Treble Boosters what matters?

[Wavelength]

Gang,

So like a year ago I built and sold a couple treble boosters based on the Mullard OC140 which to me sounded a lot like the Mullard OC44. I had enough 90-100 hfe to make these so what the hell.

I bought a Hudson Broadcast and was reading Pete Cornish site and woke up in the middle of the night thinking to myself... Brian May played through that silly 1W germanium amp why not transformer couple a Treble Booster and get some of that transformer saturation in a pedal as well.

Ok so I am making this and it sounds good input impedance a bit higher than the OC44 at 25K (more than 2x, Rangemaster = 12K). 2700pF =~ 2.5KHz -3dB tested ok. Also can switch in more cap if I want so I do a 500Hz and then full using a 0.1uF.

Nice I can set the phase to inverted like a standard Treble Booster (Rangemaster) or true phase so I added a DPDT to the design and all that works great.

I also added my LDO shunt regulator to the design to lower the noise and that works well with something with this much gain. Plus an added bonus is the output transformer really keeps the HF noise out of the output.

~~~
Now I am over thinking this a bit and these questions are hitting my head.

a) Does the 12K input of a rangemaster have an effect on the tone? Ok yes everything has an effect on the tone. Pete Cornish puts a buffer up front that has like a 500K+ input impedance and he just goes on the 2.5KHz filter and not the germanium or any of that bs.

b) I prefer Polystyrene caps for anything 0.01uF or below. Original are film caps type got me??? Cornish has a 10uF on the output of the buffer stage going into a film cap for the filtering.

c) I have a Prism dScope III that I can test this stuff with, low output impedance. Or I can use one of the AudioQuest DragonFly USB DACS I designed with really low (0.4ohms) output impedance + Faber Signal Scope Pro to look at the -3dB in action. But what, my Tele bridge is like 7K || 250K and 25 feet of Analysis Plus cable to my Treble Booster. How does that effect the -3dB point of the filter?

~~~
Anyway thoughts?
Thanks,
Gordon

[ElectricDruid]

So like a year ago I built and sold a couple treble boosters based on the Mullard OC140 which to me sounded a lot like the Mullard OC44. I had enough 90-100 hfe to make these so what the hell.

Nice. I also built a few boosters (OC44s) and they sounded good. Just sort-of "vintaged" your sound up a bit, if you know what I mean. It must be distortion at some measurable level, but not so much you hear it as such.

a) Does the 12K input of a rangemaster have an effect on the tone? Ok yes everything has an effect on the tone. Pete Cornish puts a buffer up front that has like a 500K+ input impedance and he just goes on the 2.5KHz filter and not the germanium or any of that bs.

I don't agree that "everything affects the tone". That's one of those big pedal myths. The fact is there's a ton of parts in a pedal you can swap around without affecting the sound at all, and even for crucial frequency-response-determining parts, you can still tweak stuff in quite radical ways and not affect the response (RC filters for example - boost the Rs by x4 and reduce the Cs by x4 and the overall response doesn't change, other things being equal).
But to address the question directly...Yes, 12K is a seriously low input impedance and will affect the response out of many guitars. At this point it's going to depend on the detail of what's inside your guitar, which makes life harder still and also explains why Pete Cornish is putting that buffer in there - it gives him a consistent signal with full treble to work from. Any modifications after that are by design rather than by accident. Otherwise, you're building a pedal which sounds great with one guitar and hopeless with another.

b) I prefer Polystyrene caps for anything 0.01uF or below. Original are film caps type got me??? Cornish has a 10uF on the output of the buffer stage going into a film cap for the filtering.

Polystyrene are certainly highly regarded for audio, but polypropylene and other good quality plastic film caps are also excellent. I can't talk about the 10uF without seeing a schematic, but it sounds to me like P.C. is making sure he has a -3dB point at fractions-of-a-Hz, which is generally wildly excessive (even counterproductive since you get no blocking of mains hum). However, if you want to charge Cornish prices, maybe you need to advertise a "full bass response to 0.1 Hz" or something. That's the kind of stuff that goes on in the boutique pedals world.

c) I have a Prism dScope III that I can test this stuff with, low output impedance. Or I can use one of the AudioQuest DragonFly USB DACS I designed with really low (0.4ohms) output impedance + Faber Signal Scope Pro to look at the -3dB in action. But what, my Tele bridge is like 7K || 250K and 25 feet of Analysis Plus cable to my Treble Booster. How does that effect the -3dB point of the filter?

I'm not totally clear what you're proposing here, but it definitely sounds like a case of that over-thinking you were mentioning. If the output is low impedance and the input you're feeding is high impedance, what's the problem? Low is "less than 1K", high is "more than a few hundred K" in my view. It should be possible (easy, even) to design stuff where the roll-offs are so far away from anything crucial that we just don't care. If you're talking about sending the signal from the booster to test equipment, then you absolutely should be in this situation. The test equipment should easily be able to display what you've got without any further worries.

