Anti-Log Pot for DS-1

[Christoper]

I've done a few mods to my DS-1 and now I've noticed that the majority of its time the gain control is spent between past 3 o'clock for single coils and 1 o'clock for humbuckers. Below noon the gain control doesn't do much

Am I right in thinking that if I switch the 100k linear pot in there now for a 100k anti-log it will bring the taper closer to what I want, or will it be too drastic a change?

What's the math behind the linear pot being more responsive at the top end of its travel?

[Mark Hammer]

The Gain/Distortion control on a DS-1 provides two complementary actions.  Gain in a non-inverting op-amp is dictated by two resistance: the resistance in the feedback path, that dictates negative feedback from the op-amp output, and the resistance of a ground path.  Those two resistances function like a voltage divider (i.e., volume control).  The smaller the resistance of the ground path, relative to the feedback resistance, the more negative feedback if bled off to ground, and the more the op-amp is allowed to get closer to behaving like an "open loop".

On a DS-1, the distortion pot reduces the one resistance as it increases the other (going in both directions).  That allows for a fairly wide range of gain adjustment.  I doubt whether any particular alternate taper would achieve a more desirable adjustment of gain.  It's possible, I suppose, but I'm not especially hopeful.

[Christoper]

But I would assume that if I swapped in an antilog pot with a 90% taper, noon would have the same gain value that I have with the stock linear pot set to 90%?

I guess what I'm asking, the gain control is too sensitive at the right side of the pot and not sensitive enough on the left. Is there a way to fix this?

[Rob Strand]

But I would assume that if I swapped in an antilog pot with a 90% taper, noon would have the same gain value that I have with the stock linear pot set to 90%?

I guess what I'm asking, the gain control is too sensitive at the right side of the pot and not sensitive enough on the left. Is there a way to fix this?

Mark's comments about the double action of the pot holds.   However, to get something like what you want you need an antilog pot.    However, I suspect the antilog pot will go to far and make the gain too high at the 12:00 (setting 5) position.

With a linear pot:
position gain
7:00   1.0
...
12:00  1.9
13:00  2.3
14:00  3.0
15:00  4.2
16:00  7.1
17:00  22

With an antilog pot at 12:00 the gain is around 4.7.  On on the high side of what you want.

There's a few simple mods you can make with the linear pot to push things closer to what you want.

A resistor in parallel with pins 2 and 3 of the Distortion pot.   
Since the bass cap is so large it doesn't affect the bass response much.
100k makes the 12:00 gain about 2.3; like 13:00 on the original
33k  makes the 12:00 gain about 3.0; like 14:00 on the original

An alternative mod is a resistor in series with pin 1 of the Distortion pot.
This doesn't affect the bass response at all but it increases the gain at all positions.
The gain at zero (7:00) isn't going to be a big deal.  But the gain at maximum might
not be desirable.   A value of 22k with make the 12:00 position like 13:00 on the original.
Best to do this mod in small doses.

[Christoper]

Okay this is helping me out quite a bit. The two most pertinent mods I made (which I guess I should have mentioned earlier) are the transistor booster stage, where I brought the gain down from 56 dB to 27 dB. This is still enough to get some distortion even with the gain at zero. I also changed the stock +/- 0.7v clipping voltage for the diodes to +/- 0.9v.

The take away here is that even with moderately high output humbuckers, anything to the left of noon might as well be minimum gain. So I have no qualms with losing adjustability between 0 ohm and 50k. Right now I feel that I want what is currently 15:00 to be the new noon.

I think the easiest mod to do would be to use a resistor to adjust the taper. I did a similar thing to the Filter control on my Rat2. I know putting a resistor in parallel effects the value of the pot as well as the taper. How would this effect a DS-1 given that the pot is providing two resistance values instead of one? IE, would adding a resistor in parallel have any effects other than adjusting the taper?

