Converting Smoky to a more useful 'Little Gem'

[javacody]

Peter, would a gain of 20 be ok for the reverb? I could just leave pins 1 and 8 open so that they use the lm386's internal resistor.

[javacody]

Wow, that GGG link has some good links off of it, namely, this one which decodes the Accutronics reverb tank codes.  http://ampwares.com/reverb_codes.htm

I also found out just about everything I'm going to  need to know about building my own reverb at http://members.tripod.com/~roymal/reverb.htm

[Peter Snowberg]

Peter, would a gain of 20 be ok for the reverb? I could just leave pins 1 and 8 open so that they use the lm386's internal resistor.

Yes.

The easiest way to test is to just connect your speaker to the tank send. Adjust the dwell pot so that you only hear distortion when you really bang on the strings. That should do a decent job. The final levels will depend on how much boost you have in the JFET stage and much output your pickups have.

Any idea what the dimensions are on the particular tank you pointed me to?

....grabbing a tape measure.... They're 9 1/4 x 3 3/8 x 1 1/2

I'm glad you found that reverb article. That's a great one. If you use an opamp to drive the reverb, keep in mind that you will need one with an isolated send jack. Also The only chip I would use in that case is the NE5532 as it's quite happy driving low loads with lots of current. It's my favorite all-around opamp too.

Once you're done, I hope you can post some pictures. That's going be a really smoking little amp. Just wait until you plug it into a 4 x 12 cabinet.

Take care,
-Peter

[javacody]

Ok, the AMZ site is back online and I have a buffer question.

In the following:

Before we see the circuits let us look at a circuit fragment that may be required for some of the buffer variations. As shown to the left, the resistor/capacitor network provides a reference voltage that may be used to bias the transistor or opamp into the best operating range. The point marked "Vr" is connected to the point also notated as "Vr" on the buffer schematic. If there is a reference voltage already established in a circuit to which you are planning to add a buffered input, the existing Vr can be tapped and used for the buffer's reference.

Is this neccessary? What are the drawbacks of leaving this out? Will putting it in change the sound of the buffer at all?

[Peter Snowberg]

If you use an opamp for buffering and you run it from a single 9V battery (a single-ended supply), you need a way to 'park' the opamp in the middle of it's operating range. That's what that circuit does.

If you run opamps from a bipolar (+9 / -9) supply, the opamp inputs that don't have signal applied to them are usually tied to ground (which sits 1/2 way between the + and - power), either directly or through a resistor. That circuit simulates the bipolar supply ground when you use a single-ended supply which doesn't have that reference point that sits 1/2 way between the other two.

If you run opamps from a single ended supply and you leave that Vref or Vbias generator off, the opamp won't pass any signal at all.

Take care,
-Peter

[javacody]

Thanks Pete. Picked up the rest of the parts for the buffer and the tone stack at the shack today. They didn't have any 330pf caps, so I had to buy an $8 assortment. Oh well. I'll probably never need to buy small value ceramic caps again.  

I also am taking your advice about  the powersupply cap. Hey, I sprung a buck for a 470 uF cap.   :wink:

Last but not least, I found this http://www.diystompboxes.com/sboxforum/viewtopic.php?t=11576&highlight=mpf102 when searching the forum on MPF102 JFETS. Very interesting, so I picked up a 10K trimpot for my drain resistor. I also printed out a bunch of RunoffGroove's small PC Board templates so I can play with a layout. I will post it when I get done. I'm doing the buffer tonight and the tone stack tomorrow night.

I figured to keep things simple, I would use a modular layout, with one board each for buffer, tone stack, amp, and reverb. That way I can tweak any one of them without hurting the others. I need some kind of easy connectors to connect the modules. Any suggestions?

[Peter Snowberg]

Cool.

That was a good post about biasing. I would advise adding a 1K in series with the trimpot so you don't accidentially let the blue smoke out of the JFET.

Since these are circuits that you're going to have connected in a case, I would recommend just soldering them together. Modular construction is a really good idea, but the blocks really have to be tweaked to work in conjunction so they're not as modular as something like a Moog. There aren't as many easy connector options when you're talking about small spaces and many of the smaller connectors don't have convenient 0.1" pin spacing. It might be easier to just little wire nuts while you tweak the circuit. They work great.

