Moot point to tie a buffers drain/collector to ground with a high value cap?

[Boner]

So I've gotten in the habit of putting a 10 ohm resistor and 100uF to 200uF cap on every transistors high rail.
Like this guy.

http://www.rason.org/Projects/jfetamp/jfetamp.htm

.. even buffers. Is that a huge wast of space and money?

I dont think I notice a difference noise when a buffer has it or not, but my hearing blows.

[marcelomd]

Looking at the last drawing, looks like it is a single resistor + capacitor per circuit, which is normal practice.

The resistor + capacitor on the power rail is there to filter noise coming from the power supply. Use only one at the very "start" of the power rail, after polarity protection, if you have it.

Bonus: Use one 100nF ceramic capacitor between +V and GND physically placed as near as possible every integrated circuit, such as opamps (one between +V and GND and one between -V and GND if using dual supplies).

For higher powered circuits, such as tube amps, I've seen one capacitor (or a bank) per "stage", but I don't have enough experience with those.

[anotherjim]

I dont think I notice a difference noise when a buffer has it or not, but my hearing blows.

It's one of those things that you won't miss until it's not there. You can't expect an externally powered pedal to always have a good clean power supply. However, when the pedal is battery powered, it isn't really doing anything since a battery might run down, but it doesn't go noisy.

[ElectricDruid]

It certainly won't hurt, but it's probably overkill. As Marcelomd said, the typical way is to use one at the power input, and leave it at that.

What I would say is that 10R/100u or even 10R/200u isn't actually enough to provide *any* attenuation to mains hum, which is a pretty common type of noise on a..erm..power supply. It'll block hiss, but that generally comes from elsewhere.

10R/100u = 160Hz cutoff (-3db point)
10R/200u = 80Hz cutoff

I'd look at upping the value of the resistor a bit. You could go to 47R or 68R, or even 100R if the circuit doesn't draw much current. 10R is very low.

[Boner]

Apologies for bringing up a dead thread!

So are you guys saying just do 1 resistor/cap combo and tie all transistor power connectors? No point and is a waste of money to have a resistor/cap combo on each drain?

[amptramp]

I have seen successive stages decoupled with one R-C filter when the first stage is inverting.  Thus, when one collector / drain / plate goes high (current is reduced), the other goes low (current is increased).  This means the variation of the total current over two stages is less than that of one stage.

[Kipper4]

You should only need one filter for your power supply. It's not necessary to have one on each of the "drains"

Heres an article by AMZ
http://www.muzique.com/lab/hum.htm

Apologies for bringing up a dead thread!

So are you guys saying just do 1 resistor/cap combo and tie all transistor power connectors? No point and is a waste of money to have a resistor/cap combo on each drain?

[R.G.]

It's a matter of objectives.

(1) Is there an objective? If not, you're inserting magical amulets for generalized luck/goodness.
(2) If there is an objective, what is it? This is important, as pointed out humorously by Lewis Carroll:

"Would you tell me, please, which way I ought to go from here?'
'That depends a good deal on where you want to get to,' said the Cat.
'I don't much care where -' said Alice.
'Then it doesn't matter which way you go,' said the Cat.
'- so long as I get SOMEWHERE,' Alice added as an explanation.
'Oh, you're sure to do that,' said the Cat, 'if you only walk long enough."

― Lewis Carroll, Alice in Wonderland

Here are some possible objectives for a series resistor/shunt capacitor on a power line:
1. Incoming noise removal. This happens by voltage divider action, where the series resistor and shunt capacitor are the voltage divider. The "divider" part works for AC only, since the cap should not pass DC, and noise is reduced to
the amount of Xc/(Xc+R). As frequency goes up, Xc goes down, and AC is divided more and more. This works best for high frequencies, as the capacitor size gets out of hand for low frequencies. As an example, for a 10R and 100uF, the filter is ineffective below F = 1/ (2*pi*R*C) = 159hz. That's no help for 50/60Hz hum or for 100/120Hz rectification ripple. It's also not useful for between-units filtering of audio leakage below about middle E on a guitar. But hiss and high frequencies, yep, it's good. RF circuits put a ceramic cap to ground on every stage to suck RF out of the power supplies locally to keep interference down. At RF, the caps are small and easily cover the whole band of interest.
2. Power supply impedance reduction. Some circuits are sensitive to the apparent impedance of their power supply. A local (i.e., not two inches of traces away) capacitor holds a local "bucket of electrons" for the circuit to use while it operates. Good for circuts that have sudden, sharp edges, and not a bad thing for any circuit.

The problem is, just as boner noted, each use of an R and a C costs both components and board space. The trick is to know what you're trying to achieve, do the math to find out that you are (or at least could be) doing it, and spend your parts wisely.

[PRR]

> the habit of putting a 10 ohm resistor and 100uF to 200uF cap on every transistors high rail.

Why ? ? ?

1) there's crap coming in the power line

2) the power source is a high impedance so the gain-stage can't treat it as "solid"

3) there's sneak-back through the power rail from large-signal stages to small-signal stages

Logically, if you have real crap coming in the power line, do a good cleaning ASAP, and further cleaning before any low-level stage.

Sneak-back: since all simple gain stages are inverting, sneak-back around two stages is generally benign (won't howl). So you can usually put two sequential stages on one decoupling node.

We now normally filter rails with cheap electrolytic caps. Sneak-back is also about the bass response in the forward path compared to the bass loss through the sneak-back path. Guitar amp gain stages usually do not have deep bass. And e-Caps are very cheap. In many-many cases, ONE 100uFd on the rail will swamp the sneak-back in the power line to many sequential stages.

And FWIW, I have almost *never* seen a report of a pedal "motorboating", the classic sign of insufficient power rail damping. Pedals squeal supersonically because high frequency signals leak through the air, and you can't keep output far-enough away from input, and pedal gains are insanely high.

RDH4 (70+MB PDF) section 12.10.ii suggests your "10 Ohm" Rd really can be 20% of your "plate resistor", 1.8K in your linked example. This suggests 360 Ohms, not 10 Ohms. RDH does say go high when supply is ample for the stage and go low when headroom is tight... so say 10%, 180 Ohms.

RDH suggests Rd*C can be "10,000 * 10^-6", or 0.01 Seconds (16Hz). If your C is in uFd, then 10,000/180 gives 55uFd. In Hi-Fi with extended bass gain, you may need a bigger cap. In pedalwork, 55uFd may be conservative, but 100uFd is cheap enough.

BTW, note that your 10r+100uFd has a corner at 160Hz. So it does near nothing for typical 120Hz wall-buzz or for 80Hz response in a chain of gain stages. The drop in 10r for that JFET scheme is a puny 0.020V--- you can afford more drop and more drop will give the capacitor much more leverage effect. 180 Ohms is still only 0.4V drop but 18X the effect, down to 9Hz which is far below wall-buzz or g-amp bass gain.