input capacitor?

[stormbringer]

Hi all,

Does anybody put a cap at the input of the amplifier so that DC is blocked from the guitar? If not, why not?

Cheers!
John

[Diablo]

Why would there be any DC at the input jack if the amp is wired properly?

[stormbringer]

Why would there be any DC at the input jack if the amp is wired properly?

There wouldn't be. The Ci would be added safety for unforeseen problems, like a plate-grid short.

[es345]

An input cap doesn't hurt, the cost is neglegible. Just put it in, I am doing that as a default.
And think about both case of failure : it prevents DC from inside out as you mentioned and from outside in - e.g. in case of a defective cap in an effectbox in front of the amp.

[stormbringer]

An input cap doesn't hurt, the cost is neglegible. Just put it in, I am doing that as a default.
And think about both case of failure : it prevents DC from inside out as you mentioned and from outside in - e.g. in case of a defective cap in an effectbox in front of the amp.

Do you solder your Ci caps to the input jack itself?

[TransLucid]

don't forget..

...that the cap and the pickup inductance will form a series resonant circuit which, depending on the component values, could cause a peak in response in the audio range.

[es345]

Do you solder your Ci caps to the input jack itself?

I normally do P2P, shortest connection from Input jack to the grid of the first tube. An example is attached. The ground for the input stage is isolated from the chassis driectly soldered to the Ground oif the first triode. Overall is a starground concept applied.

[es345]

...that the cap and the pickup inductance will form a series resonant circuit which, depending on the component values, could cause a peak in response in the audio range.

Well, but don't forget, that in series you have the 1M grid resistor in series to that LC. There will be no peak.

[exclamationmark]

A guitar plugged straight into an amp does form an RLC circuit... however, the R component is typically in the order of 100's of K's, which immediately flattens any peak present.

I've never actually heard of a grid to plate shorts happening in real life, though. Instead of a capacitor, you could stick a suitable voltage zener from grid to ground (which can also help shunt RF to ground, depending on it's capacitance). The only times you really need a capacitor is when you bias the input stage either really hot (which will cause the stage to draw grid current), or use grid leak bias. Usually if you do any of these two, there will be a tiny DC voltage on the grid, which can cause scratchy volume pots.

[TransLucid]

true. but that 1M resistor is in parallel with the Miller capacitance...

[es345]

... and what is your conclusion? What is the resulting effect? What do you want to tell?

[exclamationmark]

That's actually an interesting point. The cable capacitance, pickup inductance and miller effect probably form some sort of resonant circuit. However, since the reactance of the input capacitor would probably be absurdly small in comparison to the cable capacitance/miller effect, it would just act as if it wasn't there at all, not affecting the resonant peak in the slightest (or forming one for that matter).

[Steve Conner]

I've seen many tubes that leaked enough DC back through their grids to make the guitar volume pot scratchy. The input capacitor stops this, and doesn't have any sonic drawbacks that I know of.

[J M Fahey]

True.
And also consider that some Musicians use Pedals (ever heard? ), which often have indeterminate quality or output electrolytics.
Any slight loss, *through a 1M resistor* can throw bias out of whack.

[TransLucid]

Let's assume that the pickup inductance is 5H, and the combined Miller/cable capacitance is 100pf. By my calculation that puts the resonant peak at 7 kHz. Now, if you add a .01 ufd input cap, the resonant peak is lowered to 700 Hz. I can imagine that having an audible effect.

However, .01 ufd is a small value for an input capacitor, so as long as the value is large the resonant peak will be at a low enough frequency so as to not be audible. Note that even .1 ufd gives a resonant frequency over 200 Hz...