Bizarre hum problem from tone stack

[harry]

Ah, that makes sense - useful to know too! Thanks

[mykey]

I'm getting some hum that I had assumed was 60-cycle hum coming from a mismatch in the output tubes (as it was present when I accidentally cut the power to all my pre-amp tubes). However, when I turn down the treble control on the pre-amp, the hum progressively disappears until you turn the control completely down, when there is (almost) no hum. The output is fine when you do this, although the treble control still functions correctly, so all the treble is rolled off...

If the 47p is a ceramic disk capacitor, these can cause quite a bit of 60 cycle introduction to the signal path. suggest you change it to a silver mica capacitor. the ceramic disk picks up the power supply AC fields, especially heater wiring.
another dandy suggestion is to use metal film resistors instead of carbon composition. the carbon comp resistors have glued on leads and high thermal noise. the metal film leads are welded on, noise is much lower.
one more thing is to make sure that the treble control potentiometer case is grounded solidly to the chassis. did you check that?

[R.G.]

Ah, good. mykey's back.

I'm getting some hum that I had assumed was 60-cycle hum coming from a mismatch in the output tubes (as it was present when I accidentally cut the power to all my pre-amp tubes). However, when I turn down the treble control on the pre-amp, the hum progressively disappears until you turn the control completely down, when there is (almost) no hum. The output is fine when you do this, although the treble control still functions correctly, so all the treble is rolled off...

Say, mykey, can you learn to use the quoting function correctly? Your reply simply copied his whole post in and didn't set it apart so it's very difficult to figure out what he said from what you said.

If the 47p is a ceramic disk capacitor, these can cause quite a bit of 60 cycle introduction to the signal path. suggest you change it to a silver mica capacitor. the ceramic disk picks up the power supply AC fields, especially heater wiring.
another dandy suggestion is to use metal film resistors instead of carbon composition. the carbon comp resistors have glued on leads and high thermal noise. the metal film leads are welded on, noise is much lower.

Meanwhile, back at your advice. Can you tell me how it is that a ceramic cap introduces hum to the signal path? How does it get there? Is is capacitively coupled in, inductively coupled in or conductively coupled? What is it about mica that's different from ceramic in causing this?

Then there's carbon versus metal film. The man is complaining about hum, and you're telling him to change resistors to get rid of thermal noise. Why is that? That's not what he asked, it's an inappropriate response, a kind of primal non sequitur. It just confuses things. Just as an aside, the lead mounting method is NOT why carbon comp resistors are noisier than metal film. Can you figure why it really is?

Then there's the sound advice -

one more thing is to make sure that the treble control potentiometer case is grounded solidly to the chassis. did you check that?

The only problem with that is I already told him that. Did you even read the earlier responses? I said:

For instance, there is some hum coming in if the control pot bushings are not fully grounded to chassis.

You see, any of the control pots not being grounded well to chassis can to it, not just the treble pot. 60 Hz is pretty low frequency compared to what the controls filter.

Welcome back to the sandbox. But do try to keep up and don't slow down or annoy the other kids.

[Enzo]

So from zero to 1 the hum drops, and then from 1 through 10, the volume grows, as you turn the knob?

I think you are hearing two sources of hum. The circuit itself is amplifying some hum signal. The more you turn it up the louder that hum gets. But there may also be some background hum from a different source. If the two sources are out of phase with each other, the lower level second course is slowly cancelled out by the main hum . At the 1 position on the dial, the two hums are at about the same level, so out of phase they cancel. As you turn higher, the main hum gets louder and louder, the background hum has already cancelled all it could. So you don't notice that the loud hum is really ever so slightly less than it would be without the little bit of cancellation.

[mykey]

[QUOTE=R.G.;26930]Ah, good. mykey's back.


Say, mykey, can you learn to use the quoting function correctly? Your reply simply copied his whole post in and didn't set it apart so it's very difficult to figure out what he said from what you said.


it seems to be fairly discernible for those of us who read the English language.
the suggestions for using lower noise parts are quite applicable to this noise problem. he has already changed the 47p and says the noise is lower. changing to silver mica cap and non inductive (metal film) resistors will also lower the 60 cycle noise, only more. whether or not anal retentive types care about welded leads is a matter of personal taste, you tell us.
I have given some good suggestions based on actual experience. Those who wish to lower the 60 cycle noise can try these suggestions. those who wish to heckle in order to satisfy inferiority complexes may also do so.