[Rob Strand]

I think the easiest mod to do would be to use a resistor to adjust the taper. I did a similar thing to the Filter control on my Rat2. I know putting a resistor in parallel effects the value of the pot as well as the taper. How would this effect a DS-1 given that the pot is providing two resistance values instead of one? IE, would adding a resistor in parallel have any effects other than adjusting the taper?

Let's back up a step.  With cases like this there are some very fine points about what things are the same and what things are changing.

As you have it now with the 100k linear pot unchanged:  The reduction of the gain at the input means you have to advance the gain/distortion control to compensate.   The thing is, when you advance the gain the pot resistance between pot pins 1 and 2 now needs to be higher.   With a higher resistance the low-pass filter formed by the pot (pins 1 and 2) and the 250pF feedback cap on the opamp now has a lower cut-off frequency.   So even though it seems like not changing the pot doesn't change anything in reality the behaviour of the DS-1 does change because the gain/distortion pot *setting* has changed.   So we can easily get caught up in a fine detailed debate about the correct low-pass frequency we are aiming for.   Similarly the different pot setting changes resistance on pot pins 2 and 3 and that changes the filter high-pass filter cutoff (set by the 1uF cap to ground on the opamp).   The cap value is quite high so the high-pass cut-off is less of a concern.

Lets say if we add resistor on the pot we don't want the cut-off to be affected any more than what you have now by advancing the 100k linear pot to 3:00.

Just to be clear, in these examples I'm doing both mods.  Each contributes a bit.

To get approximately the same gain  and HF roll-off at noon as you have at 3:00 now:
- add 39k in parallel with pins 2 and 3 of the gain/distortion pot
- add 33k in series with pin 1

Slightly better with less change in the HF roll-off than you have now,
- add 33k in parallel with pins 2 and 3 of the gain/distortion pot
- add 22k in series with pin 1

[ElectricDruid]

Why not tweak the transistor mod to put a little bit of the missing gain back? It sounds to me like that mod isn't entirely to your taste and it's left you having to blast the Gain control to compensate.
Cutting from +56dB to +27dB of gain is a big deal - that's practically "x30" for your signal gone. Maybe put a bit of that back (maybe just 10 or 12dB) and see if that gives you a better response on the Gain control?

[Christoper]

I suppose this is another point of discussion I don't understand. ElectroSmash is quoted as saying:

The transistor stage is a Common Emitter amplifier, the approx. voltage gain in this topology is collector resistance divided by the emitter resistance (RC/RE or R8/R9 = 10K/22 = 454 = 53dB), but the effect of the feedback resistor and capacitor (R7 and C4) will lower this gain to 35dB  have to be taken into consideration.

How does he calculate the effects of R7 and C4 on the total gain? I planned to socket R5 and R9 anyway. If I bump my current R9 of 390r down to 82r and R5 of 56k down to 47k, it should give me a 41 dB boost from that stage, before the other considerations. I think this still should be low enough gain from this stage to where I don't get any clipping from this stage.

At this point I'm happy with the sounds I can get out of the pedal, I just want a more useable sweep on the gain control; so if this mod would change the sound too much, I'd prefer to not do it. But now that I think about it, this is probably the answer.

[antonis]

The transistor stage is a Common Emitter amplifier, the approx. voltage gain in this topology is collector resistance divided by the emitter resistance (RC/RE or R8/R9 = 10K/22 = 454 = 53dB), but the effect of the feedback resistor and capacitor (R7 and C4) will lower this gain to 35dB  have to be taken into consideration.

How does he calculate the effects of R7 and C4 on the total gain?

Consider R7 // X_C4 as a single feedback resistor (frequency dependent)..
(where X_C4=1/2π*f*C)

Actually, Q2 stage gain could be calculated via NFB gain general formula, A_CL = A_OL / (1 + β*A_OL), where A_OL = (R8//R10) / (R9+0.025/Q2I_Collector) and β = (0.025/Q1I_Collector) / (R7//X_C4)

The higher the open-loop gain (A_OL) the more close to 1/β the closed-loop gain (A_CL)..

