Hissing in the wind

[sepulchre]

I just finished another "Dilla" - 50 watts, EL34s, etc. They usually come out quiet as church mice; with nothing plugged in there's no hum and only the tiniest bit of hiss with everything dimed. But this one, well it doesn't have any hum but the hiss is way out of hand, very loud. Even if it only shows up with everything cranked I can't sell it like this.

I do not skimp on parts; F&T, Sprague and SoZo caps, Dale resistors and so on. So what could be hissing? A bad cap? I haven't tried to trace it out yet but was wondering if anybody knew a likely culprit. Maybe I could narrow the scope of testing.

[martin manning]

Did you try changing preamp tubes, especially V1? Another possibility is a leaky cap. To trace the origin of the hiss start by pulling tubes beginning at the input.

[R.G.]

Let's start with taxonomy - enumerate the sources. (That means, in Bogart-speak, round up the usually suspects.)

All resistors hiss to some degree, the source being the random motion of electrons simply because of the temperature of the resistive material. This thermal noise can be computed from the resistance and temperature. You can't do anything about this source of hiss other than picking smaller resistances or running at liquid-nitrogen temperatures.

Some resistors also have excess noise. It's called excess noise because it it in addition to the bare minimum thermal noise, and appears to be related to the construction of the resistance materials and how current flows through them. It's worst in materials made of many small particles that resist by having the electrons flow through the tiny places where the particles touch. That is - carbon composition resistors, which are the universal examples of excess noise. Excess noise in carbon comps is variable, depends on the carbon particles, binder, manufacturing process, and the phase of the moon. Even with the best quality manufacturing and materials, some CCs will have huge excess noise and leak through the manufacturing quality control efforts. Film resistors, even carbon film resistors, do not display excess noise, and metal film resistors have none. Neither do wirewound.

Excess noise is related to intermittent contact noise. If a connection or contact is not solid and continuous, like a well made solder joint, the electrons have to leap across many tiny contact points, much like the process in CC resistors. This can be thought of as a multitude of tiny arcs, and this all sums up to hiss. Cold solder joints can do this. So can dirty switch contacts or connectors. So can physically cracked components. Reflowing all the solder joints as a practice is is aimed at fixing the solder part of this. I have experienced a handful of cracked but not separated components in my electronics work, so it does happen.

Ideal capacitors and inductors do not generate hiss; at least the capacitive and inductive parts do not. The conductive wire and plates are low value resistors, so they can have thermal noise, and poor connections to them can cause intermittent contact noise.

Active devices all hiss to some degree. In tubes, it's related to the "graininess" and "clumping" of electrons transiting the gap between the cathode and plate with statistically varying energies. That's the minimum value. But some tubes have internal problems giving them additional hiss above that - they simply are or go noisy.

A final hiss source that is rare in tube amps but always a possibility in solid state stuff is RF oscillation. If something is oscillating up in the RF bands in your amp, the amp will detect that and represent its phase noise as an angry sounding hiss in the audio. This isn't likely to be your problem.

In tube equipment, the first step in debug and repair is nearly always to swap tubes. There is a reason they are in sockets. You could have a hissy tube. But if you build amps, you already know this, and have probably already done this.

A bad-contact hiss source like a cracked resistor or wire, or bad solder joint can make huge amounts of noise and can be anywhere in the signal chain, but hiss is always made worse by amplification. So the input stage noise is always amplified by the gain of everything after it. Input stage noise problems are usually the worst ones.

You mention not skimping on components, naming some brand names. At one time, it was possible to count on a brand name for quality. In today's world, it's not; or at least, you can't count on it as much. For instance, you can still buy electronics labeled "RCA" and "Sylvania", two icons of the tube Golden Age, but those companies no longer exist; they are only labels, owned by someone else. General Electric may not exist much longer. Brand names are just logos and labels, and they no longer mean that the company who owns the label makes the stuff so labeled. The brand name may have been sold several times. Then there is the whole mess of counterfeiting. So brand names may be an indicator, but they are very much not a guarantee of quality. It's back to the same old story - buy from supposedly trust-able sources, and then be paranoid. Paranoia in parts procurement is a survival characteristic.

So with that as background, let's find and fix your hiss problem.
1. Swap tubes until you're sure it's not a bum tube, if you haven't already.
2. Turn the volume control all the way down, and any gain controls after the volume control all the way up. Does the hiss go down? If it does, that portion of the hiss that vanishes is before the volume control. If it doesn't, it's after the volume control.

That lets you concentrate your efforts in parts of the amp, not other sections of the amp. If that doesn't help..

3. Make up a capacitive signal shunt. This is a high voltage cap able to withstand the amp's voltages, that you clip from ground to signal points. Clipped from a grid to ground, it signal "grounds" the grid, and eliminates all signals, hiss included from being amplified further on. So you can walk from the input towards the output, grounding signals at grids and plates, until you find a place where it previously didn't make any difference, but now stops most of the hiss. Your hiss source is then between this point and the previous one. Obviously, the capacitor shunt has to be discharged between trials and then recharged to the new voltage, and you can shock yourself with it if you're not careful. It makes sense to turn the amp off to move the shunt, then back on. I used this trick to localize a hiss source to a bad resistor among the 200 or so resistors in a Vox Beatle amp once, after all other efforts, including using an oscilloscope, failed.

Also remember:
- the bigger the resistor value, the more thermal hiss
- the more voltage across the resistor, the more chance for intermittent contact and/or excess noise hiss
- reverb inputs are high gain inputs, too
- amps often use a very high value resistor, 3.3M or more, to isolate the reverb signal source from its recovery path

[xtian]

Awesome post, RG. This needs to be sticky.

You didn't mention the physical size of resistors as a factor. Isn't it true that a 1/8-watt resistor might show more noise than a 2-watt resistor of identical resistance?

[R.G.]

Good question, and one I don't have a ready and definitive answer for.

I'm currently (argh!) in the process of educating myself on it. I'll be back.

[pompeiisneaks]

Awesome post, RG. This needs to be sticky.

You didn't mention the physical size of resistors as a factor. Isn't it true that a 1/8-watt resistor might show more noise than a 2-watt resistor of identical resistance?

I can sticky it but it would be best as a new thread. If you could do that R.G. I'd sticky it.

~Phil

[Stevem]

Scope out which stage is adding the most hiss and if it's not the tube then start swapping out plate resistors and then cathode resistors .
A easy way to kill the gain from a stage for trouble shooting is to place a short across the plate load resistor.

[sepulchre]

Thanks guys, especially you, rg. Your expertise is shining, and the generous dose of common sense makes the task far less daunting.

I'll let you know what I find.