Adding headroom to a tubescremar (TS-9) (changes in power supply)

[Yoshi]

Hello, I´m new in the stombox world. And incountered a big question, Can I increase the Headroom of a tubescreamer by using the powert supply dising that exist´s in the Klon Centaur?.

I notesed that the Klon Centaur uses a +18V and -9V in some of their op-amps, and I was bondering if I can apply the same power supply desing without burning the protoboard or destroying my guitar.

If it´s a "YES" thats cool, But if is a big "No", I would like to know why and how can anyone add more headroom in the clasic JRC4558.

Thanks, and again I´m new in this word of electronics. All the best.

[GibsonGM]

Hi Yoshi, welcome to the forum.   

Your thinking is on the right track. For simplicity's sake, I would first check what the electrolytic caps in the tube screamer are rated for, their voltage rating. You will not want to exceed that, and will want to stay 1 or 2 V below them for safety.

They may well be 16V rated, so that would be the absolute max you would be able to use.

Then, the easier way to obtain a higher voltage may be to use a 12V supply, or work out something between 9 and the max rating.   The 4558 can handle 18V, so a simple charge pump can work.  Those caps, though...the literal PS from the Klon is probably not a good candidate, IMO, but something similar based on a MAX1044 or what have you.  (internet search, "MAX1044 voltage doubler").    You can also find a lot of info on running on 18V (etc) on the forum search, or net search...

If this is an already-made TS, it may be difficult to 'retro fit' it, however.  No space inside!   A new build is probably the best way to go and a great learning experience, unless it's determined you CAN run at 18V and use an adapter to just do so.   

[antonis]

Better imply +18/0 V for not having to deal with various grounds..

So, a "YES" under conditions..

P.S.
Be sure that all electro cap are rated at 25V or higher ..
(actually, the only one needed for 25V is power supply reservoir but let them all be ..)

edit1: Sir MIke is by far quicker.. 

edit2: Welcome..

[Mark Hammer]

I am of the view that increases to supply voltage will improve headroom for some things but not for others.

Consider that the diodes in a Tube Screamer are in the feedback loop of the op-amp, and will begin to conduct when the output reaches around +/-600mv  The actual supply voltage will not change that.  IF there were no diodes, then yes, the maximum voltage swing of the op-amp could be increased by an increase in supply voltage.  But very very little in the supply is going to change the headroom of the circuit, since the limiting factor is the forward voltage of the diodes, and not the supply.

NOW...having said that, if one replaced those simple silicon diodes with LEDS (or several diodes in series, in each "direction"), an increase to the power supply might improve the headroom, and dynamic range of the circuit.  Would it be an improvement?  Not necessarily.  It would certainly be louder, which some might consider an improvement.

But many of the TS-derived pedals that have found favour among players have used a 2+2 feedback diode arrangement, rather than 1+1.  This gives a little more headroom and dynamics, without any change in the supply voltage.  Even a 2+2 diode complement still keeps the signal well below the maximum voltage swing of the chip, using a +9V supply.

[marcelomd]

Hi,

Check the JHS Moonshine, both V1 and V2. They are a TS808 with a 18V charge pump, modified gain range and diode options.

https://www.pedalpcb.com/docs/BootlegOD.pdf
This is the V2 schematic, which includes a clean blend.

[Mark Hammer]

Note that some people find the component values on the output buffer of the TS-808 provide a better tone.  These were changed to different values on the TS-9.

Why?

Well, remember that the Ibanez pedal uses electronic switching for bypass, with the JFET that disables the overdrive inserted before the output buffer.  That means that the output buffer affects not only the overdrive tone, but the bypass tone as well.  The manner in which the output buffer values loaded down the overdrive tone in the 808 was considered desirable, since it "smoother out" the resulting sound.  But those same values would suck tone from the clean signal in bypass mode.

Most of us here will likely make or use a TS-type circuit with a 3PDT stompswitch for true bypass.  Since that allows the clean/bypass signal to not be affected by the output buffer, that gives us the liberty to use the 808 buffer values and benefit from the beneficial effects of that loading.

[Vivek]

Normally "headroom" is used to mean that excess capability over normally expected signal levels where the output still won't get too distorted.

Suppose you expect 2vpeak signals at an output but can handle 5Vp signals with acceptable distortion, you have 3V of headroom


But for a pedal that is meant to create distortion, how do you define headroom ?

[antonis]

But very very little in the supply is going to change the headroom of the circuit, since the limiting factor is the forward voltage of the diodes, and not the supply.