P.S.
ElectroSmash Booster stage gain analysis is misleading (like some others by them..) 'cause for such a low Emitter resistor value, intrinsic Emitter resistance (0.025/I_Collector) should be seriously taken into account..
(its value is considered in series with Emitter thus dominating actual open-loop gain..)

[Christoper]

That makes it a little bit easier to understand. I think I'm just going to throw a trim pot in for the Q2 RC and RE. I just looked at my notes and apparently I already tried 82R and 47K. Unfortunately I didn't leave any note as to why I reverted it back to what I have now.

[antonis]

Rc value sets Collector voltage in conjunction with Collector current which is set by Emitter resistor value and Emitter voltage, the later set by bias configuration so, I think, you should need something like below: 



P.S.
Bear in mind the maximum gain restriction obtained from a single stage CE amp..
(which, in your case, is much lower than "ideal" 20 times Vcc..)

[Christoper]

Okay that looks like more something I recognize from college. 
I'll need to review my notes once I get home, but I think I'm starting to understand.

I just decided to heck with it and put a trimpot in there for R9 and settled upon roughly 270R, so I just replaced it with a 270R resistor. Definitely more useable range on the gain knob now.

One last question I had, not strictly related to the matter at hand, but apparently in the olden days of DS-1 they used a 250pF capacitor in the opamp feedback loop in order to tame high frequencies. Now they use a 100pF. How noticable a change would this be? I've played DS-1s from before they made the swap and I never noticed a particular difference versus afterwards, but now that I know about it I may have to see if I can't tell. My current iteration of DS-1 is a little harsh with the tone in the higher settings, but I don't think that bothers me enough to change

[antonis]

Feedback cap forms a LPF with Distortion pot effective value (lug 1-wiper)..

By implementing LPF's corner frequency formula (you DO know it, don't you..??) you can estimate the change on highs..

[Christoper]

Oh yes, I know my formulas. But if there's one thing I've learned it's that numbers are not always replacement for practical experience. But then again, if there's a second thing I've learned, it's that my own experience means more to me than the anecdotes of people on the internet. I guess I gotta try it myself then at some point 

[antonis]

How noticable a change would this be?

Could you plz explicitly quantify your own perception..??

[ElectricDruid]

One last question I had, not strictly related to the matter at hand, but apparently in the olden days of DS-1 they used a 250pF capacitor in the opamp feedback loop in order to tame high frequencies. Now they use a 100pF. How noticable a change would this be?

Well, it's more than an octave lower, so it's definitely "significant". Whether it's "noticeable" depends on where the cutoff frequency started out. If it was at 50KHz previously, then it probably didn't make any difference you'd hear. If it was at 25KHz, then it'd come down to 10Khz or so and you'd probably notice. Any lower than that, and you'd *definitely* hear it.

[Rob Strand]

I thought the 250pF to 100pF change was a tweak to make the *overall sound* of the new opamp circuit like the old one.  They obviously spent some time tweaking around here.  FWIW, the DS1w uses a different circuit again, even though people consider standard mode the same as the original.  In fact I don't think that feedback cap is there at all although there is a compensation cap in the discrete opamp which could shape the response.

Here's the results of tweaking the BJT stage to get about 12dB more gain (sort of the 3 O'clock to 12 O'clock shift).    Both this mod and the original mod affect the bass response.   It depends if you want the mod for lower gain or for the bass response change.   You can of course tweak the bass independently of gain by adjusting the 47nF cap on the base of the transistor.

[Rob Strand]

Something I forgot to mention about the 1k mod on the BJT.  It shifts the bias point on the transistor and that can change character of the clipping of the BJT stage.   It's possible to adjust the bias point independently of the gain by adjusting the 100k on the base of the transistor.

You can spend hours or weeks doing AB comparisons with small changes.

The take home message is if you change something there can be a lot of secondary follow-on affects, which often get totally ignored by modders.

[Christoper]

The take home message is if you change something there can be a lot of secondary follow-on affects, which often get totally ignored by modders.

Yeah especially with LPFs and HPFs. Feels like half the time I change a resistor I need to change a cap to compensate and visa versa