Gar Gillies' infamous STINGER adapted for stompbox use!

[whoisalhedges]

I've long been interested in obsessed with onboard tube fuzz circuits, like the Laney Klipp or the subject of my latest (freakin' sweet) pedal, the Garnet Stinger.

The Stinger doesn't really sound like anything else; or at least not anything affordable to most people. It's an enigma of sorts: full-voiced, throaty, yet articulate and never muddy. Smooth and clearly "tubey," yet raucous and unhinged (like the best fuzzes are). After about a decade of wanting one, I finally decided I had to have one. Trouble is, it's a pentode circuit built into an amplifier; and all I have are some transistors and 9v batteries....

Keeping in mind the imperfections of emulating tubes with FETs (and even that's typically been triodes), the prospect of nailing down the Stinger's pentode-derived fuzz was pretty daunting. Then I remembered ROG got a fairly decent EF86 sound out of a MOSFET, at least at high gain levels, in the Matchbox (and all the Stinger has is high gain levels); and while the Stinger used 6AN8s early, and then 6JW8s, that could be a starting point, right? Except I don't know doodley-squat about MOSFETs....

Well, why not just use the Matchbox's first gain stage pretty much as-is? I mean, this it gonna be a decent-sized build, I've got plenty of room to manipulate the frequency response - this is all about getting things going, presenting the rest of the circuit the right ingredients to mix up and cook into the hopefully-not-inimitable Stinger sound. If I clone an existing stage, at least I'll get the thing biased right, yeah?

Next is basic JFET "triode emulating" gain stage. Back-to-back LEDs to ground in front and no source bypass cap because I don't want it clipping. That leads into a Baxandall//James tone stack and a 2N5088 (or any reasonably high gain NPN BJT - hell, Ge would probably work so long as it's not too leaky) driving asymmetric back-to-back diodes to ground (I used one Ge and one Si) because hey, it's a fuzz.

Last but not least is a basic JFET mu-amp, I cribbed the circuit from GEO or AMZ, don't remember (they got it, as either Jack or R.G. freely admits, from National Semiconductor decades ago).

And you know what? I'll be damned if the final product, kicking a tube amp in the teeth, doesn't sound pretty much just like the Stinger. And just as impressively, it worked the first time I plugged into the breadboard (ok, second - I had put the safety diode between the MOSFET's source & gate in backward the first time)!

Clips will come, till then here's a schematic, and I hope y'all might have a couple pointers for me.



First: I've seen them as standalone devices, but I just wanna make sure: mu-amps are noninverting, right? I don't wanna spend hours on a vero layout, solder it up, and put it in a box just to have phase issues.

Second: other than the output (which I got from the mu-amp circuit), the pot values actually come from the original Stinger circuit - Gar seemed to have a thing for 1-Meg potentiometers - and, well, they don't seem to do much. The "color" control is subtle in an actual Garnet amp with the Stinger onboard, looks like a pretty basic variable HPF, sounds like a pretty basic variable HPF. It "tightens" or "loosens" the fuzz. It's subtle, but it works as expected.

The next three are somewhat problematic. The Stinger volume (the gain control) doesn't seem to do a whole lot of anything; and the bass and treble pots (I copied the tone stack from the amp schematic) are similarly ineffective. I popped the values in the James tab on Duncan's tone stack calculator, and yeah - it doesn't have the widest range of variability I've ever seen in a Bax stack, pretty much the opposite. And for that matter, the 100k output pot doesn't seem to do much - it's a loud circuit; if I turn it down, it's a loud circuit; if I turn it up, it's a very slightly louder circuit.

Any advice from smarter folk than I? As far as the tone stack goes, I could audition other cap, resistor, and pot values with Duncan's assistance to get a wider spread of available tones. I want the sound to remain faithful to Mr. Gillies' original vision (and golly gee, does it sound like a Stinger through my Laney AOR30 - and even, to a somewhat lesser degree, through my little SS bench amp) but if I'm going to bother putting 5 knobs on a pedal, I want them all to do something. Anyway, unless there's a Garnet Amps guru on the board who can just tell me what to do, I can figure the tone stack out myself.

The biggest thing I need advice on is the two gain pots. It would be nice to have more control over just how loud and how fuzzy the pedal is. I suppose I should increase the size of the master volume pot and decrease (there ain't many > 1M pots, after all) the Stinger volume rating, but to what? Often, gain controls aren't my biggest concern (why do they put volume controls on guitars, after all?) but this thing does NOT clean up with guitar volume. If there's any hope of taming the gain on that rare occasion I wish to, it's in the circuit. And output-wise, well, we all know how certain pedals are louder on certain amps than others - I want this box to sound FREAKING AWESOME whatever I plug it into.

