Differential / LND150 Distortion

[mdcmdcmdc]

Hi folks,

I ordered a few LND150 mosfets to see what they're all about. After messing around over the weekend, I came up with a circuit that sounds ok? There's not much to it, and it's almost certainly riddled with less-than-best practices. But! It's a fun, thick sounding distortion circuit. A LND150 with some clipping slammed into a BJT differential amp with a good ol' muff tone control.



Clipping diodes are yellow LEDs.

(sorry, noticed a few inconsistencies in the schematic vs the breadboard and re-UL'd the image)

[marcelomd]

Looks nice.

Can you post an audio sample?

Thanks!

[PRR]

Value R3?

Matching of R5/R6 R9/R10 looks critical.

[mdcmdcmdc]

Oops - R3 is 10K, Q1 biases to just over 6V iirc. I didn't really notice much difference between 6 and 7V aside from the difference in volume. I'll try to knock out a quick clip later this eve.

[PRR]

I don't think R3 or Q1 is critical. Just that without a clue your fans/followers may be baffled to pick *any* workable value.

That's not critical because Q2 Q3 only need a part-Volt to slam to the max. Which implies that R5/R6 R9/R10 matching to a small part of a volt may be critical. +/-10% on those resistors could be 300mV offset? Which would throw the diff-pair to one side or the other? And OTOH "perfect match" may not be "most musical". (I recall I had a user-trimmer last time I played with the diff-pair, 40 years back.)

But it is working for you, and is <$10 of parts, so everybody should try it. If some don't work there are ways to make it self-bias.

[Vivek]

For the first stage:

I wonder what difference would we find if we replaced the MOSFET with any other FET or Transistor or Opamp which is only in charge of linear gain, with all clipping functionality given to the clipping diodes

For the second stage:

In another thread, we concluded that if the last stage has strong clipping, it pretty much wipes out all characteristics of the earlier stages.

I'm in the food business.

There are many levels of roasting coffee beans.

Light roasting retains the aroma and taste of the original beans. You can detect the origin and varietal information of the raw coffee beans if they are roasted lightly. Kenyan tastes different than Jamaican.

Then there is ITALIAN ROAST, which is such a dark roast, the coffee bean is almost reduced to charcoal. All varietal information is lost. Any bean roasted to ITALIAN ROAST tastes pretty similar to any other bean roasted to the same level. You can hide bad quality beans if you roast them to ITALIAN ROAST levels.

Having a last stage with high gain is like ITALIAN ROASTING the pedal !




PS : I hereby reserve the name ITALIAN ROAST for one of my forthcoming high gain pedals

[Vivek]

Matching of R5/R6 R9/R10 looks critical.

If the goal of symmetrical or Asymmetrical clipping is not declared,

Would these resistors matter ?

Is it possible to produce more desireable tone by purposely mismatching these resistors ?

[mdcmdcmdc]

I defer entirely to the knowledgable folks in the room on the math and science; I'm fairly ape-eared, and my approach to this stuff is generally just to mess with a stage individually until it sounds good to my ear and then start stacking them together in different combinations and orders until something kind of clicks.

I had tried a few of those LND150 amp stages in series, but I don't love the way that sounded. They sound very nice on their own; and the LED clipping is pretty subtle with the yellows. That differential stage is something I used in a different circuit—I wasn't able to find many examples of differential BJT amps with a single-ended supply; again, it sounded really nice to my ear and distorted in a pleasant way. Slamming one into the other gave me a grin and here we are.

I did try swapping the BJTs in the diff amp with a pair of LND150s; it passed signal, but had very low output. By that point I was happy enough with how the whole thing sounded to not dive into that particular rabbit hole.

[Vivek]

Dear mdcmdcmdc,


By any chance, do you have an Oscilloscope to see wave shape and measurements at the output of each stage ?

[mdcmdcmdc]

Nope, I don't.

Here's two chords through a looper though:

https://imgur.com/yPZlRUF

First two chords are clean, then pedal on with drive at full CCW, sweeps to full CW with tone at 12:00; then tone sweeps to full CCW and then up to full CW. Volume is around noon, just over unity.

[duck_arse]

Q1, being a mosfet, needs a positive bias voltage on its gate. are you perchance missing a resistor from the gate to either Vcc or Drain?

[mdcmdcmdc]

Nope, it's working as shown in the schematic—this might be particular to the LND150?

[Vivek]

LTSPICE shows these DC bias voltages:

[mdcmdcmdc]

6.2V and .76V respectively on my DMM.

[duck_arse]

not to be rude then, but is what you have drawn what you have built? 6V2? there is no positive DC path to the gate shown on your circuit.

ahhh, the datasheet. I guess the following, from said datasheet, means it self biases ......

The LND1 is a high voltage N-channel depletion mode (normally-
on) transistor utilizing Supertex's lateral DMOS technology.

please, pardon.

[Vivek]

But other LND schematics seem to be similarly biased

http://www.hawestv.com/amp_projects/amp_solid_tube/fetpreamp_200v1.htm

[mdcmdcmdc]

not to be rude then, but is what you have drawn what you have built? 6V2? there is no positive DC path to the gate shown on your circuit.

ahhh, the datasheet. I guess the following, from said datasheet, means it self biases ......

The LND1 is a high voltage N-channel depletion mode (normally-
on) transistor utilizing Supertex's lateral DMOS technology.

please, pardon.

Not at all! I think it's still a slightly uncommon device in stompboxes.

[Vivek]

But I feel, in this circuit, we are not really using any special feature of the LND150

Ok, it has high input impedance
Modest gain of 11

Clipping by the LED, Asymmetrical due to bias

Sufficiently low output impedance to drive the next stage


And all that is standard stuff ie same stage built with FET or Opamp would have extremely similar response.

Maybe if built with transistors, we would need an input buffer feeding the clipping stage, so the LND saved one stage in comparison.

And in any case, nobody can hear the LNDness of first stage after it goes through that second stage.

[mdcmdcmdc]

lol I didn't mean to imply it was a game changing circuit design. It's just a distortion pedal that sounds pretty decent to me. It may not be for you!

[Steben]

But I feel, in this circuit, we are not really using any special feature of the LND150

Ok, it has high input impedance
Modest gain of 11

Well those are qualities. Are they possible differently? Yes. But it works.
Somehow I think a jFET might be "on paper" better choice, but try finding the "right" one...
Would I prefer an opamp stage? Perhaps. Or BJT.  But it is just taste. And I am done with FETs (unless the input of let's say a TL072, lol).

And in any case, nobody can hear the LNDness of first stage after it goes through that second stage.

Because it has its limit in flat against the rails.
All clippers which retain an increment at full gain will retain information of the first stage, be it small or large.