EHX LPB-2 has some distortion in bypass

[j_flanders]

I have a point to point wired, '70's LPB-2 that has some distortion, even when the pedal is in bypass.
It has the old style bypass that leaves the input connected and switches the output.
It seems in bypass I'm getting some of the distorted signal mixed in. But how or why?
When I put the battery switch in the off state it just passes through my dry signal without the distortion.

On the other hand,
When the pedal is on (not bypassed) and the battery is off there should not be any sound, but despite the transistor not getting any power, I'm getting some gatey, sputtery distortion at the output.

How can both be explained and the first problem be solved?

Here's the schematic how I traced it plus some voltages:


I'll add a photo of the wiring later in case that can help.

[antonis]

It seems in bypass I'm getting some of the distorted signal mixed in.

It might not be the LPB2 distorted signal but bypassed signal is distorted due to LPB2 low input impedance..
(about 20k or so..)

[j_flanders]

It seems in bypass I'm getting some of the distorted signal mixed in.

It might not be the LPB2 distorted signal but bypassed signal is distorted due to LPB2 low input impedance..
(about 20k or so..)

I'm running/testing straight into the input of a Plexi style (pre-)amp:

[Rob Strand]

The fact the transistor base loads down the guitar pickup is enough to cause some distortion but when the transistor clips the distortion will get worse.  The solution is true bypass to stop the circuit loading the guitar.

The signal getting through without power is just the collector to base junction (a diode) conducting which then creates a path from input to output.

[j_flanders]

It might not be the LPB2 distorted signal but bypassed signal is distorted due to LPB2 low input impedance..
(about 20k or so..)

The fact the transistor base loads down the guitar pickup is enough to cause some distortion but when the transistor clips the distortion will get worse.  The solution is true bypass to stop the circuit loading the guitar.

Thanks for the replies.
How or where is that distortion created? What limit am I hitting where?
The pedal is in bypass.

The signal getting through without power is just the collector to base junction (a diode) conducting which then creates a path from input to output.

Thanks. I see. I had this 'problem' before with a Fuzz Face, but this explains it.

Another question about this circuit:
In my pedal the previous owner replaced some resistors, probably also the transistor as I cannot find any reference to this resistor being used by EHX but more importantly (or maybe not) he rewired the battery leads compared to the originals.
In the original pedals the battery + goes to the collector but also to the enclosure (sleeve of the input jack).
In my pedal someone wired it as we're used to: battery+ to the collector and battery- to enclosure(sleeve of one of the jacks).
Why would EHX connect the battery + to the enclosure while using an NPN transistor?
Some refer to this pedal as positive ground but is that entirely correct?

[Rob Strand]

Thanks for the replies.
How or where is that distortion created? What limit am I hitting where?
The pedal is in bypass.

The impedance looking into the base loads down the pickup impedance.  *If* the impedance looking
into the looked like a pure resistor the loading effecting would affect the level  and frequency response
but doesn't add distortion.  However a transistor is a non-linear device so this loading is non-linear
and it adds a very small amount of distortion.   For small signal levels you could ignore it for guitar but
a hi-fi designer would take a closer look.    This mode of distortion is unlikely to be what you are hearing.

When the transistor clips due to a positive input signal, the impedance looking into the base is *very* low.
It almost blocks the signal at the base while the clipping is present.   In effect this cause hard clipping
of the positive cycles *at the base*.   Since the base is always connected it affects the bypass mode.
The clipping occurs when the base signal is positive and the transistor turns on and pulls down the collector
voltage.  When the collector voltage hits the emitter voltage you get clipping.  At the collector the negative cycle
is clipped as the transistor inverts the signal.

What to do about it?  Well the whole way that circuit works is it clips slightly so you don't really want to stop
the clipping.     You can make small improvements by decreasing the gain so it needs a high input before clipping
occurs by  increasing the 390 ohm resistor.  Or you can rebias the transistor by raising the voltage on the collector so
it needs more signal to clip on positive inputs.    You will only make small gains here and if the signal is large enough
it will still clip. Making small changes is fine but you don't want to change things too much otherwise the pedal won't do what
it supposed to do.

So you are pretty much stuck and the best way to avoid the problem is to do a true-bypass.

Another question about this circuit:
In my pedal the previous owner replaced some resistors, probably also the transistor as I cannot find any reference to this resistor being used by EHX but more importantly (or maybe not) he rewired the battery leads compared to the originals.
In the original pedals the battery + goes to the collector but also to the enclosure (sleeve of the input jack).
In my pedal someone wired it as we're used to: battery+ to the collector and battery- to enclosure(sleeve of one of the jacks).
Why would EHX connect the battery + to the enclosure while using an NPN transistor?
Some refer to this pedal as positive ground but is that entirely correct?