So... thoughts?

[TejfolvonDanone]

Just a reading for the mu-amp: http://www.geofex.com/Article_Folders/modmuamp/modmuamp.htm

mu-amps are noninverting, right?

No. Mu-amps are just a common source amp with an active (current source) load. The common source as the common emitter and the common cathode are inverting amplifiers.

the 100k output pot doesn't seem to do much

The mu-amp needs a high load around 500k. If you put a 100k pot and plug into an amp (which has high input impedance but you can't be sure how high is "high" maybe it's just 100k.) the loading is too much for the mu-amp. So a buffer stage would be really helpful (second schem on the linked article). I think putting the volume pot after the added buffer stage would help with the ineffective volume controll problem.

If you connect the taper to the LEDs and the upper leg of the put to the outpu cap of the MOS stage things will get different. As you drew it the volume isn't really a volume. It just varies the resistance between output to ground and a series resistance to the LEDs. This way if the out-gnd resistance is big enough not to load the output of the MOS stage the LEDs will see almost a constant voltage. The series resistance just alters the current through the LEDs thus not changing the signal level going into the JFET.

Last: are you sure about that MOSFET choice? It is a power MOSFET. I would put in a BS107 because it is a small signal MOS and also smaller and cheaper.

[whoisalhedges]

Thanks.

Yeah, the IRF510 is what I had on hand. From the datasheets, I infer that a BS107 or 2N7000 will pop right in the circuit, and will be what I end up using for size and cost concerns. Ultimately, I'm after results, and the power MOSFET works, but isn't needed by any means.

Good to know about the mu-amp. The output is coming off the drain of the FET doing the amplifying, but the source of the other - so I wasn't sure how that worked. I have seen mu-amps used alone as boosts, though, why is that?  Are those for people who don't have any other effects? Ultimately, I guess it shouldn't matter so long as I just have one signal chain; but people sometimes run FX chains in parallel, and I don't wanna kill it.

So. Adding an inverting buffer (in the right place, obviously) could solve both my output volume control issues and  my phase issues?

I had actually thought of connecting volume1/gain/fuzz as you describe - it didn't look right to me as-is. Either I missed something in my conversion of the original scheme, or it's just another case of transistors not being valves. So, does the MOSFET want to see 1M? Should I add some series resistance to ground, either with the 1M or a different value? Any places you see that may benefit from audio taper besides the gain pots (or for that matter, any reason I may want to use a linear pot for either of the gain controls)?

[TejfolvonDanone]

I have seen mu-amps used alone as boosts, though, why is that?

The upper JFET is an active load. It is almost a constant current source. The output impedance of the upper JFET (as a current source) serves as the drain impedance. The gain of a common source amplifier is (almost) the ration between the source resistance and drain resistance. So gain of this stage is really high. There's a really big BUT - the output impedance of the common source amp is the drain impedance. So high gain = high output impedance. The loading of the next stage will alter the gain of the mu-amp. Even loads which doesn't really in most cases can do it with the mu-amp. That's why you often need a buffer after the mu-amp.
So high gain = good booster.
Also it has a unique distortion characteristics.

So. Adding an inverting buffer (in the right place, obviously) could solve both my output volume control issues and  my phase issues?

I'm not really sure it would solve the volume issue but it would defenitely help with the phasing.

Any places you see that may benefit from audio taper besides the gain pots (or for that matter, any reason I may want to use a linear pot for either of the gain controls)?

Maybe in parametric EQ-s because we observe frequency ratios rather than absolute frequencies. But for gain/vol the usually the audio is the best choice. I don't think you should change it.

Either I missed something in my conversion of the original scheme, or it's just another case of transistors not being valves.

The valume pot is independent of the device use. The volume pot in a guitar is connected as i described. I think you missed something in the conversion.

So, does the MOSFET want to see 1M?

The MOSFET stage's output impedance is almost the drain resistor. (It is altered somewhat by the feedback resistor.) You need a loading which is higher than that because a voltage divider will be formed by the output impedance and the input resistance of the next stage. So bigger the input impedance less the attenuation. As it stands now the attenuation (loading) varies over the turn of the pot. After a certain rotation from bottom to top this gets to a level where it doesn't metter.
With a traditional volume wiring the loading is constant and really high (no signal loss to the next stage) you only control how much signal is fed to the next stage. So actually it isn't a gain control but a volume control between stages but it is only a matter of terms used.

Should I add some series resistance to ground, either with the 1M or a different value?

