3legged dog, a necessity modification

[dukie]

This mod is based on runoffgrooves.com 3legged dog distortion. It came as a necessity mod because i could not find J201 transistor here in Jakarta,Indonesia. im using MPF102 as replacement at first but the sound is very weak and barely audible. After some reading on JFET biasing materials, i experiment with the drain and source resistor but this time using 2n5458 because the MPF102 is used on my GGG RAT Rodent. Changing the drain resistor with 470Ohm and 220ohm at the source produce a unique sound even tough not the same sound of the original 3LGD. Then i changed the drain resistor with 500k lin pot and 560K at the source and add more gain plus some decent distortion and sustain. So i posted this mod to find solution on how to add more sustain, some dirt and maybe if even possible to add a drive stage.

Here is the mod layout :


I also added 4 diodes at the output stage just to experiment with the sound in order to intoduce more distortion, but the gain/volume is greatly reduced.
here is the sound samples :
3LGDtest2WdiodeDist.mp3
3LGDtest1.mp3

I hope great people here can help me improve this (necessity) mod and my knowledge in DIY effect buliding..
Cheers!

[edvard]

Don't know if this will help, but read up at http://www.hawestv.com/amp_projects/fet_preamp/fetpreamp1.htm for an in-depth tutorial on biasing the MPF102 for optimum gain.

Diodes at the end will always attenuate, because you are effectively chopping your signals peak-to-peak voltage down to the level of the diode's voltage drop.
For silicon, that's ~0.6 volts, 0.3 for germanium.
The reason why many overdrive pedals use LED's for clipping is they have a higher voltage drop (1.5 to 0.9 volts) than 'normal' diodes, which leaves mores signal and doesn't clip as harshly.
If you really like the sound of diode clipping, maybe give LED's a shot.

[dukie]

Don't know if this will help, but read up at http://www.hawestv.com/amp_projects/fet_preamp/fetpreamp1.htm for an in-depth tutorial on biasing the MPF102 for optimum gain.

i've been reading that article too.. i change the source resistor(220ohm) with smaller value than the drain (470ohm) on my first atempt but the result is still not satisfied me until i changed it with 2n5458..well im still not satisfied though but its better than using MPF102..maybe i should try using MPF102 again, when i have a multimeter (its quite expensive here) to measure the exact voltage/current on the JFET 

Led?hmm been reading this too.. i may give it a try...but will it suck more battery power? because in my current true bypass wiring i used one led indicator with 1k resistor and its seem dried my battery out very quickly?

cheers..

[edvard]

If I'm not mistaken, LED's on the output are only sloughing off signal voltage peaks, not drawing battery current, which is what your indicator LED is doing.
If output diodes were hard on batteries, then everybody on every stompbox website would advise against diode clipping on those grounds alone.

Just for kicks, experiment with higher-value resistors on your LED so it still lights but won't drain your battery so quickly.
I just did a quick test and with a 1k resistor, 3 different LED's came in at ~ 7ma.
With a 2.2k resistor, that went down to ~ 3ma which is less than half the power draw, and it was still bright enough to see in all but broad daylight.

Also try a different color LED. I just tried a green one and it clocked in at 2.3ma with a 1k resistor.

[dukie]

Also try a different color LED. I just tried a green one and it clocked in at 2.3ma with a 1k resistor.

Hi Edvard, well im using green LED with 1K resistor too and tried using 2.2k resistor but the LED is too dimmed..maybe something else cause the battery to drained?

Anyway i've done another experiment with it..changed the transistor using MPF102 with no significant improvement, subtle distortion and more cleaner sound. i played  with the drain and source resistor value from 100 ohm to 1.5M and swap it between drain and source respectively. and my conclusion is, the bigger the resistor value at  drain or source resulted in louder, cleaner sound while smaller value created more clipped distorted sound with reduced volume/gain.

as far as i know the JFET stage is actually a booster or amplifier stage isnt it? if it is then it working just fine. so what the CD4049 is use for?could it be that the CD4049 is  the distortion stage?and if it so maybe the CD4049 is malfunctioning?well those thought kept me wake all night long since i dont have a multimeter to test it 

Since i cant sleep with the mistery puzzling in my head im doing another experiment and changed the transistor again, this time im using J200 MOSFET P Channel type. Twisted the drain lug because its on the center while the N channel FET drain are on the outside. and to my suprise it works! i bougth this J200 MOSFET by mistake because i thougth it had close number to J201 without knowing that its a totaly different stuff 

And while im at it i swap the 500k pot for biasing the drain with 1M from gain and change source resistor with 1.5M.The J200 suprisingly react very well to this setup, its louder and no noise what so ever! but once again the distortion that i sought after is not present... well atleast i got my self a booster for now..

so anybody have a clue of what happened here or maybe what the CD4049 is use for?

cheers..

