Ways to remove background noise?

[vibrovox7reverb]

Any posts on any ways to do this would be greatly appreciated. I've done an AC snubber, made sure the filaments on the power tubes are placed correctly, shielded all grid wires, and star grounded: cathodes of power tubes, power supply for main and screen, and the power transformer centertap all together, and star grounded each preamp cathode to its corresponding power supply ground. Still more noise than I'd like. I realize that some of it can from noisy tubes or lots of gain, but I'd really like to hear any circuit ideas or products that you recommend. Thanks

[cbarrow7625]

The best way that I know of to reduce noise is to make sure that all of your long wire runs in the amp are at the highest potential possible. That is, if you have a voltage divider (for example) don't run a short wire to the input of the voltage divider and a long run of wire to the tube grid afterward. Do the opposite, run the high potential signal as far as possible, go through the voltage divider and then run as short a wire as possible to the next grid. Every wire &component is an antenna and pick up stray noises, not to mention the noises induced by some components themselves.

I don't think this method really has any chance of helping to any significant degree with the typical turret board-and-wire hand wiring found in many older amps. Sure, they look nice but the "orderly" layout & wires running to the tubes really precludes getting the best noise performance out of them (especially for higher gain amps / channels).

PC Boards really have an advantage for this kind of layout. You can get the components right next to the tube grids. Even better (and what I like to use for my own amps) is a ground plane. Just a good old piece of double-sided un-etched copper plated PC board material. I mount my ceramic tube sockets right to it & then hand wire point to point from socket to socket. You can use insulated turrets or some other creative method for holding the components in place. I don't know if the ground plane really helps any with "absorbing" extraneous noise, but it sure does make for a super-quiet amp. I have a rats nest of a 5 channel low wattage amp built this way and everything from my "Fender"clean channel to my hot-rodded SLO/Dual Rectifier type channel are dead quiet. I mean dead quiet, almost zero hiss even when cranked.

Of course, you have to follow all of the other rules for good wiring and ground terminations (make sure you are wiring the proper grid ground & cathode ground very close to each other, for one). A nice advantage of this approach is that you can keep your heater wires completely isolated from your signal by running them (as wires, not PC board traces)on the underside of the double-sided ground plane and just poke them through next to the tube sockets. Again, I don't know how much that in itself contributes but it all adds up to an ultra quiet amp.

This is not an approach I have hear anybody else taking so I figured I would share. It is time consuming and tedious. It also may not be the "prettiest" amp innards you have ever seen when you are done (if you are into that kind of thing - I prefer tone first).Maybe it can help someone on their next build.

[vibrovox7reverb]

Wow, thank you so much! Those are some gem tips! I'm working out an old amp and using the existing board and some ground points for my own design and that kind of limited my layout, but I can still reground where it's necessary (without looking I can think of a couple of places), and I can check to make sure that areas of the circuit that have more potential have longer wires. After hearing about your idea with the unetched PC board I think I might try something similar on my next build. Thanks again, that was some awesome info

[Enzo]

I think you are using the term "noise" indiscriminately, or at least imprecisely. Noise is any unwanted sound, of course, but in amplifiers we tend to have hum and then other noises, usually like hiss and just random noise.

Hum ultimately comes from the power source. All the things you listed as having done fight hum. Hum comes from radiated power transformer fields, sensitive components or wires picking up radiated electronic fields in the vicinity, ground currents interfering, ripple, ground loops, heater issues, unshielded preamp tubes, poor grounding of circuits, etc.

Hum is not generic. There are many sources of hum, and each has its own cure. You cannot simply apply a bunch of hum abatement techniques and expect it to cure hum no matter what. All the ground schemes in the world won;t help a ripply filter cap. And all the filter caps in the world will do nothing to help a poorly grounded input or a heater leaking into a cathode. So to reduce hum you need to understand its source or sources. If six things are contributing to your hum, then six remedies will be required. And if a particulr thing is not adding any hum, applying its cure will do nothing to assist the other things.

And there are variations. Heaters can introduce hum in several ways, which you can counter. REferencing the heaters to ground helps, grounding one side is better than no ground, grounding the center tap is better (either a real center tap or a virtual one), then elevating the heaters to a DC offset can help too. And depending on the amp, sometimes running the heaters on DC instead of AC helps. And those are all just heater related.

WE have had many lengthy hum related threads, it is a wide topic often brought up. But there is plenty of other noise as well. Hiss or white noise is everywhere, and it is an artifact of gain. The more gain you have, the more hiss you will pick up. COmponents can add to this noise, certainly most of us have found noisy tubes and replaced them. But circuit layout matters as does exposure to the electrical environment, ie lack of shielding.

And then there is random noise. COmponents themselves can make noise. That little burbling staticky noise you sometimes hear in an amp is often a noisy resistor. Replace it and the noise is gone. But tubes and caps can also contribute such noise.

Other sources of noise are loose hardware. Little charges build up on parts. A loose mounting bolt on a power transformer can lead to hum as well as crackling noises. Same with loose nuts on controls and jacks. Weak or cold solder joints can get noisy. and so on.

[vibrovox7reverb]

My apologies, the hum level in the amp is relatively low, but the background hiss is what is bothering me in this project. If hiss largely a product of gain, does that mean that lower plate voltages will reduce hiss to a degree? Last I checked the plate voltages were rather high on this project, no higher than your average blackface super or twin, but still high compared to other crunchier sounding amps (voxes, marshall plexis). Would lowering the voltages reduce hiss to an extent? I probably should do it anyway because while I like the sound I'm getting it doesn't have quite enough distortion for me.