With the schem you provided it would make the gain control completely useless.
With the other gain control wiring it would set the minimum gain (precicely the maximum attenuation).

[antonis]

I don't know how it sounds but from a first POV tonestak values seem "odd"..

Not only the ratio of R12/R11 & C10/C9 but the actual pots & resistor values - they drive an impedance lower than them (depending on pots setting..)
IMHO, a value reduction of 1/10 should be OK without changing Q2's Source resistor..

[amz-fx]

Try something like this:



The gain control and output volume need to be log (audio) taper. If they are linear, then the response will be all near one end.

regards, Jack

[whoisalhedges]

Try something like this:



The gain control and output volume need to be log (audio) taper. If they are linear, then the response will be all near one end.

regards, Jack

Thank you! I'll try that after work.

Should I add some series resistance to ground, either with the 1M or a different value?

With the schem you provided it would make the gain control completely useless.
With the other gain control wiring it would set the minimum gain (precicely the maximum attenuation).

Thank you, yeah - I understood that part. Was referring to whether or not I should try doing so after re-orienting the pot correctly as a volume control.

[TejfolvonDanone]

Thank you, yeah - I understood that part. Was referring to whether or not I should try doing so after re-orienting the pot correctly as a volume control.

If you'd like to have signal when the gain is turned all the way down then yes you should do it.

[whoisalhedges]

Try something like this:

Do you (or anyone else) have any advice as to a good clean PI stage to get my output back in phase with the input? Minimal parts count, transistor preferred (I have a hell of a time getting op amps to work as yet), and neutral - again, the sound of this thing is right on the money, I just need to fix the spread of the controls. I love how it sounds, I just want the controls to do something and for the circuit to play well with others.

[PRR]

Why does Q2 have a Drain resistor??

Q3's Base bias is drained away, varibly, through the tone stack.

R12 R13 R11 Bax values should be in the ratio 10:100:1 or 100K 1Meg 10K, not 100K 1Meg 68K. (That's where "variability" is lacking.)

C15 C13 are loading-down the Treb pot especially on Boost.

The whole scheme is too much to brain-plan. You need to breadboard it.

Phase is the least of your issues. In a single channel, absolute phase is nearly inaudible. When mixing processed and direct signals, phase can cancel only if the two paths are nearly the same amplitude; here this path is sure to be very different than any other path you mix it with.

[whoisalhedges]

Why does Q2 have a Drain resistor??

Q3's Base bias is drained away, varibly, through the tone stack.

R12 R13 R11 Bax values should be in the ratio 10:100:1 or 100K 1Meg 10K, not 100K 1Meg 68K. (That's where "variability" is lacking.)

C15 C13 are loading-down the Treb pot especially on Boost.

The whole scheme is too much to brain-plan. You need to breadboard it.

Phase is the least of your issues. In a single channel, absolute phase is nearly inaudible. When mixing processed and direct signals, phase can cancel only if the two paths are nearly the same amplitude; here this path is sure to be very different than any other path you mix it with.

Well, in that case, perhaps phase won't be my biggest concern - the amplitude, yes - also I can't for the life of me imagine what I'd *want* to run parallel to this.

It is breadboarded, that's how I know what it sounds like. I don't break out the iron till long after I run something through its paces.

The Bax values... I've never seen anything like it. I copied the tone stack exactly from the original circuit. So it's a bad tone stack to begin with.

As to Q3 - whoah, yeah. Even my beginner eyes should've caught that. So what, you think I should figure out best as I can where it sounds best (after improving the tone stack), take a DMM reading on the base, and then plug in a voltage divider to hold the bias at that point? Unless, of course, I for some reason want it that way....

[whoisalhedges]

Last but not least is a basic JFET mu-amp, I cribbed the circuit from GEO or AMZ, don't remember (they got it, as either Jack or R.G. freely admits, from National Semiconductor decades ago).

Oh hey, it's the AMZ mini-booster (more or less)! Now I remember where I got it!

