Is this a good idea for a mod? TS 68k input resistor

[Guitarfreak]

I put a 68k metal film input series resistor in a TS derivative project of mine which is just for fun.  I thought it would be a good idea because Marshall and Marshall based amps use this value and it works well enough for them.  I also modded the parallel resistor to 1M as well, to further clone the voltage divider.  The sound seems dulled a bit though from before when using the stock 10k ser and 1M par combo.  The input cap is 100nF, and as a fix I was thinking that it might be a good idea to make it smaller to see if it cleans up.  The pedal is buffered, and the input components being discussed are pre-buffer.  Should I work around the value, or just get rid of it?

[defaced]

68k is chosen in tube amps to act as a filter with the internal capacitance of the tube to reduce RF interference.  So, when you changed your input resistor from 10k to 68k, you only changed half of the equation which is why it's duller now. 

[Guitarfreak]

68k is chosen in tube amps to act as a filter with the internal capacitance of the tube to reduce RF interference.  So, when you changed your input resistor from 10k to 68k, you only changed half of the equation which is why it's duller now.  

Ok, that's an interesting point.  Of the amps I look at, I notice that they don't have input caps, though pedals do.  Does this mean that the input caps on pedals are supposed to simulate the capacitance of a tube in addition to removing DC?  Do you think that lowering the value of the input cap alone could remedy the situation?

[merlinb]

Ok, that's an interesting point.  Of the amps I look at, I notice that they don't have input caps, though pedals do.  Does this mean that the input caps on pedals are supposed to simulate the capacitance of a tube in addition to removing DC?  Do you think that lowering the value of the input cap alone could remedy the situation?

Valve amps don't need an input cap because the grid of the first valve rests at zero volts, whereas pedals need to bias things up to 4.5V.

[Guitarfreak]

Ok, that's an interesting point.  Of the amps I look at, I notice that they don't have input caps, though pedals do.  Does this mean that the input caps on pedals are supposed to simulate the capacitance of a tube in addition to removing DC?  Do you think that lowering the value of the input cap alone could remedy the situation?

Valve amps don't need an input cap because the grid of the first valve rests at zero volts, whereas pedals need to bias things up to 4.5V.

Again thank you.  I love learning this stuff.

I understand that the input cap if made smaller will have a network effect with the resistances/impedances present and roll off some low end.  What are the exact (read: approximated) calculations that I could use to calculate the impedance so that I can select an appropriate cap?  Learning the process is more important to me than finding the right value cap for this particular project.

[defaced]

I'm pretty sure it's f = 1/(2*pi*c*r), but memorizing equations was never my strong point.  I'm much more of a conceptual person.  Don't ask me about the units.  I seem to remember it being something like in ohms and micro farads or some other odd mix. 

[jaapie]

You're right, it's

Code Select Expand

f=1/(2*pi*R*C)
R is in ohms
C is in farads

make sure you've got the right order of magnitude for your capacitance; if you're using a 22nf cap, your value for C in the equation would be 22x10^-9, or 0.000000022, not 22.

[Guitarfreak]

I'm pretty sure it's f = 1/(2*pi*c*r), but memorizing equations was never my strong point.  I'm much more of a conceptual person.  Don't ask me about the units.  I seem to remember it being something like in ohms and micro farads or some other odd mix.  

You're right, it's

Code Select Expand

f=1/(2*pi*R*C)
R is in ohms
C is in farads

make sure you've got the right order of magnitude for your capacitance; if you're using a 22nf cap, your value for C in the equation would be 22x10^-9, or 0.000000022, not 22.

Thank you both.  I have been using this formula for a while, in the form of this online calculator: http://www.muzique.com/schem/filter.htm

The problem I have is that I think the R value must be a calculated impedance in this case, and not a simple resistor value or something of a more simple nature.  There is series resistance as well as a number of parallel resistors.  Should I calculate the impedance value for R by using the formula for parallel resistances Z = (R1*R2/R1+R2) ?  Does the series resistor come into play as well, being that it completes a series RC filter with the input cap at the start of the circuit?  This conversation may well get complicated very quickly and I may not be ready for the correct answer... but I am willing to educate myself one way or another as I am very interested in this type of thing.

[merlinb]

The problem I have is that I think the R value must be a calculated impedance in this case, and not a simple resistor value or something of a more simple nature.

You need the total resistance in series with the cap, that is, imagine a loop that goes from one end of the cap, to ground, back along the cable shield to the guitar, up through the pickups and along the cable again to the remaining side of the cap. There maybe many series/parallel resistances that make up this path. Usually you can ignore the pickups/guitar and concentrate on the resistance you can see in the pedal itself, as they will usually be the significant ones.