and while im on the subject it's quite helpful to use small gauge shielded wire here in the tone circuit, ground the shield only on one end of the wire (the side connected to the treble control). i like to shrink wrap the other side of the connection to keep it neat...
and twist your heater wires to help cancel the AC field (this is the old style method)
then keep the signal wires (to control grids) as far away from the heater wires as possible, it's the AC field bleeding into the signal path that causes this buzzzzzz...
to take it a step further use DC heaters, this is the most noticeable improvement of all...to get the AC out of the circuit altogether.
and also you can increase the size of the filter capacitors so that there is less ripple in the DC supply, it is this ripple that is being amplified along with the audio signal.
its usually a combination of several things (as above) that will quiet the AC ingress to an acceptable level.
what's really happening is that the parts in the circuit such as wire, caps, resistors induct AC from the power supply and heaters, and anything you do to reduce the AC
entering your audio path will lower your noise. ceramic disk caps are the worst offenders, carbon resistors the second worst, unshielded wiring third. This is why the
best amp designs with the lowest noise use DC heaters. So then the more gain your preamp has, the more you need to reduce the AC ingress to obtain an acceptable
noise level.
In older Marshall tube amps (plexis) for example, the bias supply filter cap was connected to the tone and volume control ground. by moving that filter cap ground to the opposite side of the chassis, the noise was noticeably reduced. this is why in many amps the power supply grounds are separated from the signal path grounds, moved far apart but still connected to the chassis. connecting all the grounds to a central point (star grounding) does not always produce the lowest noise!
then there is the consideration of the ground loop, which is another factor in reducing the AC noise of the audio path. (and that subject is another story).

[R.G.]

it seems to be fairly discernible for those of us who read the English language.

Taken a poll of your readers, have you? Or is it just you who can read the English language?

the suggestions for using lower noise parts are quite applicable to this noise problem. he has already changed the 47p and says the noise is lower. changing to silver mica cap and non inductive (metal film) resistors will also lower the 60 cycle noise, only more.

You're missing the issue again mykey, either because you're not picking up the content from what I wrote or because you just missed it. Fact is there is nothing about capacitor construction or materials that can reach out and pull in AC hum from the air. There is no way which is not dependent on where and how you attach the leads for a ceramic cap to pick up more electrical hum than a mica cap can. The cap's dielectric material does not care about hum in the air. Where the leads go and how it's wired make the difference, not whether the cap is made of ceramic or laminated belly-navel lint. Contending otherwise indicates that you don't know how caps work.

Then there's that note of yours -

non inductive (metal film) resistors will also lower the 60 cycle noise, only more.

The problem with that is that metal film resistors are NOT non-inductive. They are a layer of metal oxide on the surface of a cylindrical ceramic pellet that's etched off in a spiral around the pellet. Conductive spiral... hey, that's an inductor! They are inherently inductive unless you pay extra for specially made noninductive ones that have fancy spiral patterns to cancel the inductance. Beyond THAT issue, the inductance is so tiny that it's almost unmeasurable at 60Hz so even if they're inductive it makes no difference. What you just told us is that you don't know about the construction of metal film resistors and that you apply parts based on superstition.

whether or not anal retentive types care about welded leads is a matter of personal taste, you tell us.

Well, you're the one who brought up glued on leads, right?

Besides, "anal retentive" means that a person has learned to control their bowel movements to when they choose to do them. Look it up.

I have given some good suggestions based on actual experience. Those who wish to lower the 60 cycle noise can try these suggestions. those who wish to heckle in order to satisfy inferiority complexes may also do so.

Let's be clear - mykey, you're putting out a mix of fact and fiction that you may or may not understand; it's certain that you don't understand parts of it as I've pointed out. I don't care whether you are confused or not, but I don't like it when you spout half-truths to the beginners who could take that stuff to heart. If you're going to advise beginners, at least have the courtesy to get the details correct. Not everyone who tells you where you've messed up is heckling you, nor has an inferiority complex. It may just be that they know what they're talking about.

Here's an offer specially for you, just to help you learn the facts. If there is any part of what I say that you think is factually incorrect, say so. I will post substantiating references here for it, or publicly retract it. I challenge you to do the same.

Please - get your facts straight. It's not helpful to confuse the beginners who need help.