[Gus]

None of the plans so far look good to me
First work out what is going on in the stinger preamp.  I made the mistake of looking at the schematic of the stinger a web search finds. I suggest you post a link

Figure out the gain the output resistance and headroom(B+ voltage related) of the two upper triode gain stages

Next the pentode triode stage note the 200k in the cathode circuit that gets shorted to ground to add the distortion
It looks like contact bias for both the pentode and triode Note the 10meg and .005uf at the pentode g1.
Contact bias has its own sound when overdriven
The plate out from the triode goes into a tone shaping network before getting mixed back in at the top right triode

I think you can do this better with BJTs or opamps.because you can control the gain better when you scale down the voltage you need to scale down the gains

[whoisalhedges]

None of the plans so far look good to me
First work out what is going on in the stinger preamp.  I made the mistake of looking at the schematic of the stinger a web search finds. I suggest you post a link

Figure out the gain the output resistance and headroom(B+ voltage related) of the two upper triode gain stages

Next the pentode triode stage note the 200k in the cathode circuit that gets shorted to ground to add the distortion
It looks like contact bias for both the pentode and triode Note the 10meg and .005uf at the pentode g1.
Contact bias has its own sound when overdriven
The plate out from the triode goes into a tone shaping network before getting mixed back in at the top right triode

I think you can do this better with BJTs or opamps.because you can control the gain better when you scale down the voltage you need to scale down the gains

There were two versions, mine was based more on the second (page 2):
http://www.thermionic.info/amps/garnet/BTO.pdf
Though I did begin with this whole preamp as the starting point, it's not meant to be a 1:1 conversion - I don't trust that transistors are ever going to behave exactly like tubes; my goal was the sound and (to my ears, anyway) the sound is there... except, as noted in my original post and some of the reasons why explained by several others who know a lot more than I, it's only one sound. My circuit isn't correctly designed & built to have a properly-functioning tonestack or volume controls.

I'll definitely be applying a lot of these suggestions on my breadboard - probably not tonight, I work Friday-Tuesday from 5-1:30, so today was my "Monday" and I'm tired. But even so, I'm champing at the bit to get this doing more of what I want it to.

I mean, maybe I completely missed the mark and just got lucky with the one sound my build makes as-is. Maybe once I can control the tone & gain, that'll be the only usable, Stinger-like tone that comes out of this thing - and if so, hey, learning experience. But in any case, I remain pretty stoked that the biggest circuit I've put together so far a) made sound straight away when I put it between a guitar and an amplifier, and b) made the sound I wanted it to, far beyond my expectations.

[Gus]

There is an interesting link in reply 3 of this thread http://www.diystompboxes.com/smfforum/index.php?topic=114035.msg1057516#msg1057516

I looked at the 2nd page of the link for the other stinger circuit

If I was to do a 1:1 at 9VDC I would first try x10 gain stages for the triodes this depends on what signal level at the input. However note the pentode wiring  no DC path(except leakage in the tube body, tube socket)to ground

It might be interesting to build something like this with 3 opamps.  Search this site for  wording like "scaling tone control values" you can change the values of the RC networks keeping the same RC but at lower resistance, higher capacitance to a point.

[amz-fx]

But in any case, I remain pretty stoked that the biggest circuit I've put together so far a) made sound straight away when I put it between a guitar and an amplifier, and b) made the sound I wanted it to, far beyond my expectations.

Good for you!  It is a great feeling to build a circuit and get it working, and even more so when the sound is to your liking. When you are pleased, that's all that counts.   Tweak it a little bit, then think about what you want to build next!

Best regards, Jack

[whoisalhedges]

OK!

I breadboarded Jack's suggestions, mostly to great success.

Also adjusted the tone stack: changed R14 to 68k and halved R15/treble pot to 500k. Left the other odd values in place for now, will probably adjust further for a wider spread, but this is a step in the right direction.

A few concerns: now that the R7 volume pot is hooked up correctly, it does control the volume - but now somehow the gain has gone up even higher, and I can't for the life of me figure why. As per Jack's suggestion, I changed Q1's feedback and bias resistors from 1M to 10M... but that shouldn't do that, should it? And turning down the volume control does turn down the volume, but doesn't tame the distortion as much as I'd hope - maybe I should move the pot to the other side of the JFET? The LEDs keep the FET from clipping, but I want the BJT and diodes thereafter to clip, so perhaps I'll try it that way? As my goal is the final sound of the thing, not a piece-by-piece structural replication, I only care where the pot is because of what it does.

AND, for that matter - my ears could be playing tricks on me. It might not actually be any fuzzier; and if it is, maybe it's because my initial incorrect orientation of the pot attenuated the signal. Perhaps I could pop in some series resistance. In any case, that's what breadboard are for.

A couple questions as to the whys and wherefores of Jack's changes.
1) why the 10M resistors around the MOSFET? Do 1-megs not cut it for biasing stability or static protection?
2) The 100k resistor in series with the output pot - I understand the theory of putting a resistor between the pot and ground (which isn't yet in my circuit) to keep it from being turned full-down to silence; but this one is between the pot and the circuit. Is this a trick to prevent the mu-amp's nasty habit of increasing impedance along with increasing gain?