[edvard]

The CD4049 is where your distortion happens. It's meant as a digital device, but it works on a range of voltages, which makes it nice for 9v operation.
What does it do? It's a 'hex buffer/inverter' which means it is meant to take a negative pulse and turn it into a positive pulse and vice versa.

Serendipitously, with proper biasing it will become a linear device which will pass plain old sine waves and invert them just as handily... much like an inverting op-amp or transistor will do to your guitar signal.
The 'magic' happens because the 4049 is a CMOS device, which is related to the technology FET's and MOSFET's.
It's well known that FET operation is close to how tubes handle signals and so they are popular for capturing a tube-like sound at modest power levels... like 9 volts.
However, FET's are notoriously difficult to bias correctly (as you found out) due to wild variations in their specs, so the 4049 is basically a way to get FET-like operation (and therefore, sound) without the biasing headaches.

For the record, the original idea came from Craig Anderton's book "Electronic Projects for Musicians" and it was called the "Tube Sound Fuzz" (the circuit can be found here) and only used two stages of the 4049. In that book, he addressed the problem of excessive battery drain in the notes.
Thus I quote...

Excessive current consumption (eats batteries):
Some 4049's draw more current the others. If you have a couple of 4049's around, try plugging in different chips (turn off power while changing chips!!) and reading the current consumption with a meter. Choose the chip with the lowest current consumption.

Sounds like you really need a meter...

[dukie]

The CD4049 is where your distortion happens.

Well in this case maybe the CD4049 is malfunctioning, because the sound get amplified trough the JFET stage and the distortion i found earlier is caused by non proper biasing of the JFET. that is actually attenuation not distortion isn't it?.the sound i got now is clean loud sound, but when the 1M pot that biasing the drain is set at 12 o'clock and the 500k Gain set at full or 12 o'clock give a tubey sound especially when using double single coil mode in my humbucker pickup..but again no distortions.so right now im using it as abooster for my RAT.

Serendipitously, with proper biasing it will become a linear device which will pass plain old sine waves and invert them just as handily... much like an inverting op-amp or transistor will do to your guitar signal.

wait a minute, then the CD4049 is still need "proper biasing"?i dont get it edvard?

Excessive current consumption (eats batteries):
Some 4049's draw more current the others. If you have a couple of 4049's around, try plugging in different chips (turn off power while changing chips!!) and reading the current consumption with a meter. Choose the chip with the lowest current consumption.

then i'll need a 9v DC adapter then..

Sounds like you really need a meter...

..and a multimeter.. and maybe i should buy another CD4049 to test it...

cheers..

[edvard]

Serendipitously, with proper biasing it will become a linear device which will pass plain old sine waves and invert them just as handily... much like an inverting op-amp or transistor will do to your guitar signal.

wait a minute, then the CD4049 is still need "proper biasing"?i dont get it edvard? Huh

Sorry, didn't meant to jump the gun.
What I'm referring to is the feedback resistors on the inverter stages in the 4049. THAT is what makes the "proper biasing".
Don't worry, it's all part of the circuit already.

I'm puzzled as to what's happening to your circuit.
The 4049 SHOULD be distorting quite a bit, not just boosting, and you shouldn't need (at least I've never needed) more than 10k on the drain of your FET. I've never seen or used 1M even on tubes, so I don't know what is happening there.