Now is the right time to argue with Mark..!!

Of course max clipping headroom is determined by clipping diode(s) forward voltage drop but there are additional conditions like slew rate (Vivek, plz let it pass as it's written..) or shape of signal which should be taken into account..

[Mark Hammer]

Slew rate CAN matter...when the bandwidth of the signal  is wide, the signal complex, and the degree of amplification required substantial...but it doesn't always matter.  If I was designing a D/A converter, or a mixing board, I'd likely want to use op-amps with a decent slew rate.  But for guitar, and especially distorted guitar,  High slew rate won't add anything, and may even detract.  Keep in mind that we're talking about a signal that quickly rolls off above 6khz, by the time it hits the speakers.

We keep forgetting that neither op-amps or diodes are designed and developed for the purposes of guitar or even music, but rather for a possible range of purposes that can include signals and times wayyyyyyy beyond the range of musical instruments, and way way wayyyyyy beyond the bandwidth of guitar.  Somewhere, in the overall universe of signal processing, the difference between 5V/us and 13V/us slew rate matters a great deal.  But the reason why so much of the music products industry directed at guitar uses op-amps that don't slew any higher than 1V/us (and often less) is because higher rates don't matter.

RE: Supply voltage and headroom.  For those parts of a circuit that are NOT constrained by diodes in the feedback loop, supply voltage CAN play a role in raising headroom.  So, 18V won't really do much of anything for the clipping stage of a TS-808/9 or Boss SD-1.  It will increase the headroom of the 2nd op-amp that provides the tone-control stage, but again there will be negligible impact because that stage is not set for very much gain; only a max of about 5x at max treble, which still keeps it well within the voltage swing at +9v.

Now, consider the Distortion+, DOD250, OCD, and similar.  All of these have a gain stage whose output is clipped by diodes or similar GOING TO GROUND from the output of the chip.  At higher gain settings, these op-amps are being asked to provide signals in excess of the voltage swing possible with a +9V supply, with none of the limits that feedback diodes would provide.  The signal is actually being clipped at many times, before it even reaches the diodes, by running out of headroom.  Increasing the headroom via the supply voltage can reduce the clipping derived from headroom limits, while still permitting the clipping elements on the op-amp output (diodes, etc.) to do their work.

[ElectricDruid]

I'm with Mark on this one.

TBH I don't see much point running the Tubescreamer unmodified at 18V. The signal from the clipping stage will be the same, so you gain extra headroom for the tone control stage, but unless clipping in the tone control was a big problem you're not going to gain anything there.

Of course, you can tweak the rest of the circuit too - change the clipping, change the tone control, whatever - and maybe that makes sense if you've got headroom to use. But just "18V tubescreamer"? I don't see any benefit.

[amptramp]

The Tube Screamer uses a non-inverting op amp stage as the clipping stage and part of the unique character of the tone is that whereas the stage may be set for a gain on the order of hundreds going to the clipping diodes, there is always a gain of one for the input signal even when the diodes are clipping.  So if you add the input signal to the clipping level and multiply it by the possible 5x gain of the tone control stage, a hot input signal of a guitar that peaks at 0.4 volt added to the 0.6 volt clipping level of the diode multiplied by the 5x gain of the tone control stage gives you 1.0 x 5 = 5 volts and that will surely hit the rails on a 9 VDC supply since the output is centered on 4.5 volts and this is without considering the op amp headroom.  So yes, the Tube Screamer would change characteristics if the supply voltage was increased.

[Vivek]

The Tube Screamer uses a non-inverting op amp stage as the clipping stage and part of the unique character of the tone is that whereas the stage may be set for a gain on the order of hundreds going to the clipping diodes, there is always a gain of one for the input signal even when the diodes are clipping.

To illustrate the above

Here is part of the TS schematic:



and here are the waveforms when 1Khz signal is injected at input of TS:



Red = waveform at point Vclip

Green = waveform at point Vbuf

Blue = mathematically generated curve of Vclip - Vbuf

We see that (Red) Vclip = (Green) input signal sine wave going into Opamp clipper + (Blue) squarish waves at Vf of the clipping diodes.

