected) can be suggested. A private message would be good.

Best wishes,
Charles R. Couch
“Designing Vacuum Tube Amplifiers and Related Topics”

[R.G.]

There is that plate resistance thing, and the hiss from the non-analog nature of electrons hitting the plate, which is much worse at low currents.