HTH

[Wavelength]

Tom,
Thanks for the reply. In a circuit as simple as this is... correct not much can make a difference in sound I thought till yesterday. So as I said I use a shunt regulator on this and it's an adjust to voltage type like this:



So basically the JFET runs at IDSS into the PNP which gives you a output Z of about 7.4ohms and lowers the noise from the supply between -26dB and -40dB from the testing yesterday. BUT!! I used eight 9V supplies which ranged from 8.6V to 9.2V that I had on hand and of course that shunt regulator would vary as well. The result is the transconductance changes and the input Z changes and the filter knee changes. So I will probably go to something a little more fixed like this:



~~~
I have a couple Fuzz pedals I make and there even the smallest change can end up with big changes.

In high end audio everything makes a difference. I have people asking me about power cords and why USB cables sound different. Hell even the difference between bit true playback applications or the difference in sound between macOS and Windows. There it really depends on your system and how it responds to change.

Here in pedal land it is less of an issue because there are less specifications to follow.

I am going to run some tests today on the effects of the knee of the filter point and the output Z of sending signal. Put like a 10K pot in series with the DragonFly and see what happens. I guess caps in the cable and the tone circuit will have a difference as well, but I will keep that out of the mix for now.

Thanks,
Gordon

[Wavelength]

Ok all this is interesting...

So basically a treble booster is a high pass filter. So I did this: Low output impedance oscillator->10K Pot->Treble booster.

Pot = 0             Straight into Treble Booster   -3dB = 2150Hz
Pot = 3K           Charlie Christian type pickup   -3dB = 2325Hz
Pot = 5K           Low wind Strat and Tele           -3dB = 2450Hz
Pot = 7K           Single coil bunch of options   -3dB = 2675Hz
Pot = 8.5K   P90 Jazzmasters etc..           -3dB = 2890Hz
Pot = 10K           Humbucker, high wind SC           -3dB = 3200Hz

Now this doesn't even consider using 2 pickups in parallel which would yield much lower impedance.

Of course some cable losses and of course some add capacitance from the cables but let's throw that out and of course this assumes vol = max, tone = max so the impedance output is not as much effected by those.

So the input Z of the pedal does have an effect on the -3dB point. Since this is an NPN with higher input impedance than the standard Rangemaster by 2:1, then your effects will probably be even more of a change.
Thanks,
Gordon

[ElectricDruid]

I think you might have discovered why Pete Cornish put a buffer in front of it. That way, the input impedance is fixed, the guitar sees a nice high impedance so you get all the treble and tone you're after, and then you can mess with the signal without screwing thing up further back up the chain.
Those "it responds to your guitar" pedals *sound* like a really cool idea, until you discover that they only sound decent with one of the three guitars you have and the other two sound hopeless! Hence a bit of decent electronic engineering to make sure we get a reliable result with the wide variety of guitars, pickups, and signal levels that someone could plug in. I think Pete got it right on this one.

[blackieNYC]

Prism dscope III? Nice unit.  Now everyone's going to want one for pedals!

[Wavelength]

Tom,

Exactly! Pete took the 2.5KHz from the original Rangemaster and said ok how do we do this all the time. I built up a bunch of buffers for Mick & Dan on That Pedal Show and Pete's bootstrapped buffer was one of my favorites: Bootstrapped NPN 511.875K (input impedance) 114 (output impedance) -0.075dB (gain)  -135dB (signal to noise ratio) 2.5vrms (max output voltage).

The only problem with PC's is that it doesn't change with guitar. I have a T type (Peto Guitars) with Lollar BS pickup in the bridge and now I understand why it sounded so different than my Fano.

~~~
Blackie, hope you are well! Been talking to Joe Roberts the last couple of days. He just moved from DC to Philly

You know I have this test equipment because I design audio for other companies and just need that stuff. With a really good AD/DA converter and I am sure some software that is free you could do at least 70% of what the Prism can do.

Thanks,
Gordon

[Harry Muff]

So like a year ago I built and sold a couple treble boosters based on the Mullard OC140 which to me sounded a lot like the Mullard OC44. I had enough 90-100 hfe to make these so what the hell.

Nice. I also built a few boosters (OC44s) and they sounded good. Just sort-of "vintaged" your sound up a bit, if you know what I mean. It must be distortion at some measurable level, but not so much you hear it as such.

a) Does the 12K input of a rangemaster have an effect on the tone? Ok yes everything has an effect on the tone. Pete Cornish puts a buffer up front that has like a 500K+ input impedance and he just goes on the 2.5KHz filter and not the germanium or any of that bs.