[slacker]

You're right, it's

Code Select Expand

f=1/(2*pi*R*C)
R is in ohms
C is in farads

make sure you've got the right order of magnitude for your capacitance; if you're using a 22nf cap, your value for C in the equation would be 22x10^-9, or 0.000000022, not 22.

Or you can use C in uF and R in Meg Ohms, this probably makes more sense given the values that we normally use, ie:- 1 uF = 1 and 1Meg Ohm = 1.
This then makes 22nf = 0.022 and 100k Ohms = 0.1

[Guitarfreak]

The problem I have is that I think the R value must be a calculated impedance in this case, and not a simple resistor value or something of a more simple nature.

You need the total resistance in series with the cap, that is, imagine a loop that goes from one end of the cap, to ground, back along the cable shield to the guitar, up through the pickups and along the cable again to the remaining side of the cap. There maybe many series/parallel resistances that make up this path. Usually you can ignore the pickups/guitar and concentrate on the resistance you can see in the pedal itself, as they will usually be the significant ones.

Thank you.  So you ignore anything after the coupling cap then?  Is this the way to do this for the input cap of a circuit alone, or is this the way of calculating every coupling cap throughout the design?  How do you factor in gain/buffer stages?  Do they isolate from the previous stages?

[merlinb]

Thank you.  So you ignore anything after the coupling cap then?

Um, no... stuff that comes after the cap will (may) include various resistances to ground, which are all part of the current loop.

Is this the way to do this for the input cap of a circuit alone, or is this the way of calculating every coupling cap throughout the design?
How do you factor in gain/buffer stages?  Do they isolate from the previous stages?

Every coupling cap. If there are gain stages/buffer then the output impedance of those stages is in series with the cap. Usually it will be negligible, you can concentrate on the resistances to ground.

[Fender3D]

I put a 68k ... Marshall and Marshall based amps use this value and it works well enough for them...

Just a little side note:
Marshall (and Fender) use 2 68K resistors,
When you use 1 instruments in higher jack socket you have 68k/2 (2 resistors are paralleled) in series with signal, and 1M to ground.
When you use 1 instruments in lower jack socket you have 68k in series with signal, and 68k to ground (1M res is shorted by jack's switching).

[PRR]

> in a TS derivative project of mine

This is too vague. Draw (or link) a picture. Preferably your circuit, with part values and part-numbers so as we can discuss it clearly.

[Paul Marossy]

> in a TS derivative project of mine

This is too vague. Draw (or link) a picture. Preferably your circuit, with part values and part-numbers so as we can discuss it clearly.

That's like asking a cook to give away his secrets. 

[Guitarfreak]

> in a TS derivative project of mine

This is too vague. Draw (or link) a picture. Preferably your circuit, with part values and part-numbers so as we can discuss it clearly.

That's like asking a cook to give away his secrets.  

This design is a bit of a lovechild of mine and it's been the work of about a year now.  More of a tone chase really, and it's in a constant state of liquid flux with components changing all of the time.  Each time I draw up a schem I have to scrap it and redraw it in a few weeks.  In short, I am building it within a SD-1 and it can be thought of as an SD-1 with less clipping, more gain, less clean boost, and a modified voicing.  Like I said though it's a constant battle trying to perfect that which cannot be perfected, and learn something along the way

[Paul Marossy]

That's my problem when I try to do my own circuit. I can never seem to settle on something being "good enough". 

[PRR]

I don't think the 68K and the 100nFd have much interaction, but I won't pry.

[Guitarfreak]

I don't think the 68K and the 100nFd have much interaction, but I won't pry.

I came to the same conclusion.  I wanted the input cap to cut just below 70-80Hz, and the 68k/100nF cap is somewhere around 20Hz iirc.  I was battling muddiness a few days ago so I was panicking thinking that was it.  I found the problem, but I ended up replacing the input cap with a 33nF cap anyways, and I have no problems with the sound of the unit today so... it stays!  Lets see how I feel about it tomorrow.  That's the true test!  

P.S. The original combination of 100nF cap and 10k resistor was quite tonal so I found out at a later time, which partially caused my panic when I switched to the 68k series resistor.  But it's all good now.  Everything has been worked out.

[PRR]

The usual tube guitar input is series two 68K parallel (34K) against the shunt ~~100pFd of the tube grid. This is a 47KHz high-cut. No effect on sound, but reduces radio station reception.

100nFd is more like a coupling cap. Tubes don't need that; many non-tube inputs do. If you have 10K shunt (not series) then indeed a low-cut at about 160Hz. Appropriate for de-mudding guitar. However you say you copied the 1Meg... alone, or still with the 10K shunt?

FWIW: 10K input is heavy load on e-guitar's treble resonance. That could be the "dull".