For the tone stack, I just build it directly on the pots. It makes for a smaller and cleaner build and in high gain circuits, it makes for a cleaner signal environment since you have less wire running around. Just my $.02.

One more thing to test out is a diffuser over the speaker. Weber sells nice ones, but you can make something fairly effective by taking an old CD and suspending it over the center of the cone. Give it a try. (they also work well as coasters)

Take care,
-Peter

[javacody]

OK, I'm officially frustrated. All I'm getting out of my buffer is a constant tone, a very, very annoying constant tone. What am I doing wrong?

I used the schem here:

http://www.till.com/articles/GuitarPreamp/

I left out  C2 because it says it isn't needed, and I only had a 10uF for C1, so that's what I used. Help!!  :oops:

[javacody]

Is the schematic in the link above wrong perchance? I ask, because I switched C1 from the drain to the source, and viola, it worked. This is how Jack Orman has all of his buffers.  Is the schem in the link known to be wrong?

[Peter Snowberg]

I'll bet it's the lack of C2 in there causing the oscillation.

As the 386 draws current, the power rail gets pulled on. If you don't have a well regulated source of power for the buffer, any noise on the power rail will show up in the buffer output. If that signal is out of phase, you have just created a really fast version of the circuit that runs a Fender tube tremolo. The other name for it is a "phase shift oscillator".

If you switch the resistor back to the drain and then switch what you have applied to pins 2 and 3 of the 386, it should also go away. That will reverse the phase just like swapping the drain and source.

In either case, I would not only add C2, but you might also want to put a resistor in there between the 386 and the buffer. Try something like 470 ohms. You will want to re-bias things to make the drain voltage roughly 1/2 of the voltage you see across C2, or a hair higher. You might even want to make C2 a bit larger like 47uF or even 100uF. You need solid power in there or it's going to whine at you.

If you look at tube amp schematics, you'll see that they have several high voltage supply outputs, with the most filtered one feeding the first tube. You need to emulate that topology.

What resistor values did you finally end up using?

[Narcosynthesis]

not too sure about any mods or anything for a smokey

but i can say the biggest improvement would be to upgrade the speaker

i have one and with the speaker it comes with it basically shounds shit, all buzzy and nasty
so i tried plugging it into the 12 inch speaker in my marshall and WOW what a difference, it turned it from a piece of crap into something pretty nice and very useable, next time i go to a rehersal place im gonna try it out with a 4x12 to see what it sounds like

David

[javacody]

Pete, I used a 3.3M to set up my input impedance to the gate, went with the 2.2K on the source and I'm not sure what the  resistance is to the drain, but I have to turn it almost off for it not to howl (I'm guessing 1 to 2 k?)

What does altering C1 do to me? I went with a 10uF instead of 4.7uF.

So. let me spit back out the instructions you just gave me so I can see if I understand. First, I'm going to add back in C2 as a 100uF cap, next, I'm going to add a 470 ohm resistor between the buffer and the amp.  Last, I'm going to tweak the trimmer on the drain to rebias the FET?

Does this sound right Peter? Once I do all of this, can I put C1 back on the drain without changing my input to pin 3 on the 386? Are there any ill effects of switching the phase like I unknowingly did?

I have to admit, even with the buffer working like it is, I have my sparkle back, which is a good thing.

[Peter Snowberg]

You have the sequence correct. The best way to bias when you know you're close on the meter is to switch to ears. Just to make sure on the 470 ohm... that's between the + side of C2 and pin 6 of the 386.

With stable power, you won't get oscillation, but with the circuit you have now, it's inherently resistant to oscillation which is a good thing for sure.

The best place for C1 will depend on the resistor values you end up with. One side will have more gain than the other. I would pick the side you want based on boost level, and if that means C1 on the drain, try swapping pins 2 and 3 for an inherently (more) stable circuit. No difference with phase here. Either is fine.

