Categories
Breadboarding Overdrive

Minimal and Scaled Tube Screamer Version

Let’s continue covering arguably the most famous overdrive pedal out there – Tube Screamer. This post continues building on previous posts: breadboarding the pedal and how it shapes the tone.

In this post I’ll try getting Tube Screamer sound with completely different set of components. Well, different values of the components – it is still exactly the same topology of the circuit.

I’ll cover quickly the methodology, then propose the alternative circuit, and then do a quick analysis of it. I’ll breadboard it (there’s also table with bill of material) and then compare the sound of it to my TS9 pedal.

If all goes well, we get a very close sounding circuit. While looking like a possibly pointless exercise, it is important to try things out ?. This will either confirm or disprove the points I made in the previous post. Fingers crossed I don’t embarrass myself ?.

If you have no patience to read all of this – just watch the video straight away.

Methodology

How to approach this? Let’s say I want to use 10K pots everywhere for some reason. How do I preserve all the filter cut-off frequencies for example?

Let’s do this exercise on the overdrive pot:

Section of Tube Screamer schematic with Overdrive potentiometer
Original Clipping Section

First of all – I want to preserve maximum gain of 41dB. If I replace 500K pot with 10K pot, I need to scale R4 to preserve the gain. The new pot has 50 times lower resistance, so I need to use about 50 times smaller R4 resistor.

If I use 100 ohm resistor, and I change R6 from 51K to 1K I get the gain range from 20 to 41dB. That’s what I want. So I effectively scaled down all resistors 50 times. (I don’t need to touch any other resistors mind you, since they are not affected by this).

Now, in order to preserve 720ish Hz cut-off frequency, I need to do something with C3. The theory is – if I scale down my resistor, I need to scale up the capacitor. If I do this 50 times, I get – 47nF*50 is about: 2.35uF, let’s call it 2.2uF (for a standard value). Will this work?

Here’s what the calculator says:

Diagram showing frequency response of the original component values for the Tube Screamer
Original Frequency Response

Note the scaled version:

Diagram showing frequency response of the scaled component values for the Tube Screamer alternative
Scaled Component Values Frequency Response

That is pretty close, so at least in theory, this should work! I need to actually scale C4 as well, since the high pass filter using it is now affected.

After that, I just need to go and scale all the other potentiometers in a similar fashion. This will mean that I’ll probably need to adjust lots of components, but I’ll be just following this methodology and I should be more or less fine.

Alternative Circuit

I had to scale down Tone pot 2 times and Level pot 10 times. This is the alternative circuit I came up with:

Schematic showing alternative version of the Tube Screamer circuit with scaled component values.
My alternative version of the tube screamer (Click for the full picture)

Note on the designator values. I tried keeping the component designators in such a way that I can relate them between the “original” and my alternative version. So if you see component R101 in the diagram – the component is equivalent of R1 in the original schematic, R116 is equivalent to R16 etc.

Just so I keep the track what I’ve changed and what were the original values compared to the new – alternative values. And when I start writing about different pedals using the same circuit topology, I’ll do similar thing, I’ll just identify them as R201, R301 etc.

Analysis

How close did I get the scaling?

Well, if I add all the relevant filter cut-off frequencies here’s what I get:

Minimal schematic for Tube Screamer with all sound shaping sections annotated
Minimal & Scaled – with filters annotated (click for full image)

Just a quick glance and all looks pretty good. Let’s have a closer look.

Clipping Section

Not much to say here for the clipping section:

Clipping Section of the circuit only with notes on tone shaping sections
Focus on Clipping Section

Minimal differences, and very unlikely that they’ll cause any noticeable difference in sound.

Tone Section

Let’s have a look at the tone section:

Schematic of the tone section with all filters and gain section annotated
Focus on Tone Section

The same story here – pretty much everything the same. I did put a warning there at the filter formed by output coupling capacitor – C107 and the Level pot (RV103). I also decided to use linear pot instead of audio pot. This means that for only 5-10% of the rotation there will be some effect on the audible spectrum. Just something to test out before committing.

I’m in two minds about Level pot being linear or audio. Boss pedals, for example, use linear pots for Level control. Also – I do have to note that I’m using linear pot for Tone – the taper in Tube Screamer is G (or 4B). I’ll have slightly bit less of a control, but should not affect the sound itself.

Component Selection

When I said in clipping section that everything looks pretty much the same, that isn’t really true. What is not normally seen on any of the schematics is the component selection.