[iainpunk]

my black guitar has enough output to clip red led's without amplification, (hot humbuckers in series) and using a tube screamer, it does not clip to rails according to the o'scope. tone section was set neutral, in this experiment.

the tone control doesn't matter by the way, since the frequencies it boosts at max boost, are also cut beforehand, passively, to the same degree or more than the boost.

the slew rate (1.7v/us) only affects frequencies above 90kHz in a meaningful way, if the signal is rail to rail, and the smaller the signal is, the higher that affected frequency becomes. a more important parameter is the GBP (Gain Bandwidth Product, not Good Boy Points) it affects how much high frequencies are able to be amplified, and has an audible effect in audio circuits

cheers

[ElectricDruid]

So if you add the input signal to the clipping level and multiply it by the possible 5x gain of the tone control stage, a hot input signal of a guitar that peaks at 0.4 volt added to the 0.6 volt clipping level of the diode multiplied by the 5x gain of the tone control stage gives you 1.0 x 5 = 5 volts and that will surely hit the rails on a 9 VDC supply since the output is centered on 4.5 volts and this is without considering the op amp headroom.  So yes, the Tube Screamer would change characteristics if the supply voltage was increased.

Don't agree, sorry. The tone control doesn't show this amount of gain, and even if it did, you can still get a 7Vpp signal out of a 9V op-amp before you hit clipping.

[GibsonGM]

It's something that's "done" - people ask for it...why would they do it if 'nothing changes'?   (scratching head emoji) 

(Have to guess they have heard 'nice sounding 18v TS' and didn't know they also added diodes to raise the clipping threshold/output)

[Mark Hammer]

So if you add the input signal to the clipping level and multiply it by the possible 5x gain of the tone control stage, a hot input signal of a guitar that peaks at 0.4 volt added to the 0.6 volt clipping level of the diode multiplied by the 5x gain of the tone control stage gives you 1.0 x 5 = 5 volts and that will surely hit the rails on a 9 VDC supply since the output is centered on 4.5 volts and this is without considering the op amp headroom.  So yes, the Tube Screamer would change characteristics if the supply voltage was increased.

Don't agree, sorry. The tone control doesn't show this amount of gain, and even if it did, you can still get a 7Vpp signal out of a 9V op-amp before you hit clipping.

When the tone control is set to max treble, the 1k feedback resistance, and 220R ground leg, DO provide a gain of about 5.5x.  But here's the critical element: that gain is for content above 3.3khz.  What is the signal level of content up there; especially after the lowpass filter of the 1k/220n network?  Not very high.  So multiplying that voltage by 5.5x still gets you something well below the maximum voltage swing of the op-amp, powered by +9V.

I stand by my assertion, and Tom's, that an 18V supply won't yield any headroom advantage over a 9V supply.

[Elektrojänis]

It's something that's "done" - people ask for it...why would they do it if 'nothing changes'?   (scratching head emoji) 

Does it really change or do they just do it and think it changed... Effect also known as experimenters expectancy or observer expectancy.

It's very hard to judge sound quality if you dont have a real side by side comparison and if you do, the two unists could just be different. It could all be just component tolerances. Even if they would be exactly same you could be subconsciously change your playing when you expect something is different.

The test for subtle differences really needs to be a double blind test.

Then again... There could be subtle differences with higher power supply voltages, but does it really matter if even your pick attack changes the sound more?

[ElectricDruid]

When the tone control is set to max treble, the 1k feedback resistance, and 220R ground leg, DO provide a gain of about 5.5x.  But here's the critical element: that gain is for content above 3.3khz.  What is the signal level of content up there; especially after the lowpass filter of the 1k/220n network?  Not very high.  So multiplying that voltage by 5.5x still gets you something well below the maximum voltage swing of the op-amp, powered by +9V.

I did a sim of the tone control section, including the 1K/220n network, and I get a maximum gain of barely over 1dB for the mid-bump. So the gain of x5.5 only puts back a little bit more than the LP takes away. While it's not quite true to say "there's no overall gain", it's almost true.

[iainpunk]

It's something that's "done" - people ask for it...why would they do it if 'nothing changes'?

probably for the same reason people change the OP07 with the LM308 (despite double blind tests show that guitarists don't notice the difference, and most preferred the 741 i threw in as well)

(edit: this ones for vivek  ) running a TS circuit on 18v does change the slew rate affected frequencies (point of -3dB of RMS value) from 90kHz to 45kHz.

cheers

[Elektrojänis]

(edit: this ones for vivek  ) running a TS circuit on 18v does change the slew rate affected frequencies (point of -3dB of RMS value) from 90kHz to 45kHz.

Well that could be important if the amount of dog avatars on this thread is any indication.