Adjusting a variable mid hump control in a Tubescreamer?

[mordechai]

Brian Wampler provided an excellent overview in an article regarding installing a mid hump control in the Tubescreamer, it's here:

http://www.premierguitar.com/Magazine/Issue/2012/Oct/5_DIY_Mods_to_Perfect_Your_Ibanez_TS9_and_Boss_SD_1.aspx?Page=6

I would like to do this mod, but do the values of the caps or resistor need to be adjusted if I've already done a gain mod (lowered the 4.7k to 2.4k and upped the .047uf cap to .1uf)?  Or do these changes not affect the functionality of Brian's suggested mod?

[Mark Hammer]

The changes you've done would not reduce the midhump, which means the variable midscoop is still appropriate.

[mordechai]

Thanks Mark.  Brian also notes in that article that volume will drop with this mod.  I have subbed in a red LED for one of the diodes (the other is 1n914) which I think may help with that, but I wonder if changing the 1K resistor between pins 6 and 7 on the op amp (I'm using a regular JRC4558D) will help with that too.  Will lowering the value of that resistor offer a little more volume?

[gritz]

*caveat* I've not built this mod, so take my comment with a pinch of salt [/caveat mode]

It doesn't make any sense. Replacing the 1k fixed resistor with two 10k resistors forms a lowpass filter with C5 at about 36Hz and it progressively eat signal above that frequency. The 6n8 bypass cap won't do much to mitigate this. Setting the potentiometer leftwards just exposes the opamp output to a 4n7 cap to ground and capacitively loading an opamp output isn't usually a terribly desirable thing. That cap might as well not be there.

Nope, I don't get it.  

[mordechai]

I see what you're getting at.  There may be something else at work here that we're not considering.   I'll build it up and report back with whether or not it functions as intended.

[Mark Hammer]

The solution would be to replace the 1k feedback resistor in the tone control stage with a higher value to provide some gain recovery.  

Note that the tone control provides three complementary actions:

1) it adjusts how much the second .22uf cap adds to the existing filtering action of the stock .22uf/1k lowpass filter,
2) it  adjusts the gain of the tone-control stage,
3) it provides some highpass filtering (bass rolloff) of the tone control stage as its gain is increased.

Increasing the gain of that stage as bass content is lost results in approximately equal output levels at each extreme of the tone control.  Changing the gain of that stage via the feedback resistance will have no impact on the tone itself, but may have some impact on the feel of the tone control.

Stock, the tone stage has a gain that ranges from 1.05x at max treble-cut to 5.54x at max treble.  If you made the 1k feedback resistor 2k2 instead, that would range from 1.11x to 10x.  At the midpoint of the tone control (assuming it is "textbook" linear pot), the gain should be 1.1x, stock, and 1.22x with the suggested 2k2 feedback resistor.  So clearly the gain increase occurs largely as one rotates the tone control in the direction of more treble, which makes it worth considering, but not a general purpose panacea.

I've also found that choice of buffer transistors can make a difference in output level, and aim for something with hfe> 300.  That may be superstition on my part.

Gritz' comments about the filtering properties of the network make sense to me.  T'wer I, I'd adjust the component values to provide a high-end rolloff comparable to stock, and probably use a different value pot than the 100k unit suggested in the article.  So, maybe drop those 10k resistors down to 3k3 each, and make C5 a .068uf cap instead of the stock .22uf.  The 100k pot can probably be  replaced with 10-25k.  The value of the added .0068 cap may need to be changed too, but I have no immediate ideas about a suitable value.

As a final comment, by simply upping the value of the .047uf cap in the clipping stage, you'd be surprised how much that detracts from the mid-hump.  remember, the squawkiness of the TS is generally a result of what it forces you to focus your attention on.  And part of that is via the removal of the low end.  Introducing a little more low end can even out the tone quite a bit.  And since a big chunk of the guitar signal "lives in the basement", subbing a .1uf to .22uf for the stock .047uf will let you extract more grind for the same gain setting.  I recently made a TS with a Tonepad layout and used a .1uf value, along with a 4-diode combo of 2x1N4148 and 4148+1N5817, and was very pleased with the evenness of it.  Still plenty of grit should you need it, but able to produce a very smooth humpless tone.

[slacker]

To me it looks like that mod would make more sense if you stuck it after the existing 1k/220n filter, with the 220n replacing the 4n7 in the mod, I'd rip out R16 as well or replace it with something bigger to reduce the volume loss. That way with the pot turned so the wiper was touching the bottom of the 220n it would effectively be stock and turning it the other way would give a mid scoop or just flatten out the response depending on component values.

Having said that Mr Wampler probably knows what he's doing.

[mordechai]

my report:

I tried the mod following Mark's suggested values (for the .0068uF cap, I installed a socket and tried a few cap values).  It "worked", but not well, and it really stole a ton of signal resulting in virtually no available gain.  So, I took it out. 

I'm going to try something else instead.  For the cap that sets the midrange corner frequency, I will wire one pad to lug 2 of a 25K pot carrying two caps (let's say a 100nf and a 33nf) on lugs 1 and 3.  I will connect the free legs of each cap together and wire them back to the other pad on the PCB...so, just a blend between the caps.  With the resistor lowered to 3.3K, I think that sweep should be usable.  But would the resistance from the blend pot interfere with this?

[gritz]

...  But would the resistance from the blend pot interfere with this?

At anything other than either extreme of the pot's travel it's likely that it won't do much at all. There are ways to synthesize a variable capacitance, but perhaps a pragmatic solution might be to just have a switch to select different caps, from the stock 0.22 down. Even a 0.1 opens up the sound a lot and smaller caps than that only provide an incremental increase in the amount of treble - if you're using the TS to drive an amp, rather than as a distortion in it's own right (and the TS is nothing to write home about in that department imo...). Mebbe add another switch to increase the 0.047 in the feeback network a bit too, if it sounds a little "thin" - but all this depends on the characteristics of the amp(s) that you want to drive.

Perhaps the load placed on the output of the clipping opamp by the 1k / .22uF filter is part of the reason why some people claim to hear such radical differences between different opamps, so perhaps monkeying with the filter's impedance mght affect the sound in other ways. But then I've read a huge amount of anecdotal evidence about the humble Tubescreamer and have never seen a graph to back up any of these tales.