I’m not sure what exact capacitor is used in Tube Screamer for C103, but I expect it does not have tighter than ±10% tolerance. I’m using 1% tolerance resistors, but I expect that original TS808 used something like 5% tolerance. I’m just speculating, the tolerances could be even worse.

The whole point I’m getting to is that I’m using electrolytic capacitor which has ±20% tolerance. So 720Hz cut-off really translates to 600-900Hz range for ±20% tolerance if I’m not mistaken. ±10% cap tolerance gives 650-800Hz. This might actually make some audible difference if the difference is at the extreme ends. Just saying.

Frequency Response Diagrams

Let’s see how the frequency response looks like for the two versions of the circuit. It’s dangerous to infer anything about sound from this, but if they at least look alike I’m on the right track.

Here’s how frequency response looks like while sweeping Overdrive knob:

Overdrive sweep frequency response

The above diagrams look very much alike. Very little difference – which would indicate that the math and logic behind is correct.

If I look at 3 different Overdrive levels and do the Tone sweep, let’s how does this looks like:

Three Overdrive levels with Tone Sweep

Again, looking pretty similar, I’ll need to test the circuits, but looks like I’m not far off at any rate.

Spice

In case you want to check this out in SPICE, here’s the simulation diagrams:

Breadboarding It

Let’s do the breadboarding of this, here’s a diagram I came up with:

Breadboarding diagram for minimal and scaled version of TS808
Minimal and Scaled version of TS808 (click for full image)

I used pots with PCB pins so I just plugged them into the breadboard.

Here’s the DIY file:

Bill of Material

Here’s the list of components I used for breadboarding:

DesignatorsComponentNotes
BreadboardAny breadboard will do, I used full size breadboard. You probably don’t want anything smaller than that.
Jumpers and wiresAs many as you need. I got some online, but 24 AWG solid core wire will do just fine.
9V batteryI used 9V battery snap with Dupont wire.
DPDT push buttonI used ALPS SPPH410100 latching push button for bypassing the effect – totally optional.
INSwitchcraft 12BIn – stereo jack (mono will do too)
OUTSwitchcraft J111Out mono Jack
RV101, RV102, RV103Alpha 16mm 10K linear potAny 10K linear pot will do, I used 16mm with PCB pins – for Overdrive, Tone and Level control
C10110n (0.01uF) Metal film PET cap
C1032.2uFElectrolytic cap 25V
C1042.2n (2200pF)Metal film PET cap
C105, C106470n (0.47uF)Metal film PET cap
C1071uF Electrolytic cap 25V
C110100uF Electrolytic cap 25V
C11147uF Electrolytic cap 25V
C112100n (0.1uF)Ceramic X7R
D101, D1021N4148Small signal diode for clipping
R101, R1061K1% 250mW metal film
R104, R108, R111100 ohm
R116, R11710K
R105, R110, R1182.2M
R107, R109470 ohm
U101TL072op-amp
List of components for Tube Screamer (scaled version) breadboarding

Trying it Out and Final Thoughts

Here’s the video of me trying it out:

Alternative Tube Screamer Circuit

I’m sure some keen ear has heard a difference in sound. The question is how much in a band situation you could hear it, and if you’re not just imagining it in the first place ?. At any rate – I unilaterally declare it a Tube Screamer.

There seems to be some faint, high pitched oscillation noise heard around 16:15 mark in the video. I only noticed it when editing, I did not notice it at all while recording. I’ll investigate and report back at some stage.

9 replies on “Minimal and Scaled Tube Screamer Version”

I’m not sure if you’re aware, but by scaling the circuit impedance down you’ve changed the noise characteristics in a favorable manner because the thermal noise from the resistors is reduced. I’d recommend trying a better opamp like OPA2210 to take full advantage of the reduced voltage noise. You should be able to safely omit C107 with this opamp as well due to its low offset voltage.

Hey Michael,
Great point, I may have mentioned lower noise in the video, but can’t remember now. I haven’t looked into OPA2210, just went through datasheet briefly, but I had used similarly less noisy OPA2134. I don’t remember hearing huge difference to be honest. While those noise values are clearly better than for TL072, I’m really not sure how that translates to the real life.
You would still need C107 – reference value for the op-amp is 4.5V, so that would be DC component on the output without C107.

Right, I wasn’t thinking of the single-ended supply with regard to C107. I’ve been designing with bipolar supplies lately.