Hopefully it's not blown, so let's try a few things:

On the FET stage, try a 10K on the drain and a 5k pot at the source with the wiper terminal connected to ground. Adjust it for loudest sound.
When you get a multimeter, measure what the pot is set at when you find the 'sweet spot' and replace it with a resistor of the same value.
Touch your output wire (the one that is attached to the output jack) to the output of each stage of the 4049. Pin 6 first, then pin 4, then pin 2.
It should get louder as you go down the line.
Lastly, try going direct into the 4049. Go directly from the input to the 10nf capacitor that is connected to pin 7.

Let us know what happens...

[dukie]

Changed the tansistor to MPF102 and drain resistor to 10k also used 10k trimmer at the source and  it works like a charm!thank you for the advice edvard!
I'll post the sound sample soon after i have time to buy some batteries..yes it eat battery like a mad dog!

Once again thank you for the help Edvard!


Cheers!

[edvard]

Glad I could help! 

So does that mean the pedal sounds like it should now?
I've been considering building one of these myself, maybe I ought to do just that...

[dukie]

well edvard that is something you must decide your self by listening to the clip i attached here

3LGDtest3.mp3

FYI i recorded this sample directly to the computer soundcard, while the original 3legged dog use a bulldog cab sim before the effect. but i think something is still missing i mean the sound is lack of sustain. the note dies ungracefully, and i think this maybe just maybe caused by non proper biasing.  im using 5k trimpot at first but change it to 10k because the distortion sounded too shallow. yet im still not satisfied but i got to live with it and find another way to tweak it.

Edvard, looking at what happen here i have a question to ask you about the 10k drain resistor and 5k pot choice you adviced me. why did you come up with those resistor value, i mean ive been trying to find the exact value for 2 weeks and still cant find the perfect match. i did read the biasing article here and there but still cant find an exact formula to use. why im asking this because there are so many effect layout out there that i want to build is using J201! so that will be a tedious task to do every time i want to properly bias a replacement trannies. is there any general value that i can use or is it depend on the schematic?or maybe..just buy a multimeter  to make my life easier

Btw im glad that this mod is working but now im thinking to use the J200 Mosfet again. well it got more tubey sound -IMHO- than MPF102 that is louder and  noisy. the J200 mosfet is quiet yet have enough punch and warmth of a tube. what i have in mind is to somehow keep the tube characteristic but in the same time can drive it to overdrive that can be control from a pot. what i need now is, again A Proper Biasing to make it work isnt it?

Well what do you think Edvard? can it be done? 

cheers!

[edvard]

Yes, it can be done. Welcome to the exciting world of FET's where no two are alike, ever...

Now I guess I should explain myself a little...
The 10k/5k values came straight from my own experience, not any secret data sheet, engineering wizardry or actual thinking on my part.
The thing is, I've never had any given FET not work with a 10k on the drain and something between 0-half the drain value at the source.
That said, my suggestions here may or may not be "proper" at all, but will probably work in most situations. YMMV, IANAEE, Ad Nauseum, et cetera.

The big difference you'll find between the J201 and the MPF102 is the amperage they "want" for proper operation.
From the datasheet, we see that the J201 will be perfectly happy with 0.1 - 2.0 mA (that's called IDss on the data sheets).
The MPF102 is a little hungrier and wants 2.0 - 20 mA(!).
So if we want to do this (somewhat) properly, we first need to figure out how much (or how little) juice to let it have.
REMEMBER; The less the resistance, the more the current draw and vice versa.

Using ohm's law (resistance = voltage / current) we can start figuring out some ballpark values for the two transistors.
We know we have 9 volts, and we know we need a certain amount of amperage.
So let's figure out the TOTAL resistance you have to work with. We'll get to biasing later.

OK, for the J201:
9 / .0001 = 90k ohms to 9 / .002 = 4.5k ohms. I do NOT recommend running any given JFET at any more than 75% of it's IDss rating, so we'll crank it down to 9 / .0015 = 6k ohms.
So as long as the drain + source ends up between 90k and 6k, you won't be starving it or burning it out (although you might want to allow a little more on the low end to prevent cutoff in extreme voltage swings, say 50k or so).

And, on to the MPF102:
9 / .002 = 4.5k ohms to 9 / .02 = 450 ohms (whoa!). Once again, going 75% gives us 9 / .015 = 600 ohms for the safety margin.
It's now immediately apparent that my 10k/5k tried-and-true formula should NOT work with the MPF102, but for some KA-RAZY reason, it usually does, although probably not at all optimally.
Once again, welcome to the wonderful wacky world of FET's.