[vibrovox7reverb]

on a related subject, should the ground of the input jacks be grounded to the chassis?

[R.G.]

My apologies, the hum level in the amp is relatively low, but the background hiss is what is bothering me in this project. If hiss largely a product of gain,

It is and it isn't. Some hiss is unavoidable, but it can be minimized. I just did a polemic on the sources of hiss in another forum, so this is fresh on my mind.

There are three sources of hiss-like noise: thermal/Johnson/Brownian noise, contact noise, and shot noise. Shot noise is contributed by active devices and diodes, so you need to get better tubes to minimize that one. Thermal noise and contact noise are properties of resistors and contacts.

You minimize thermal noise by (1) using the lowest possible resistances, (2) the lowest possible temperatures and (3) the lowest possible bandwidth. You're kind of stuck here. You can't change the resistances without dramatically altering circuit operations, you probably can't affect the resistor temperatures much, and you need the bandwidth, so it's going to be what it is unless you want to redesign the circuit and possibly the signal source.

You can minimize contact noise, though. Contact noise depends on (1) the materials carrying the current and (2) the DC current through them. The primary offenders on contact noise are ... yep, you guessed! Carbon composition resistors. They have significantly worse contact noise (it's called "excess noise" in the literature) than any other resistor type. You *might* be able to lower the DC current in some places, but generally not without a big circuit redesign. You certainly can lower other contact resistances by making sure that all switch and bolted-down contacts are clean and shiny where they contact, and that all solder joints are well and properly soldered. Crummy solder joints can make really bad noise.

So to lower the contact noise, (1) remove all the jacks, switches and pots from the chassis. Clean the mounting locations and the bushings, then put them back in, tightening them down firmly. This will sometimes reduce hum substantially as a byproduct.
(2) Clean the contacts of the jacks, switches and pots with a good, non-residue cleaner.
(3) Clean and retension the tube socket contacts. Every contact counts!
(4) Re-flux and re-melt every solder joint, making sure each is clean and shiny.
(5) Clean the tube pins with a rubber pencil eraser; these are slightly abrasive.
(6) When you're done with that, consider what parts to change for lower hiss.

Now you're down to picking which and what to change. You mentioned gain and hiss. These are linked, because thermal hiss cannot be reduced for a fixed resistance level and temperature. And because it can't be eliminated, the hiss contributed by each stage is amplified by all stages after it. A quick bit of thought comes to the correct conclusion that the noise of the first/input stage completely dominates the thermal noise for a non-defective amplifier. It's hiss has more gain after it than any other sections'.

So you start by worrying about lowering hiss by working on the parts in the input stage.

It makes sense for lowest hiss to replace all the resistors in the input stage with equal value resistors that are not prone to excess noise. That's as good as you can do without redesigning the amplifier circuits for lower currents, lower temperatures, etc. You can't do anything about thermal noise, but you can get low excess noise. The lowest common types for excess noise are metal film and wirewound. Hiss should go down if you use metal film for the input resistors, bias resistor(s), and plate resistor of the input stages, and if you've done a good job on the contacts.

There is another noise source that can get you. Notice I mentioned a "non-defective amplifier". There is a soft failure mode in plate resistors of tube amps. These are often 100K to 270K carbon comp resistors. It is common for a carbon comp plate resistor to develop really bad "popcorn" or "shot" noise; it sounds like popcorn or crackling sheets of paper. In the beginning, it can mimic rising thermal and excess noise. It's not a bad thing from a noise point of view to arbitrarily replace plate resistors with metal film or wirewound.

"Butbutbut..." I can hear you thinking, "what about carbon comp mojo?"

There *is* a carbon comp mojo. There is a very small second order distortion caused by the voltage coefficient of resistance of carbon comp resistors. This only affects the signal in noticeable ways for signal voltages of over 50-70V peaks across the resistors. It's not noticeable in the first few tube amp stages because of the lower signal levels there. Metal film at the input won't detract any carbon comp mojo.

does that mean that lower plate voltages will reduce hiss to a degree?

Excess noise is proportional to dc current, so lowering plate voltage will lower hiss a bit, but it's not a huge thing. Better to sidestep it with lower noise resistors.

[R.G.]

on a related subject, should the ground of the input jacks be grounded to the chassis?

Hmmm. The signal ground must connect to the AC-safety-grounded chassis somewhere and somehow. Exactly how that happens depends on your grounding scheme. There is one scheme that is known ahead of time to produce the minimum hum/noise; that is, there should be exactly one (1) wire connecting the chassis to signal ground in a perfect world. With only one connection, it becomes impossible for currents in the chassis to cause signal interference in the signal path. People get acceptable results with other schemes involving multiple connections, but it may take substantial experimentation to find a good way. You will get horrible hum if the chassis is not connected to the signal ground in some way.

[vibrovox7reverb]

Oh, now that you mention it I have really noticed the noise increase since I switched from carbon film to carbon comp throughout my amps, especially in the reverb section. I think my solder techinique may have something to do with it too, I generally try to get the joints to look like some of the nice joints I've seen in Fender amps and I do not leave cold solder joints. However, I may be using to much solder, or perhaps I'm heating the components too much before I apply it and causing damage. I have only worked from vintage amps because they're more fun to work out of, easier to afford (sometimes) and because I don't have the money to pour into a fresh cab and chassis etc, so cleaning should probably help a LOT too. Thanks so much for all of this info on hiss! This board is awesome, and I love coming here because eveybody is extremely knowledgable, friendly, and helpful. You have no clue how grateful I am for your help.

[Chuck H]

I'm only posting to track this thread... Nothing to offer at this time.