Little Jim - a Marshall inspired MOSFET distortion pedal design

[jonny.reckless]

One other thing I was wondering about: the 100 ohm resistor (R15) in the PSU input stage, does it serve any purpose for the dynamic response (sagging) of the rest of the circuit? Or is it purely for keeping out the PSU noise?

I put it there to get rid of a little bit of hum I was getting. I have no idea if the sag affects the sound or not

[Phend]

Call me ET101, I have searched for this answer, I believe I know, Here are my questions:
On the Little Jim circuit drawing I recon that on the main circuit:
1) VDD shown in 4 places indicates 9v supply (D for drain).
2) VBIAS shown in 3 places indicates ground.
3) The 2 smaller circuits are for 9v dc supply (if used) and an LED indicator (if used).
4) The last is the switch wiring for true bypass.
5) So for battery operation, connect pos to VDD and gnd to gnd, use the main circuit and switch
    circuit and you are good to distort ?

Thanks for any comments , Regards

Paul

[rutgerv]

Hi Paul,

1) VDD shown in 4 places indicates 9v supply (D for drain).

Yes, you could see it that way, though it's not entirely correct. As you can see in the left bottom of the schematic, VDD is in fact the 9V supply *after* it went through some filtering (R15 + C12). As a consequence of the current required by the circuit, and the presence of R15 in this filter causes a slight voltage drop on Vdd, making it 8.56V as indicated in the schematic. The real 9V would only be found on the socket (input) to the right of R15 and gets filtered and then is called 'VDD' for the rest of the schematic.

2) VBIAS shown in 3 places indicates ground.

No!.... the ground is labeled 0V in the schematic and occurs many more times. VBIAS is a voltage that sits between Vdd and 0V (ground), and is created by two resistors (left bottom) R7 and R8 forming a voltage-divider. Capacitor C5 filters the VBIAS supply line.

3) The 2 smaller circuits are for 9v dc supply (if used) and an LED indicator (if used).

Yup! Also take note of a nice example of engineering experience: the LED is fed from 9V rather than from Vdd. If you understood what I tried to describe with respect to current draw and voltage drop, and you know how much current a LED draws, this is a smart thing. Also... switching LEDs induces noise in supply lines, so supplying the LED with a voltage from the 'other side' of the supply filter (R15 and C12) is a clever thing to do.

4) The last is the switch wiring for true bypass.

Correct!

5) So for battery operation, connect pos to VDD and gnd to gnd, use the main circuit and switch circuit and you are good to distort ?

To be exact: for battery, connect - of your battery to 0V and + of your battery to 9V, but leave out R15. In battery operation R15 is less important as a filter, and getting rid of it avoids the voltage drop you'd otherwise get from this resistor. With batteries running empty the voltage will already be on the low side, and you can't afford to lose more voltage. R7 and R8 are crucial to keep! C12 might also be beneficial (I don't know how well batteries deal with load variations).

And then indeed... rock and roll!

Rutger

[Marcos - Munky]

Paul, don't be afraid to ask any kind of question. Nobody is a born know-it-all. Some things that may seems basic-ish were questions everybody had when we all started, and we got the answer either by asking (smart way) or by doing mistakes the finally asking (hard way).

Way back then, I built some circuits that had a "Vref" indication but no drawing of a voltage divider (because it was assumed it's a basic knowledge), I tied all the "Vref" points together and as expected the circuit didn't worked. Now I know the reason, but back then I didn't.

A few tips:
- gnd is usually labeled gnd (written), gnd (the gnd symbol), 0V or an arrow pointing downwards. This schematic uses 0V and the arrow pointing downwards. But be careful on symetrical power supplies (+9V and -9V, for example), the arrow may mean the negative voltage instead of gnd.
- the 9V power supply is usually labeled V+ (or +V), 9V, +9V, 9VDC or by an arrow pointing upwards. In this schematic, you can see both 9V and 9VDC. But take a look at this schematic (https://www.diystompboxes.com/analogalchemy/sch/shocktave.html), you see +9V by an arrow pointing upwards, and all arrows pointing upwards are also connected to +9V even they're not labeled.
- Vref, VA, VB, V1, V2, Vbias or "V something", or BA, B1, B+1 or "B something) are usually voltages based on some reference that's not the full power supply. In this case, you have two voltage references: VDD and VBIAS, which are both below the full voltage (+9V). If you check the schematic for a tube amp (like those ones http://revolutiondeux.blogspot.com/2013/04/trainwreck-amplifiers-express-liverpool.html), you can see at the power supply a B+1 voltage, then a resistor that drops the voltage a bit, then a B+2 voltage which is lower than B+1. This is pretty common for tube amps. For pedals, usually you'll see the full voltage and a "Vref" or "Vr" voltages (http://www.runoffgroove.com/tubereamer.html for example)

