Touch sensitivity

[cminor9]

What makes a guitar amp touch sensitive? By this I mean when you roll off the guitar's volume or play lightly, the sound cleans up, and when you turn up the volume or play harder it gets dirty?

I was a/b testing my deluxes, tweed and blackface. I noticed the the tweed is MUCH more responsive and sensitive. If I turn both amps about halfway up, the tweed sounds nice and clean when I turn my guitar's volume down to about 4. The blackface just sounds thin and gets quiet until I really turn the guitar up to about 8. Then it goes from cleanish to dirtier on that last little bit of the guitars vol knob.

I checked out the schematic and the main difference in the preamp is the tone stack. The tweed goes from anode to a coupling cap to the volume control. The blackface goes from anode to the stack to the vol knob. Is the tone stack responsible for this lack of responsiveness? or is this just the cumulative effect of the overall different design?

[R.G.]

This is one of those places where I know what is happening with absolutely no previous knowledge of exactly how it works in the real world.

For something to be touch sensitive, it has to have a graded response to more signal, since "touch" generally equates to signal size after the forum discussions have come and gone. That is to say, it has to have a whole range of signal levels where increasing the signal level changes the distortion in some fashion, since distortion is generally what people have in mind, consciously or not, when they say "tone".

I once spent a weekend where I had to be camped out in a motel in a remote city waiting for some technical work that was to be done by Monday. I was good an tired of eating out and whatever else you do in a remote town, and have such dull taste in "fun" that I decided to just spend the time messing with thinking about distortion.

People think this or that distorts in too ugly a fashion if the distortion onset is abrupt. It's perceived to be better sounding if it begins as a little distortion and gets more distorted as you hit it harder. Every known distortion device has at least one "knee" where it changes from mostly linear and non-distorting to nonlinear in some way. The size of that region where the change from linear to distortion is what I call the knee; a region where the distortion is increasing toward some maximum value.

High feedback opamps and triodes fed from high source impedances can clip at some razor sharp level. That is for signals even 1millivolt under the clipping level, there is almost no distortion, and 1mV over the signal flat-tops.

Silicon diodes have a knee of about 0.2V, starting at about 0.45V where they conduct a little harder and being in full-on clipping at about 0.65V. So for the range of signal at the diode with peaks at 0.45V to 0.65V, silicon diodes act touch sensitive. Whether or not this comes out as sounding touch sensitive depends on how much of the whole signal cycle is encompassed inside that knee. Feed a silicon diode clipping pair 0.4V peak signals and it will not have perceptable distortion at all. Feed it a signal with peaks always less than 0.65V, and it will spend the top third of the signal range in the knee, and sound touch sensitive (that is, the tone changes with the signal level) within that top third.

Germanium diodes start conducting at about 0.1V and are in full conduction by 0.3V. so signals between 0 and 0.3V spend 2/3 of their excursions inside the increasing distortion zone, and indeed we find that germanium diode clippers are perceived to have more soul than silicon.

Notice that if you feed either a silicon or germanium diode pair a 10V signal, the amount of signal transition through the increasing-distortion range is about nil, and all you get is square waves in either case. So what matters is the amount of the total maximum signal level that is spent within the increasing-distortion range. Inside that range, bigger or smaller signals produce changing distortion tone. Push it to far outside the knee and no change in tone results. It loses touch sensitivity.

So how does that matter in tube amps? Good question. I would look for the range of signals over which a 12AX7 stage has a smoothly-increasing distortion level and deliberately run the amp so that the signal ran inside that range for most signals. Exactly how to do that?

Sorry. I haven't gotten that far yet.

[jrfrond]

I can tell you EXACTLY what makes this fly (in my opinion): A relatively short signal path with a minimum of gain shunting.

For instance, a 5E3 is essentially a straight line with gain and no negative feedback, with only a couple of gain stges between the input and speaker jacks, with a low component count. No gain lost through NFB, nor dropped/recovered through the tone stack. There are also other factors involved, e.g. rectifier type, carbon comp resistors and speaker choices, but for the most part, these are the factors that make an amp touch sensitive.

I've been building an amp design for several years that is not exactly rocket-science, but is extremely touch-sensitive and you can hear all the nuances of the guitar at the speaker. it consists of a pentode preamp, LTP PI and power amp. Simple, right? Yes, and the ONLY tone control on the amp is a brilliance control across the power tube grids, as a high-cut control. When an amp sounds great on it's own, you don't need a lot of knobs to twiddle to make it sound good.

[Chuck H]

I agree with all the above posts. And will add... When you have a tonestack between V1a and V1b your not getting any clipping at all from V1b. If you have only a coupling cap and load your probably able to clip V1b a little. Now... Most amps graduate enough so that with a typical guitar pickup the power tubes are the first to clip. Followed by the PI and then the last preamp stage. In modern uber gainers all bets are off. So, it just happens that in a 1-ts-2-PI amp there are fewer clipping stages. In a 1-2-ts-PI there is enough gain at the end of the tone stack to clip the PI, so you actually have the complexity of one more clipping stage. Not to mention that having all the clipping jive a little closer together seems to make the effect more dynamic. I build a 1-ts-2-3-PI amp that is VERY good about cleaning up with the guitars volume control and produces very similar volume both clean and distorted. Even for that, it's not as touch sensitive as a 5E3 which just always seems so crazy good at going from clean to hard crunch with just your fingers.