As far as giving the ground a name if the ground goes to positive it is positive ground.

As far as what is better, you will usually get less noise problems with negative ground on NPN circuits and positive ground on PNP circuits.

I don't know why EHX wire the pedals like that.  I have seen it.   At first I though someone modded the pedal from PNP to NPN and didn't re-wire the grounds but it's not the case.  I think they are really wired like that.   On the web I've also seen PNP transistors in positive ground units.   When you add onto that people modding units who knows what you get.

Maybe some of the EHX geeks can comment.

[PRR]

> How or where is that distortion created? What limit am I hitting where?

Fuzz is like hitting the walls of your garage with your car.

When that happens, you don't have to see the dents, you can feel the impact in your steering wheel.

Even though your too-simple switching does not connect the fuzz output to anything, the input is connected, and with a hot guitar it will be distorting. And you can "feel" that in the Base terminal.

It may have been acceptable with older lower-output pups and honestly less fussy stylings. Today's pups spit big signal and everybody is used to precision sound.

This is why 2p and 3p switches have become fashionable.

[j_flanders]

you don't have to see the dents, you can feel the impact in your steering wheel.

Great analogy. 

When the transistor clips due to a positive input signal, the impedance looking into the base is *very* low.
It almost blocks the signal at the base while the clipping is present.   In effect this cause hard clipping
of the positive cycles *at the base*.   Since the base is always connected it affects the bypass mode.
The clipping occurs when the base signal is positive and the transistor turns on and pulls down the collector
voltage.  When the collector voltage hits the emitter voltage you get clipping.  At the collector the negative cycle
is clipped as the transistor inverts the signal.

This reminds me of when I was audio probing a JCM800 preamp and totally unexpectedly found clipping right before the grid of a tube.
Someone at tdpri 'discovered' the same thing and an explanation by Merlin Blencowe posted over here:
https://www.tdpri.com/threads/preamp-signal-through-my-rr800v.1032352/#post-9918816

I guess this is the same thing. Combined with the simplistic bypass I'm sending the 'base current clipping (?)' straight to the output.
It would also mean the clipping I'm hearing when on (not bypassed) is the same clipping but amplified and inverted and coming out of the collector.

So, I assume in other pedals (Fuzz Face, Muff Fuzz) where the input impedance is equally low I would have experienced the same thing, weren't it for the true bypass I have in those.

What to do about it?  Well the whole way that circuit works is it clips slightly so you don't really want to stop
the clipping.     You can make small improvements by decreasing the gain so it needs a high input before clipping
occurs by  increasing the 390 ohm resistor.  Or you can rebias the transistor by raising the voltage on the collector so
it needs more signal to clip on positive inputs.    You will only make small gains here and if the signal is large enough
it will still clip. Making small changes is fine but you don't want to change things too much otherwise the pedal won't do what
it supposed to do.

So you are pretty much stuck and the best way to avoid the problem is to do a true-bypass.

The circuit acts somewhat like a Fuzz Face. I only have this distortion with the guitar volume on 10. It greatly cleans up when dialing it back, even just a little. I defintely don't want to lose that aspect.
True bypass is unavoidable then.

Thanks for all the help.

[Rob Strand]

I guess this is the same thing. Combined with the simplistic bypass I'm sending the 'base current clipping (?)' straight to the output.
It would also mean the clipping I'm hearing when on (not bypassed) is the same clipping but amplified and inverted and coming out of the collector.

Yes, *exactly* the same mechanism.   The only difference is the tube example has the 100k input resistance whereas in your case the source impedance is the guitar impedance.

So, I assume in other pedals (Fuzz Face, Muff Fuzz) where the input impedance is equally low I would have experienced the same thing, weren't it for the true bypass I have in those.

Yes, to some degree they will all do it.  The Muff perhaps the least since it has a largish resistor between the input jack and the transistor base.

The circuit acts somewhat like a Fuzz Face. I only have this distortion with the guitar volume on 10. It greatly cleans up when dialing it back, even just a little. I defintely don't want to lose that aspect.
True bypass is unavoidable then.

Exactly that.   I don't see a way around it.

Even without the distortion issue the input impedance of the LPB is alone going to affect the tone of the guitar.   You should be able to hear the difference between the LPB on bypass and the LPB completely removed.   The tone will be a little brighter with the LPB removed.    This type of loading is what inspires people to go true-bypass.

[MG]

I don't remember the LPB- booster circuit connecting directly to the base of the transistor.  That's a bad idea, even with lower gain pickups.  Kinda contradicts the L in LPB, when the pedal is on or off.