Now we get to the part where we scratch our heads and plug things in to try and make the damn thing work, starting with the J201.

Here is the part where the EE's will tell you you're doing it all wrong, and the guitar players will tell you it sounds great.
Here is where we should be using advanced calculus and a data sheet to figure the gain, but instead we waste time swapping resistors until it sounds great.
Here is where all the math in the world may be marginally helpful, but a multi-meter will be of immense value.
Here is Holy Ground where angels fear to tread.

My general beginning shot in the dark is the drain resistor should be 1/2 - 2/3 of the total resistance.
That's not based on any scientific knowledge, just raw experience.
I also don't like it to be greedy with the milliamps, so let's pick out something like...

Hey! How about 10k for the drain and a 5-10K trimmer at the source? Yeah that oughta do it.

That gives us .6 - .45 mA which the J201 will be completely happy with, and leave us with some room to play.
Set up your test circuit, install the resistors you picked out and tweak the trimmer until you get a nice volume out of your setup.
That's all there should be to it. You're good to go. Really.

When you get a multi-meter, you'll find something curious that you'll want to shoot for every time you do this.
The voltage at the drain (when biased to full gain) will be somewere around 4.5 - 5 volts with a fresh battery.
That's the Holy Grail, man... drink deeply.

On to the MPF102...
Let's start with a sane current draw; it's got to be more than 2.0 mA but not much more or it'll be slurping up scraps of battery under the table where your 4049 eats.
Say 3 mA which would leave us with 9 / .003 = 3000 ohms to work with.
Going with the 1/2 to 2/3 rule that I totally made up out of thin air, you should try a 1.5k - 2.2k resistor at the drain and a 1k trimmmer at the source.
Season to taste.

WOW! I just said WAY too much, but I hope I've helped.
Sorry, I've never done much with MOSFET's but it should be similar. Read your data sheets and use your calculator.
I'm going to listen to your clip in the car on the way home...

Happy biasing!

[dukie]

Now thats what i call a reply! thank you Ed, you open my eyes to new wacky world of Fets biasing knowlegde 

And wow what a wacky world of Fets indeed! ive been messing around today with the drain and source resistor of mpf102 using your reffered value, 1.5-2.2k and 1k trim. and.......it didnt work! im using 5k trim at the drain and 1k pot and nothing good happened..the sound is low and weak , typical syndrome of non proper biasing.

So its still a mystery that it was worked using your first formula (theoretically wrong but technically proven) and i think im going to stick to that for now. because just like old saying in DIY stompboxes building "if you got stuck, leave the damn thing for a while" and hopefuly comeback later with clear head.

Im done tinkering the MPF102 for now and start to explore the capability of J200..it got character and and warmth of a tube -IMHO-. the last setup i done with the J200 is 1M pot at drain and 500k pot at source and give more a tubey feel with nice breakup. as it is it can be used as a booster but still not engouh for me, i want to push it to overdrive more. by my expirience with it the larger the drain resistance the more it breakup, so theoretically it will need bigger pot! 

The J200 datasheet add more puzzling fact..one of them is that the Drain-Source Voltage is -180V! Idss undefined and max at -1.0ma! the Vdss(on) is  -1.5 and -5.0v max! now can you beleive that? later i found out that this product is a Complementary to 2SK1529, and what a shock to know that the 2sk1529 has opposite ratings!maybe this pairs just like matched tube?. another feat is that the 2sk1529 has Vdss 180V! so its a high power application amplifier! at this point i still got no clues what this pair of transistor used for.

So Edvard its time for Googling! 

cheers!

[edvard]

ive been messing around today with the drain and source resistor of mpf102 using your reffered value, 1.5-2.2k and 1k trim. and.......it didnt work! im using 5k trim at the drain and 1k pot and nothing good happened..the sound is low and weak , typical syndrome of non proper biasing.

Hmmm... very strange.
Tonight I'm going to clear my breadboard and pop in a MPF102 with some trimpots and see what happens.
There has GOT to be an easier way to figure these things out...

BTW- thanks for the clip. I'll probably be building a 3LD for myself pretty soon...