The larger value of C1 will allow for more bass. If you have too much bass, reduce the value by ear. Take it too far down and you have a treble booster. You could use a 3 position toggle or a rotary to switch multiple caps to taste if you wanted.... more and more options. I like the 3 position toggle with a small cap for C1 and the option to select either of two larger ones. That's a great cheat if you don't want a full tone stack.

-Peter

[javacody]

Peter, I've thrown together a schem of what we've talked about so far. Does it look right to you?

[B Tremblay]

The 100uF cap from 9v to ground in the Tillman stage is redundant since you've already got one in the Little Gem.  Also, the 10uF output cap of the buffer can be omitted.

I'm not sure how the 51k resistor will affect the circuit.  The Tillman stage has a bit of boost and does sound good, but I'd use the Dr. Quack input buffer instead (with both 1M resistors increased to 2M2), to keep the level at unity and for increased headroom before the amp.

Here's the AMZ buffers article, the Dr. Quack one is the third from the top.
http://www.muzique.com/lab/buffers.htm

[javacody]

Or I could probably just duplicate the Little Gem MKII buffer, right?

Why omit the output cap of the buffer?

[B Tremblay]

Or I could probably just duplicate the Little Gem MKII buffer, right?

Why omit the output cap of the buffer?

Yes, the MkII buffer would work fine.  The Tillman stage 10uF output cap is directly followed by the .01uF input cap of the amp.  Both caps are present in their respective circuits to block DC voltage (and shape the frequency response in the Little Gem).  You don't need both, so just use the .01 cap so the frequency respose is the same.

[Peter Snowberg]

If you're using a tone stack, that will also provide DC bocking. If you're not, the 386 has in imput impedance of roughly 50K to ground. With a value that high and your other resistors being just a couple K it won't really interfere much. It will just adjust the bias voltage a tiny bit.

Two schools of thought here....

The Zen side says unnecessary component and electrolytics eat tone. I must agree.

The academic side says modular building blocks and you only want to pass AC between those blocks. Nothing wrong there either. This also gives you the ability to shape the tone at one more place. It has good and bad sides.

As Brian mentioned, you don't really need to boost what the 386 is seeing unless you want mega distortion, but adding a tone stack will eat a LOT of the signal. You may want it boosted in that case.

Yes, and if you have the 51K in there from the Tillman schematic, you can omit that for sure.

Ahhhh.  now your schematic has loaded.

Yes.... remove that 51K and without the stack in there you should have either the .01uF or the 10uF, but not both. Play with that cap to shape the tone. I didn't know you had the .01uF in there. Also remove the connection from the 9V to the buffer's 100uF cap. You want the power to be sourced from the 386 pin 6 through the 470 ohm for decoupling it. I would still keep C2, but with the non-inverting buffer the Zen camp says unnecessary. The whine without it and with the inverting buffer shouldbe evidence enough that additional filtering is necessary.

Take care,
-Peter

[javacody]

The tone stack is going in next. I figured as a beginner, it was easier to get each seperate section up and working before moving on to the next piece. Once I get all the bugs worked out and have finalized my components, then I will do a layout and etch a board.

Given the fact that the tone stack IS going in, what components should I change and what would stay the same? I'm assuming I should leave the 10uF cap.

Here is the schem as it stands with Peter's latest recomendations.

buffered_gem_v1.1.gif

[Peter Snowberg]

For a booster to amp stage design, the only change I would make would make would be to eliminate one of the two coupling caps connecting the two sections.

For now, drop the 10uF and play with the value of the .01uF. Try using the amp some with .01 in there, then add that 10uF back in parallel with the .01uF and play that for a while. Also try values like 0.1uF and 1uF. That's four orders of magnitude for comparison.

When you add the tone stack, add it in place of the coupling cap. The purpose of the cap is mainly to stop DC from flowing out of the buffer and into the 386. The stack will perform the same function. Without the stack in place, it's advantageous to use a smaller value cap to limit the bass. Think of it as a bass control turned down with the treble all the way up.

At this stage it's a good idea to play with booster options too. Try out as many circuit configurations as you have parts for.

Take care,
-Peter