[j_flanders]

Yes, *exactly* the same mechanism.   The only difference is the tube example has the 100k input resistance whereas in your case the source impedance is the guitar impedance.

The JCM800, which the OP and I were probing doesn't have that 100k resistor there. But there's some before the grid leak resistor.

As for the clipping in the LPB-2, I see GGG calls this 'ugly distortion'.
They propose to move the volume control in front of the circuit instead of at the output.
http://www.generalguitargadgets.com/pdf/ggg_cb1_mods.pdf
Similar to what is often done to Fuzz Faces: a variable input resistor or voltage divider at the input.

Yes, to some degree they will all do it.  The Muff perhaps the least since it has a largish resistor between the input jack and the transistor base.

I really meant the Muff Fuzz (or Double Muff) not the Big Muff, which does indeed have a 30k-ish input resistor. The Muff Fuzz or Double Muff doesn't, similar to a Fuzz Face.

You should be able to hear the difference between the LPB on bypass and the LPB completely removed.   The tone will be a little brighter with the LPB removed.

In bypass, without a doubt it has the worst tone suck of any pedal I ever tried.

I don't remember the LPB- booster circuit connecting directly to the base of the transistor.  That's a bad idea, even with lower gain pickups.  Kinda contradicts the L in LPB, when the pedal is on or off.

[Rob Strand]

As for the clipping in the LPB-2, I see GGG calls this 'ugly distortion'.
They propose to move the volume control in front of the circuit instead of at the output.
http://www.generalguitargadgets.com/pdf/ggg_cb1_mods.pdf
Similar to what is often done to Fuzz Faces: a variable input resistor or voltage divider at the input.

Input control is certainly a possibility.   At the end of the day the input volume pot will still load the guitar in bypass mode.
I guess the problem is how much dirt you want in Effects mode.

I really meant the Muff Fuzz (or Double Muff) not the Big Muff, which does indeed have a 30k-ish input resistor. The Muff Fuzz or Double Muff doesn't, similar to a Fuzz Face.

Ah,  forgot about that one. 

[PRR]

LPB _not_ clipping has a 36k input impedance, pretty constant over the audio cycle.

LPB _clipping_, input impedance varies from 2k to 39k. The guitar's internal impedance may be 10k or more, so this is a real load. Since only part of each cycle, a distortion at the input, upon the guitar's signal. Before we even ask about collector clipping.


The severe dips to zero are math burps.

This does not happen at 100mV(peak) input, but does happen at 300mV(pk). So this sits in the zone between older pups and newer pups (and strings and picks and styles and dietary supplements...)

[MG]

Note that the low resistance values of the base bias resistors will also load down guitar pickups.  Pickups are inductive, so the top end will roll off.  This is similar to using a 250k vs 500k pot in a guitar.  Except the base bias resistors are as low as 30k+ in the schematics above.  That's R2 in some of the drawings.

Aside from that, the base-emitter junction is really a diode.  It will break over (conduct) around 0.6 volts or so.  Transistors are current amplification devices, so that's just how it is.

Tubes are not as much a concern at the input of an amp.  Tubes are transconductance devices (voltage to current, not current to current), and grids are normally high impedance.  The tube schematic above shows around 1.6 volts on the cathode.  That's the relevant voltage with regard to grid conduction.  In other words, the bias is effectively -1.6 volts (that's a minus sign cause it's measured with respect to the tube's grid).  The input voltage swings up and down from ground, and would need to reach +1.6 to equate to 'zero bias'.  The internal stages are another story, of course.  Singing compression at the 'sweet spot' on an amp's volume control is due in part to conduction of the output tube grids.

If the 100k resistor that you refer to is on the input (you'll usually see 68k or so on Fender and other amps), it's intended for RF suppression, not prevention of grid conduction.  That's another subject.  You could search for Miller capacitance and come up with some references.

If I were looking for "Linear Power Boost" with the key word 'linear', I'd not use anything like this circuit.  It's horrible.  You could try a relatively low power opamp in non-inverting mode.  High input impedance, as much gain as you want, linearity, low output impedance.  On the other hand, some may find the combination of pickup loading (loss of treble, even in bypass), distortion, etc. to be the sound they're looking for.  ??

PS: When I referred to a 100k resistor above, it was in reference to the Marshall (or Fender or whatever) -input- circuit that was being discussed.  I was not talking about Blencowe's sample circuit that someone linked to.  In the latter case, he was using a 100k in series with a tube grid to show the difference before and after.  Right at the grid, you do see the result of grid conduction, presuming that voltage levels are high enough.