[Phend]

Thanks much rutgerv and marcos,  I missed the 9v in the left circuit. It being a 8.79v battery.
So using the entire circuitry,  sans 9v dc plug, one would connect all VDD as together, all VBIAS 3.88 together and of course all v0 together.
But as marcos mentions this might be a mistake, should I connect all VDD together except for VDD 8.56 (2) which would be connected separately .
(Being from the mechanical world, this circuit method reminds me of assemblies and Sub-assemblies.)

[Marcos - Munky]

But as marcos mentions this might be a mistake, should I connect all VDD together except for VDD 8.56 (2) which would be connected separately .

No, all VDD points are the same and should be connected together. the 8.56 is just to indicate the voltage reading Jonny got at that point, to help on possible troubleshootings. So let's say your build doesn't work, if you measure the voltage between ground and VDD you know you're expecting a reading around 8.56V.

This is not a so common but not so unusual practice. In fact, it should be done more times, so we all have more info on what to exactly look for on a non working build. But yeah, it indeed can be confusing when you look at it for the first time.

Btw, I should add something I didn't said yet. The circuit sounds amazing! I made a 1590A layout for it yesterday, and if things goes well it should be tested some time next week.

[cab42]

As mention earlier I did a quick vero layout. Currently, it is not verified and I maybe should have waited until I have built and verified it. However, the BOM and my parts collection seems to be completely out of phase, so I need to make a parts order first.

Please let me know if you see something that should/could be changed!



BTW, there are some really good and helpful answers in this thread, thanks to all of you!

[Steben]

One other thing I was wondering about: the 100 ohm resistor (R15) in the PSU input stage, does it serve any purpose for the dynamic response (sagging) of the rest of the circuit? Or is it purely for keeping out the PSU noise?

I put it there to get rid of a little bit of hum I was getting. I have no idea if the sag affects the sound or not

Class A stages. No dynamic current draw.

[teemuk]

^ This. I would expect no significant "sag", given moderately low current draw (preamps merely need to amplify voltage), constant current draw of class-A stages in average, low filtering resistance (100 ohms) and large-ish filtering capacitance (470 uF).

For example, to generate any worthwhile current draw to generate sag for FET preamp's power supply rails Roland Blues Cube amps had to feature a whole "mini" push-pull power amp driving a dummy load. Most preamp-based "sag" schemes feature somekind of envelope follower that modulates reference voltage for clipping.

While no considerable sag is produced using a power supply RC filter is a good idea for decoupling purposes so that signal does not couple from one gain stage to another via supply rail modulation. This circuit design does not decouple any of its four gain stages but does, however, isolate them from the variable impedance of the power supply. (Without the RC filter the increased source impedance of a dying battery might cause stages to couple through the supply rail).

[rutgerv]

I'm simultaneously working on some Amt jfet preamps and see they have a habit of adding a generic RC supply filter for the pedal, and (!) an extra RC supply filter for the first stage (perhaps it requires the highest clean gain?).

[Steben]

^ This. I would expect no significant "sag", given moderately low current draw (preamps merely need to amplify voltage), constant current draw of class-A stages in average, low filtering resistance (100 ohms) and large-ish filtering capacitance (470 uF).

For example, to generate any worthwhile current draw to generate sag for FET preamp's power supply rails Roland Blues Cube amps had to feature a whole "mini" push-pull power amp driving a dummy load. Most preamp-based "sag" schemes feature somekind of envelope follower that modulates reference voltage for clipping.