I don't agree that "everything affects the tone". That's one of those big pedal myths. The fact is there's a ton of parts in a pedal you can swap around without affecting the sound at all, and even for crucial frequency-response-determining parts, you can still tweak stuff in quite radical ways and not affect the response (RC filters for example - boost the Rs by x4 and reduce the Cs by x4 and the overall response doesn't change, other things being equal).
But to address the question directly...Yes, 12K is a seriously low input impedance and will affect the response out of many guitars. At this point it's going to depend on the detail of what's inside your guitar, which makes life harder still and also explains why Pete Cornish is putting that buffer in there - it gives him a consistent signal with full treble to work from. Any modifications after that are by design rather than by accident. Otherwise, you're building a pedal which sounds great with one guitar and hopeless with another.

b) I prefer Polystyrene caps for anything 0.01uF or below. Original are film caps type got me??? Cornish has a 10uF on the output of the buffer stage going into a film cap for the filtering.

Polystyrene are certainly highly regarded for audio, but polypropylene and other good quality plastic film caps are also excellent. I can't talk about the 10uF without seeing a schematic, but it sounds to me like P.C. is making sure he has a -3dB point at fractions-of-a-Hz, which is generally wildly excessive (even counterproductive since you get no blocking of mains hum). However, if you want to charge Cornish prices, maybe you need to advertise a "full bass response to 0.1 Hz" or something. That's the kind of stuff that goes on in the boutique pedals world.

c) I have a Prism dScope III that I can test this stuff with, low output impedance. Or I can use one of the AudioQuest DragonFly USB DACS I designed with really low (0.4ohms) output impedance + Faber Signal Scope Pro to look at the -3dB in action. But what, my Tele bridge is like 7K || 250K and 25 feet of Analysis Plus cable to my Treble Booster. How does that effect the -3dB point of the filter?

I'm not totally clear what you're proposing here, but it definitely sounds like a case of that over-thinking you were mentioning. If the output is low impedance and the input you're feeding is high impedance, what's the problem? Low is "less than 1K", high is "more than a few hundred K" in my view. It should be possible (easy, even) to design stuff where the roll-offs are so far away from anything crucial that we just don't care. If you're talking about sending the signal from the booster to test equipment, then you absolutely should be in this situation. The test equipment should easily be able to display what you've got without any further worries.

HTH

i like silver mica ...you can get cheap lots on ebay

[mac]

a) Does the 12K input of a rangemaster have an effect on the tone? Ok yes everything has an effect on the tone. Pete Cornish puts a buffer up front that has like a 500K+ input impedance and he just goes on the 2.5KHz filter and not the germanium or any of that bs.

In general the low input impedance and the tiny input capacitor of a RM interact with the guitar coil and its series resistance giving a fixed low Q wah-like frequency curve with the peak at  about 1khz, depending on input cap, guitar coil, etc.
It is more a bandpass than a treble booster IMHO.
A buffer kills this effects.

LTSpice simulation of Silicon Booster with/out guitar mic model. No model "is like" adding an input buffer.



mac

[ThermionicScott]

So treble boosters were really midrange boosters all along... funny how we keep coming back to that. 

[Wavelength]

So treble boosters were really midrange boosters all along... funny how we keep coming back to that. 

Scott, actually it's way above a guitars midrange. Almost to overtones. Standard guitar range is 80->1200Hz, the Treble Booster -3dB point is 2.5KHz

Thanks,
Gordon

[amptramp]

I like the idea of a high-impedance input since the inductance of the pickup forms an L-R lowpass filter with the input impedance of the booster.  I also prefer to have all the frequency-determining parts isolated by buffers on both input and output or you get frequency response variations with every change in setup.

I would use a large coupling capacitor into the booster so the input impedance is not isolated from the impedance of the guitar.  This way, the resistive noise voltage of the booster is shunted by the output resistance of the guitar and is reduced that way to the noise of the parallel resistance of the guitar and booster right down to the lowest frequencies.

As for the lowpass at the output, there is probably no need for overall pedal chain response beyond 5 KHz and any feedback stabilization capacitor can be sized with that in mind.  But this is a -3db cut and if you have a long pedal chain with a lot of -3db cuts at the high end, you may find yourself more than 20 db down at the high end and you have to modify all your pedals to recover the high end response.  I generally go with 10,000 Hz response in the buffer.  If you want less high end, it can be chopped out later.

Fun fact: The first album to have no high end above 5 KHz was "No Jacket Required" by Phil Collins and it did alright in the marketplace.

[teemuk]

How often is a "treble boost" just a treble boost and not also voicing for distortion? In later case it also helps to filter out very highest frequencies from contributing too much to intermodulation distortion.