[mykey]

now we have looked at the cause of 60 cycle ingress
1. A. coming from cheaper parts such as disk capacitors
B. or carbon comp inductive resistors
C. or unshielded audio paths
which induct 60 cycle from the power supply and heaters, then the gain stages amplify the 60 cycle along with the guitar audio...
and we have looked at those who try to compensate for inferiority complex by proclaiming themselves superior to all others...
A. by demanding theoretical explanations and mathematical formulas
B. by proclaiming that certain solutions offered CAN'T possibly solve the problems described.
C. or by persistent general heckling of others who join the discussion with differing viewpoints.
then third we must also consider the possibility of this hum being caused by high frequency oscillation, it seems to be in the tone stack because when the tone control is turned the oscillation occurs.(hence the "bizarre" classification) REALITY IS THAT the problem is not really the tone stack at all, but the tone stack ALLOWING high frequency to pass thereby TRIGGERING the oscillation.
In this scenario (which is well understood by experienced builders) the actual layout of the components and wiring will cause self regeneration in the signal path, triggering a high frequency which is too high to be heard by the human ear or to high to be passed by the output transformer and speaker. The oscillation can be detected with an oscilloscope.
the resulting high frequency sine wave overloads the output tubes taxing the power supply to the point of increased AC ripple, and the listener hears this amplified AC ripple (HUM) from the speaker.
Yes the oscillation occurs only when the tone control is operated in a certain position, allowing the high frequency to pass. when the tone control is turned down, the high frequency cannot pass and the oscillation stops.
So then here are some of the methods employed to prevent this high frequency oscillation from occurring:
1. Again using higher quality component parts as I have described, and by shielding the wires which connect to the grids of each stage.
2. keep the input (grid) and output (plate) wiring separated. reroute wiring to archive this separation.
3. separate grid resistors from plate resistors with physical distance. separate the tone stack components from the plate resistors.
4. A. employ negative feedback which limits the high frequency response of each stage. try connecting a small silver mica cap (example 10- 50 pf) between plate and grid of each preamp gain stage until you find the stage that is oscillating. (do not rule out any stage, it's usually V1)
B. connect small cap (try .001) between each preamp grid and ground till you find which stage is oscillating.
C. or another handy method is to connect a cap (try 50 pf) between the output tube control grids. this method is used by Marshall and other manufacturers.
(have a look at some older tube amp schematics to get a good idea in high frequency limiting methods)
we have not seen the complete schematic and this may be helpful in providing some other suggestions, however following the suggestions provided with some experimentation (to find the best combinations of method) should solve your hum problem.
be patient in experimenting with layout, shielding and negative feedback to "tweek" the amp into behavior. there is no real "formula" to achieve a working layout other than experimenting with the general guidelines. This is why all successful tube amps are built in a certain layout pattern.

[mykey]

the parts mentioned are most definitely non inductive and check the manufacturer's specifications for yourself. this is the reason that highest quality amp manufacturers are using these types of parts for lowest noise highest gain performance.
I have not claimed any "magic", only lower noise and less oscillation.
readers don't be afraid to experiment and try suggestions offered by the amp builders such as myself. YOU can decide what works best, no need to be corralled by "rules" from self proclaimed know it alls.
building a guitar amp from an old PA head is a really cool project, and the more you experiment the more will be learned, try new ideas. Patience and determination pays off.

[harry]

1. Again using higher quality component parts as I have described, and by shielding the wires which connect to the grids of each stage.
2. keep the input (grid) and output (plate) wiring separated. reroute wiring to archive this separation.
3. separate grid resistors from plate resistors with physical distance. separate the tone stack components from the plate resistors.
4. A. employ negative feedback which limits the high frequency response of each stage. try connecting a small silver mica cap (example 10- 50 pf) between plate and grid of each preamp gain stage until you find the stage that is oscillating. (do not rule out any stage, it's usually V1)
B. connect small cap (try .001) between each preamp grid and ground till you find which stage is oscillating.
C. or another handy method is to connect a cap (try 50 pf) between the output tube control grids. this method is used by Marshall and other manufacturers.
(have a look at some older tube amp schematics to get a good idea in high frequency limiting methods)
we have not seen the complete schematic and this may be helpful in providing some other suggestions, however following the suggestions provided with some experimentation (to find the best combinations of method) should solve your hum problem.
be patient in experimenting with layout, shielding and negative feedback to "tweek" the amp into behavior. there is no real "formula" to achieve a working layout other than experimenting with the general guidelines. This is why all successful tube amps are built in a certain layout pattern.

Thanks for the advice, I will definately try these if changing to higher quality components doesn't help.

[mykey]

Thanks for the advice, I will definately try these if changing to higher quality components doesn't help.

i think from what you describe, you are getting a high frequency oscillation
and this is what is really causing the hum you are hearing.