A loaded opamp can do this as well I guess.... I thought about those things before. Problem is: what do we want with sag? 5 a 10 % rail supply fall? Or induced crossover? Mostly the latter and that can be achieved easier. Recently I suddenly appreciated the fact VOX AC amps hardly have power sag given the rather low amount of average current rise. It's all about cathode bias shift there. And no one calls AC amp "not tubey"

[Phend]

So as to not interfere with jonny.reckless s circuit I made a simple connect the dots diagram.
Connected VDD and VBias and 9v added battery removed 9vdc plug connected input and output. These are just for Reference.  The image is light, you need to zoom and squint. So, as a check, are the connections correct ?  Thanks in advance...Paul

[Marcos - Munky]

Almost correct. Since you're using a battery, you don't need R15, so you can skip it and connect the battery directly to VDD. The other parts are all correct.

[Phend]

Thank you very much marcos and rutgerv.....got it, drawing my own circuit to work from sans R15. Monday plan on ordering the few parts I need. Will make Little Jim....special thanks to jonny.reckless for sharing advanced circuit and you all for some circuit design education. Also will post results or the need for HELP. LOL. (I will fill those hiss making holes in the mosfets with some silicone cement, that should fix that)

[rutgerv]

Happy to report I have a working version of this circuit on my breadboard as of today!

I'm not quite sure yet if it sounds exactly as intended though. Probably because I made some part substitutions (I didn't have all the original parts, but was still eager to try). I used 2N7000 MOSFETs and was able to find 3 with roughly matching Vth of 2.6V. I didn't have sufficient 4K7 resistors, but approximated them with parallelled 10K resistors (resulting in 5K), and didn't have 22nF capacitors, but substituted with two times 47nF in series (coming to 23.5nF). I noticed a couple of things in the sound:

- the overall Marshall tone is there and rocks! Especially with AC/DC-like riffs. I play it into a NUX Solid Studio IR loader (1960 cabinet) with Power Amp sim (EL34) enabled.
- when playing rather dynamically with neck strat pickups the sound goes rather suddenly from clean to distorted. It doesn't feel very natural.
- There's quite some noise, and turning the gain knob makes the noise change 'colour' as well. As if there's additional filtering happening
- the low end behaves a bit odd in terms of dynamics... Somehow I struggled to get a nice "chug chug" out of palm muted lower E-string riffs. It's worse for pick-ups with a lot of low-end, and becomes better when the sound is high-pass-filtered before entering the Little Jim.

I suspect a lot of the above has to do with my suboptimal choice of parts, and I'm intending to look at the individual gain stages with a scope tomorrow to see if it needs a change in bias.

@Jonny: can you provide some guidelines on what to aim for, when adjusting the bias on scope? How "asymmetric" should things be, to make it sound as intended?

Rutger

ps. In the configuration going straight into my NUX CabSim the tone control of the Little Jim is much needed, as it can be extremely bright! I also noticed quite a shift in the mid-range EQ curve when turning the tone control to maximum bright... is that intended?

[jonny.reckless]

@Jonny: can you provide some guidelines on what to aim for, when adjusting the bias on scope? How "asymmetric" should things be, to make it sound as intended?
Rutger
ps. In the configuration going straight into my NUX CabSim the tone control of the Little Jim is much needed, as it can be extremely bright! I also noticed quite a shift in the mid-range EQ curve when turning the tone control to maximum bright... is that intended?

The MOSFET drains should sit around 2/3 to 3/4 of VDD, i.e. somewhere between 6 and 7.5 volts is ideal. Looking at the signal on a scope with a 440Hz sinewave, the top half should be rounded and fat, and the bottom half should be hard clipped. Adjust R8 (VBIAS) to get the drains where you want with the MOSFETs you have. Also make sure the source of TR3 is sitting at at least 2VDC so you don't get unwanted hard clipping on the input stage.

Voicing the circuit is a highly personal thing and depends on the MOSFETs you used, your guitar, pickups, playing style, the EQ and voicing of your amp's clean channel, and choice of speakers.  Here are some ideas to play with.

• Reducing C8 will reduce the bass before the gain stages. This will make the distortion more crunchy and bright, and less fuzzy.
• Make sure you use an audio taper pot for RV1 to get a reasonable gain response. As I said in the video, it's not subtle, and doesn't do the tube screamer thing, it's a full bore high gain distortion!
• You can play around with the output filter comprising C10, R14 and C11. These are all interactive, you can simulate the response with LT spice if you like. Reduce C10 to get less bass. Reduce C11 to change the roll off turnover frequency.
  Add C4 back in to soften the high end a bit, start with 4n7 and experiment.