[Wavelength]

Ron,

Basically your describing the Pete Cornish version of the Treble Booster. Buffer input->High Pass and out. The problem like the fuzz problem is that the interaction of the guitar and the pedal is just not there. Basically pick your poison and go with what you are looking for. The PC worked for Brian May so I guess if that is what your going for then do that.

~~~

teemuk,

Yes there is a distortion in this kind of pedal. With Germanium there is a soft clipping of the negative portion of the waveform. Plus massive 25-30dB of gain of the overtones so this is basically a distortion machine.
I think on my Prism test set it was showing like 40% thd.

I may look at that again as I compare a couple other designs in the next few days.

Thanks,
Gordon

[iainpunk]

teemuk,

Yes there is a distortion in this kind of pedal. With Germanium there is a soft clipping of the negative portion of the waveform. Plus massive 25-30dB of gain of the overtones so this is basically a distortion machine.
I think on my Prism test set it was showing like 40% thd.

i totally believe the tester would show 40% thd, but im certain its not quite correct to claim that there actually is 40% distortion. the thing with such tests is that it relies on a fully flat frequency response to work correctly, but in its ridiculous tone shaping, a lot of fundamentals and low end is taken out, which increases/skews the ratio distortion to fundamental.

cheers

[Wavelength]

Iain,

Treble Boosters will soft clip the lower portion of the waveform. The Prism, AverTech, Ap or other high quality test set will show massive distortion like that because of that function.

Basically a Treble Booster is boosting harmonics not even the fundamental of the waveform. Guitar range is 80-1200Hz and the -3dB high pass is ~2.5KHz so the fundamental part of the output is pretty low and the harmonics are lifted. That is why the thd is so high.

Thanks,
Gordon

[teemuk]

Ofcourse we largely employ them to enforce intermodulation distortion of higher frequency spectrum, not lowest harmonics or fundamental. That would just obfuscate chords and overall make the distortion sound "flabby".

My point is that treble booster is hardly as is, but usually part of distortion processing, whether it distorts by itself, overdrives following stages or all of the above. Doing that a nice side effect is if it also reduces the contribution of most highest frequencies to IMD, the kind of harsh sizzle and hiss. When we look at distortion drsigns in general - other than fuzzes - they kind of revolve around boosting circa 1kHz as pre-distortion emphasis.

Rangemaster is pretty handy because it hi-passes at treble range and - probably by happy accident - also lo-passes at upper end of the treble range because of low input Z interacting with the pickup circuitry.

[amptramp]

Just from the standpoint of selling a treble booster, I prefer the use of high input impedance and frequency determination being safely protected by a high-impedance input and low-impedance output.  If the guitar characteristics become a big part of the sound then you can't really tell the guitarist "This is what our treble booster will sound like" because it would be dependent on the customer's setup.  A number of guitarists change pickups for various reasons and it would pay to have predictable response regardless of the pickup used rather than having something that sounds different with every setup.

It may be true that some of the greatest guitarists have used pedals that interact with the guitar such as the Fuzz Face and Cry Baby but if your name is Clapton, you will be presented with a bushel basket of "identical" pedals to choose from.  Not the case if you are buying online or you are just a minor artist buying a pedal at a local shop.

[Wavelength]

Teemuk, Amptramp,

Yes a Rangemaster is first in a chain used to drive a high gain amp (or some high gain OD) to the point of getting that sound. As a EE we were always taught high in low out. BUT!!! Effects do not fall into anything that electronic engineers can wrap their heads around. Sure the buyer has a real problem in the world of Rangemasters, Fuzz and OD because of this. How many of us have a pile of pedals that we don't like? All of us... heck bought my third analog reverb the other day and can't stand it can't send it back.

So remember with these pedals (i.e. Rangemasters, Fuzz and some discrete OD's) is the fact that we are distorting the input to output so EE side of the brain is there just to make it better a little quieter and a bit true to the original.

After I kick these amps out the door today I am going to change my biasing which will lower my input Z and change my input caps and see what comes of it.

Thanks,
Gordon

[iainpunk]

Iain,

Treble Boosters will soft clip the lower portion of the waveform. The Prism, AverTech, Ap or other high quality test set will show massive distortion like that because of that function.

Basically a Treble Booster is boosting harmonics not even the fundamental of the waveform. Guitar range is 80-1200Hz and the -3dB high pass is ~2.5KHz so the fundamental part of the output is pretty low and the harmonics are lifted. That is why the thd is so high.

Thanks,
Gordon

a perfect square wave has only got 48% THD, it seems extremely improbable that soft clipping on only one side of the wave would get you over the 15% mark. if such high numbers of THD are measured, there's something wrong with the setup or method of testing.
a THD tester can only give reliable results when the frequency response is flat. if it isn't its measuring other types of distortion as well, mainly frequency response distortion.

TL:DR the distortion SEEMS higher that it actually is.

cheers