[Gtr_tech]

A well designed 3 stage cascaded preamp can be very touch sensitive as well. In fact, the design I use in all my amps does this so well that it as much an instrument as the guitar is. I use 1000p bypass caps on the guitar's vol pot to combat loading effects and this will actually cause a slight brightening as you roll down past halfway. I like this, and with a fairly hot bridge humbucker you get a very wide range of usable tones. I can roll down to 4 for a nice punchy rhythm tone, then roll up full for more balls/sustain for solos. With a neck single coil, it cleans right up when the vol is rolled down, but its not *too* clean. Theres just enough (very) subtle distortion to make it sound interesting. Swirl? Yes. Its very sensitive to picking attack so it can dirty up a bit just by digging in. SRV heaven there....plus I like the natural compression offered by the 3 stage design for cleans....evens things out.

[pdf64]

What makes a guitar amp touch sensitive?

My thinking is that sag on the screen node can play a big part in this, eg on a 5E3 the screen node drops significantly with big signals, thereby reducing the gain and output of the power tubes.
Undoubtedly plate node sag and cathode bias squish also contribute to that compression effect, but my perception is that when the 4k7 screen grid dropper is replaced by a choke (or even a 1k dropper)/ 470 ohm screen grid resistors, a 5E3 becomes a different amp, in terms of its touch sensitivity. Still good, but different.
Pete.

[cminor9]

well, thanks guys. So it sounds like while the tone stack definitely comes into play, there are a lot more factors involved.

I wonder how I could look at a circuit and predict how the amp would respond? Are there any other amps that you can think of that share the same responsiveness of a 5E3? I could study the circuits and see if I can find any correlations. Like the 5F4/5E5/5E7?

[exclamationmark]

Something that can help is sticking a pentode or 12au7 (something with low miller capacitance) at the input stage, so it retains a lot of treble from the outset. I feel a lot of people associate the sparkle when they roll down their volume control with touch sensitivity.

[Merlinb]

I checked out the schematic and the main difference in the preamp is the tone stack. The tweed goes from anode to a coupling cap to the volume control. The blackface goes from anode to the stack to the vol knob. Is the tone stack responsible for this lack of responsiveness? or is this just the cumulative effect of the overall different design?

There are quite a few more difference than that! The blackface has no cathode follower, the other two have them, but in different positions (assuming you're talking about the same schems I am.... Then you have LTP versus cathodyne, different bias points etc... It's a overall design. Operating point players perhaps the biggest part, as it controls the DC component of the distorted signal, which causes dynamic (e.g., time and level dependant) compression, and changes in distortion products.

[J M Fahey]

Thanks guys.
To all of you.
Incredible insights into *good* sound and design.
Nowhere else available. Period.
I LOVE this Forum.

[Chuck H]

Here, here. +++

[MWJB]

Merlin B wrote: "There are quite a few more difference than that! The blackface has no cathode follower, the other two have them, .." No Fender deluxes had a CF tone stack.

Cathode bias over fixed bias has an effect in the "touch sensitivity" stakes.

[Merlinb]

Merlin B wrote: "There are quite a few more difference than that! The blackface has no cathode follower, the other two have them, .." No Fender deluxes had a CF tone stack.

Oh yeah. For some reason I saw "Deluxe" but read it as "Blonde".
I also wrote 'players' instead of 'plays'. Must be having a dyslexic day,,,

[J M Fahey]

Oh yeah. For some reason I saw "Deluxe" but read it as "Blonde".

Don't worry.
I'm also obsessed by gorgeous Blondes too.
Also see them everywhere.
Er ........... hope didn't misunderstand you ....

[jmaf]

Are there any other amps that you can think of that share the same responsiveness of a 5E3? I could study the circuits and see if I can find any correlations. Like the 5F4/5E5/5E7?

The Trainwrecks don't have channel switching or very many controls. They're known to be very touch responsive, allowing you to just mild down the picking and get clean, then whacking the strings and getting lots of distortion. I recall spending a whole sunday looking at Youtube videos of Trainwreck amps, downloading large gut photo sets and so on.

What Trainwrecks do is something I've immitated in my experiments and which really works and has already been mentioned here - to develop distortion mildly from one stage to the next. So they use larger cathode resistors on the gain stages, biasing the 12AX7's colder and giving more headroom before positive grid region. Part of their recipe seems to be that by not distorting everything all at once, it allows small signals to go through small and when you whack the strings the gain is multiplied so you go from nearly clean to full distortion by just altering your playing and by using the guitar volume control(no channel switching).

Purported Trainwreck schems are all over the place so you can just google them if you haven't. But from gut photos I've seen around the WWW I think KFischer used to make some changes from amp to amp, but you'll probably note they have in common the cascaded and progressive gain stages, instead of high gain Peaveys, Marshalls and alikes especially those made for heavy metal which tend to distort to oblivion already on V1A.