You may or may not notice a difference with OPA2210 vs OPA2134. It has >8x better voltage noise performance than OPA2134 in most of the audio band, although the current noise is worse, so it’s best to simulate your specific use case. In all the common audio circuits I’ve simulated or tested them in OPA2210 tends to outperform OPA2134 by a substantial margin though, and it’s also much cheaper and has lower Iq, so I’ve stopped using OPA2134 entirely. You do have to make a lot of circuit optimizations to actually achieve the low noise performance though, much like you’ve done here, as OPA2210 is almost entirely limited by the thermal noise of other components. In general I would recommend at least using OPA2210 for the input stage and the first gain stage, as 107dB SNR is achievable for a 100mV input signal in the noninverting 1M input impedance stage I use, compared to 99dB for TL072 or 102dB for OPA2134 in the same circuit. You won’t realize much noise benefit at all over TL072 unless OPA2134 or OPA2210 is the first thing in your signal chain because the damage is already done if you add the noise before buffering and amplification.

It’s worth mentioning too that the cost savings can be quite substantial for some circuits if they’re converted to be DC-coupled, at least internally, to take advantage of the 100x better DC performance of OPA2210 over TL072 or OPA2134, and it’s very attractive for headphone amplifiers and reverb drivers because it has about twice the drive capability of OPA2134 after factoring in the rail-to-rail output. OPA1637 is also very interesting for drive purposes if you want to make a bridge-tied load. Anyway I know I’m rambling a bit but I just thought I’d mention these things because most people don’t seem to know about these newer opamps and I’ve done a lot of digging for affordable low-noise performance. Best wishes from the states!

Interesting comment 🙂 I’ll look into a few things you mention there.
Realistically, both OPA2210 and OPA1637 are only available as SMD, which is the case with most of the new op-amps, and that will always make them less interesting in DIY projects.

What happen if we omit R110?

Nice post! Learned so much from this.

Btw, any chance for power amp circuit like TDA2030 with aux in and headphone out? Have some preamp-ish circuit but didnt have power amp schemo. Was used something like mooer power/ ehx magnum. Hahaha. I want build power amp for some friend as gift. Although i know they have multifx like helix, nux mg, etc. But, gifting analog circuit always make me happy because im solder junkie, addicted to fumes. Eh.

Hey Schenkerian,
If you omit R110 – nothing will probably happen and the pedal will sound pretty much the same. Output of the previous stage is already biased, so should be OK. I can’t remember now which one of the other TS like pedals doesn’t have it. Should not affect sound either. If you breadboard it you can try it out and see :))

I don’t have TDA2030, I have LM1875 laying around somewhere. I’ll do something with it at some stage, but that could be months away. Last time I tried LM1875 I had a debris from my stripboard causing a short somewhere. That thing went soooo hot in matter of seconds :)))) I found the short later and just scraped it off, but nearly burnt the whole thing :)))) What I learnt though is that it requires proper layout planning to get decent sound. It’ll be a while before I get back to it.

Hmmm, I would be careful though when looking at those pedal amplifiers. EHX magnum has Class-D amplifier which is way more efficient than Class-AB like TDA2030 or LM1875.

This is what datasheet says: The high efficiency, ~92%, of the TPA3106D1 (apparently used in magnum) eliminates the need for an external heat sink when playing music. LM1875 requires a hefty heatsink and proper air circulation I would say. Just something to be aware of.

I hear you about gifting analog circuits 🙂 I love doing that too.

Got chance to build your scaled version for my friend. Look like this. My friend known to be masochist. 1590B top jack, minimal wire. XD

Sound good. Push my mesa emulation wannabe. Hehe.
https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiG9O19b6UyKhHvTNEiWiC_UzTVeH2vv241yHeEP5FzQ3QNntYRYp4BJulvK_ngqr4icQ77TuQfIHahvUYI0qJW3_MXE_m3rJoGDC7xZtf56NSL63hrcfkkLU-_h2KhDC0hIwIUhtYFH0zAh6lHw5CXlVoNG8ZDW2yLGQJ6lePBF69kAC2x-08YZXT6MWI/s1600/scaled-TS.png

Btw, playing with spice. Why always error if use Log Pot?

And request. How to simulate impedance (IN and OUT) in spice.

Thanks.

Febri. DIY Junkie.

Hey Febri, that looks great, I’d love to hear the sound of it :))
For log pot, are you using my library? Are you using LTSpice? I had a comment that one of the commands in that library is causing issues in PSpice and other simulators. I have a fix, just need to update the files. Darn, I have to update all files with SPICE 🙁
For IN and OUT simulation, let me have a look. I do it somewhere in my SPICE files 🙂 but can’t look for it right now. I’ll reply again with more details, I promise :)))

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.