[Rob Mercure]

mikey

Ah, the "English language" - well, correct technical usage is that you "induce" electromagnetic radiation not "induct" but I'm sure you really don't care much about that. The funny thing is that you seem to want to portray R.G. as some sort of elitist who wants to censor you when over the years his advice has been well written, technically correct, and self-corrected/updated when detected/needed. You on the other hand seem to be totally unable to admit, even to the slightest degree, that you might be incorrect or that you might simply poorly state a response/opinion - why is that "Mr. Expert?"

Obviously you read - and I do sincerely congratulate you for that as so few people want to be "bothered" to actively pursue knowledge. But then you seem to not take the second part of that "challenge" and attempt to integrate the "lore" that you've accumulated with real world fact and experiences - both those you've accumulated and those of others. "Experiences" that are assumed to be "knowledge" based on a single occurance are referred to as "superstition" - it's only by repeated trial and error that accurate conclusions are drawn. So, you know a lot of good stuff but you slather it on thickly without reason.

In particular I find the premise that higher-than-audible oscillation can so load an amplifier's output stage (and the output stage is the only part that dissipate sufficient power for this to occur) that the power supply would "pulled down" below the filter recharge capabilities thus creating "hum" - or at least for any length of time - to be a specious response to the thread. While such freak occurances happen the usual result is a red plated output tube and a blown fuse or rectifier. This isn't the cause of long term hum problems - and you keep mentioning "60 cycles" - unless you've got a 1/2 wave PS the increase ripple would be 120 Hz! And it's little technical issues that take your good suggestions and tend to make them look like BS! So, are you simply "sloppy" or technically deficient? If the former - and I hope this is so - then you "owe" it to those you advise to edit/revise your posts for accuracy! (I'll often start a reply and then realize that I'm just not clearly responding - then I simply leave the partial reply on the screen and go off and do something else until I think I can reply clearly. So what if I'm not the first one to respond to a query - if we're here to help others and not stroke off our egos then an accurate response from another member accomplishes the same thing. But somehow I think that you keep a roll of paper towels near your keyboard!)

And I am - as others are - still waiting for you to give any reply concerning current being induced in a capacitor and what effect either the electic or dielectric material has on the degree of induction/coupling. Umm, do you understand the difference between electrostatic and electromagnetic fields?

Well when I attempt to find something like common ground with someone that I may not be in total agreement with on this forum I usually close my response "with respect" but in this case you've not only not earned it but seem to taken pains to ensure that none of us give it to you. And that does pain me cuz I'd rather have another "virtual" companion in this hobby/obsession than someone whose posts I ignore or perversely review when I'm grumpier than usual (simply curious today actually).

Rob

[mykey]

the problem of HF oscillation which occurs when the tone stack allows the HF to pass is quite common, even in factory production amps. the suggested methods do often cure the problem. and yes, it does cause audible 60 cycle hum when it occurs, usually because the output tubes, power supply are being driven to limit.
at times it will cause the output tubes to turn red, but not always.
and this is the difficulty of trying to cure that problem when a o scope is not available, its hard to verify when the user can't hear anything except the hum.
but actually there is nothing rare about it, many of the newer Fender and Mesa Boogie amps suffer from the same symptoms, Seymore Duncan amps had the exact same problems.
many techs these days do not recognize the problem and often blame it on loose soldering connections, etc...the older tube techs were very familiar with these symptoms, and how to cure them.
yes, the parts in the amplifier circuit do radiate and interfere with each-
other. the parts "talk" to each other! this is why component layouts need to be separated, "tweeked", and why wires need to be shielded.

[harry]

Well, here's a funny story: I figured out that one of the speakers I was using was wired out of phase. Correcting this reduced a lot of the hum.

The remaining hum was then coming from before the volume control - this is the embarrasing bit - caused by a dodgy lead.

It still hums a tiny bit, but it's almost there. The high frequency oscillation's interesting. At high gain level's I'm getting screeching feedback, I assumed something was microphonic though.

The next thing I want to address with the amp is it's output - it's cathode biased but the tubes are running far too hot (orange glow) and the output doesn't seem like a true 50 watts. Could a high frequency oscillation cause this?

[harry]

...oh, and I should mention that the tubes were running hot before I started modifying it.

[Alex R]

In the absence of a scope, jumpering a 50pf cap to ground at various points in the circuit should show us if oscillation is causing this. I have to say that whilst oscillation is a common fault that causes all kinds of bad stuff I've never seen it make power tubes pull so much current they glow red. From my experience therefore I'd say this seems an unlikely explanation. But try the cap to earth - from the signal path before and after the tone controls, and maybe across a plate resistor, to dump any HF oscillation out of the circuit.