[deadastronaut]

rutgerv , have you tried it with humbuckers?

[jonny.reckless]

The problems with palm muting and chugging suggest the bias voltage is not quite right. My prototype chugs nicely. Once you have got the bias voltages set up correctly, you can also reduce the pre gain bass further by reducing the value of C6. Both C6 and C8 affect the low end response of the first gain stage, they behave slightly differently under clipping so it might be worth experimenting, maybe start by halving the value of one or both of them and see how it behaves in your circuit.

The output tone control starts to roll off at 70Hz and varies the point it stops rolling off, so you get a nice bass boost with the tone control in the middle. You can change this by reducing the value of C11 to 10n, this will start the roll off around 700Hz which is more what you get with a tube screamer tone control. Note that by doing this you will lose a lot of low end, but there's plenty of opportunity for tweaking!

[POTL]

This is actually an interesting idea.
At the moment, this is certainly not a marshall, but a good platform to start with, I have seen a similar one with Catalinbread RAH and Jack Orman Mosfet Booster.
Connecting gate resistors to bias is a good idea to get away from Zvex style BOR / BOM / SHO etc. biasing.
We have no feedback and no Miller effect, which means our high frequencies are safe.

MOSFETs have a number of advantages over the classic AIMP Jfet choice.
1) We can change the drain resistor and affect the gain, just like in the anode resistor of a lamp (JFET either works or not).
2) We can assemble a circuit without trimming resistors, which means we can simplify our work.
3) There is no need to select transistors
4) These transistors are not obsolete and can be easily found on the market, Jfets are gradually becoming a rarity, as are germanium transistors.

But there is nothing perfect in life and we have a number of disadvantages.
1) Transistors are afraid of static voltage (although some newer models have built-in protection)
2) Noise and this is the main problem, JFETs hiss and noise can be reduced or eliminated altogether, MOSfets make noise like a transformer booth and I don't know a real way to remove noise, they are just noisy.
3) Several years ago I made breadboards, but I could not make a really pleasant sound with high gain, it was loose and fuzzy, maybe this is a lack of knowledge in those years or the MOSFet is really unsuitable for high gain. But the clear sound is much better than the JFET.
4) I tested only ZVEX style circuits and Miller effect got bored pretty quickly, maybe your CB RAH / Jack Orman style circuit will be much better.

Now let's get back to your circuit.
We know that the sound of the preamp is formed in many parts of the circuit, you have not conveyed much
1) Tone Stack forms most of the sound, you don't have it, which means there is no mid cut
2) There is no follower, although the follower forms the sound, adds compression. Yes, you can also hear it on transistors, I checked.
3) Where is the proprietary filtering Marshall 470k // 470p? This is part of their sound.

[rutgerv]

Wow, just spent a nice hour of playing several guitars and pick-ups through the Little Jim into my IR loader + poweramp sim! Very inspiring! With the tips of Jonny I was able to figure out that all something strange was going on in the last gain stage, probably a bad connection on my breadboard. It's fixed now, and suddenly it has "the sound" I was expecting. It really rocks!

So please forget my negative comment made earlier. This was really a case of "user error". Palm muted chugging is nice now, though not "heavy metal" in any way, but more classic rock (a la Led Zeppelin's Communication Breakdown). I also love how easy harmonics come out on upper strings, and how much detail/movement can be heard on 3-note chords played somewhere in the middle register. As mentioned, this pedal isn't for subtle gain, and not for 'dynamic' playing. It's indeed amp-on-11 style, though I discovered there's an interesting sweet spot in the gain knob travel. First, the Little Jim starts distorting in a fairly 'normal' way, but beyond a certain point a second type of distortion comes in on top of the first, and that really adds 'richness'. Very soon after this point, distortion becomes uncontrollable for rhythm/chords and only still serves a purpose on leads, but it's that spot just before that I think is very much unique to this circuit!

Anyway, thumbs up Jonny! I think I', going to try and put this circuit on a proper PCB, instead of the breadboard. I noticed that near maximum gain the circuit was on the brink of internally oscillating, and a better circuit board/layout might prevent that.

Rutger