I'll continue to massage the tone stack, I'm by no means done. Thanks for everyone's help so far!

[amz-fx]

A couple questions as to the whys and wherefores of Jack's changes.
1) why the 10M resistors around the MOSFET? Do 1-megs not cut it for biasing stability or static protection?
2) The 100k resistor in series with the output pot - I understand the theory of putting a resistor between the pot and ground (which isn't yet in my circuit) to keep it from being turned full-down to silence; but this one is between the pot and the circuit. Is this a trick to prevent the mu-amp's nasty habit of increasing impedance along with increasing gain?

1. The input impedance of the shunt feedback input stage is less than you think since it is divided by the gain of the circuit. The 10M gives you a higher input Z so you get more of the guitar tone. Might not be what you like but worth trying out.
2. The extra 100k is to lower the output volume since you said it was too loud. With the new resistor, the output level is max 50% of what it was originally. You could change the volume pot to 10k to load down the jfet stage but that might impact the tone - however, you might try that instead of adding the 100k resistor on the output.

regards, Jack

[whoisalhedges]

Haha sorry, that was my mistake - I didn't mean to imply that it was too loud overall, just that it wasn't variable enough. It was either loud or off, which is probably owing to several things, but most of all that there's a pretty good amount of gain in the circuit and the fact that I've got little linear trimpots on my breadboard. A proper logarithmic-taper pot will probably help that.

I'm still trying to digest some of the copious material out there about output impedance concerns with mu-amps.

[whoisalhedges]

As of this evening...


Once the trimpots are replaced by pots with a proper taper....

I'm preferring the 1M resistors on the MOSFET - with my guitars, and with no other effects preceding the circuit. More testing is necessary. Also, perhaps something in-between, a 2.2M or 4.7M, might satisfy my requirements and sufficiently boost input Z...

The tone stack, particularly the treble network, is still unorthodox but sounds much better. I plugged it in Duncan's calculator with reverse log pots, and it has a really neat mid scoop with bass & treble full up, and a mid hump with both down around 1-2. And of course, settings in the middle offer a wide range of different frequency plots. Even the linear trimpots sound good at the upper range of their travel.

I'm not 100% sure which direction I'll go in first as far as adjusting the level of distortion. As some have pointed out, the first volume control isn't necessarily a distortion pot; and looking at my design, it's right in front of a JFET that I've structured specifically not to clip. Right now, it really acts like more of a - well - volume control. And I don't need two of those. Although these definitions are pretty subjective, with the 1M resistors on the MOSFET, it's a lowish-gain fuzz/fuzzy overdrive... like the Stinger. Doesn't get quite as fuzzed out as the original, but gets most of the way there. Meanwhile, with the 10M resistors, it's a lot tighter, more what I think of as a distortion - I understand Jack's explanation that this is for input impedance, but it seems that something in the biasing or in the feedback of that MOSFET is changing, and not in a (for the sound I'm chasing) desirable way.

So, it seems to me that this is my last major hurdle. There'll be fine tuning, sure... but I feel like I'm around 90% of the way there, and a gain control that will back it off a little from where it is, and also jam it up a little more into gritty glory, is the missing piece.

So:
I have a couple of thoughts, neither of which is perfect.
1) The good old input control. Is most of my dirt coming from the MOSFET? If so, slamming it or conversely teasing it could be the way to go. Of course, if we're walking an input Z tightrope as is, it mightn't be the best idea to play fast & loose with impedance there. Especially, oh especially if it won't make it sound any better.
2) Before the BJT. After the tone stack, of course. Even with a calculator, I'm having a hard enough time with that. But the 2N5088 is driving a back-to-back pair of clipping diodes. Now that, that could sound tasty

So, does either one sound preferable? Or maybe something I haven't considered? The way I figure it: the MOSFET increases gain quite a bit. It's an effective signal booster with its own style of distortion. I don't understand MOSFETs enough to know whether or not it's contributing a great deal of distortion in this circuit. The following JFET has limiting diodes (the LEDs) at its gate and no source bypass cap - so it's pushing the signal along through the tone stack, but it's not clipping. I might be way off base technically, but I see it as analogous to a cathode follower in a tube amp. After the tone stack is the BJT and its clipping diodes. Definitely some fuzz being generated there. And finally the mu-amp, shaping the final tonal characteristics before sending the signal along to the next effect or the amplifier.

There are three main sources of audible gain, then: the MOSFET, the BJT, and the mu-amp. There are also, in the LEDs and the clipping diodes, two limiters. Somehow, in that mess, there has to be a place to put a simple, effective gain/fuzz control. Right?