Converting Smoky to a more useful 'Little Gem'

[Peter Snowberg]

One more note: If you want some free and easy to use schematic drawing software, check out http://www.expresspcb.com/

They also make great boards.

-Peter

[javacody]

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.

I'm not sure what you mean by that Peter? Honestly, wired as it is, it sounds very good. It sounds very Fendery. I'm going to remove the 10uF, add the 100 uF, the 470 ohm resistor, and change the power filter Cap on the amp to 470 uF and see how it sounds.

[Peter Snowberg]

I was talking about booster topologies mainly....

With the circuit you posted, you have the option of shorting the trimpot out and then adding bias via the gate. If you then wire a 100K pot across C2 with the wiper connected to a 10uF cap to ground as well as to the now grounded end of the grid (whooops...), er uh, gate load resistor (removing the ground connection to it), then you have essentially the Dr. Quack buffer stage.

It's all about what sounds best to you. There is no single "right" way. It's a lot like cooking.... let's see.... what do I have in the fridge.... a pinch of this and a dash of that.... It's more about the method rather than an exact recipe.

Happy cooking!
-Peter

[javacody]

Cool. Thanks Peter.  By the way, I just finished making the changes you suggested. Here is what it looks like currently:



I have to say that it sounds even better now! It reminds me a lot of my old '73 SF champ, except with a little more sparkle. Even without a tone stack, it is an awesome little amp. Will I stop here? Heck no! Well, for tonight, yes.   :lol:  It's time to sit back and enjoy the fruits of my labor (and your help of course) and play a little while I watch the snow fall outside. Peter, if I ever have another kid, I'll name him after you.  Thanks for your help.  

By the way, ExpressSCH is awesome! It took me ten minutes (being brand new to the program) to throw that schem together! That program should be listed in the FAQ somewhere!

[Peter Snowberg]

Congratulations!

That looks great. I hope you can post some pics of your build.

Once you play with the tone stack in there, you should also switch back to the cap for a little and compare. You might want an a/b switch to bypass the mid-cut inherent in the stack for a different tone.

Take care,
-Peter

[javacody]

No pics of the build until its all the way done. I'm only half way there.  

Here is what I think it will look like with the tonestack, please let me know if I'm correct:



I'm not sure if the way I have the bottom connection of the tone stack looping around into the wiper of the treble pot is correct.

[Peter Snowberg]

Three changes....

(1) You can get rid of the 1M R(load) as it's not needed. In the TSC, that value represents the load from the volume control. You could always add one by making R(load) a pot and connecting the wiper to the 386, but it isn't needed. You already have one on the guitar. In the simulation stage I should have said something about making R(load) 50K which is the input impedance of the 386. That doesn't really change any frequencies, but it will reduce the level. You already have plenty so no worries there.

(2) Eliminate the wire from the junction of the slope resistor and treble cap to the bottom of the mid pot

(3) Connect the bottom of the mid pot to ground.

Take care,
-Peter

[javacody]

Does this look right? I'm having a hard time picturing how I'm going to wire the tone stack. Everthing seems so interconnected.

bufferedgem_tonestack_v1.2.gif

I need some help. I'm becoming obsessed with this.   :shock:

[Peter Snowberg]

I need some help. I'm becoming obsessed with this.   :shock:

uh oh.... you hit the threshold. Just remember to resist treatment.

Does this look right? I'm having a hard time picturing how I'm going to wire the tone stack. Everything seems so interconnected.:

That looks perfect! Now that you have it drawn out... just drill the holes, set the pots in place, and use the component leads to connect the dots. Each time you install a part, trace over it with a highlighter.

Start with connecting the wipers to the ends of the pots for the mid and bass, then connect the ends of the pots together, then install the two .022uF caps and twist the opposite ends, then install the pot end of the 330pF, then attack the other end of the 330pF to the 100K, then attach the other end of that to the .022s, then add the input wire and the output wire and ground wire. Done. It's really easy to leave the connections floating in air and just bend the part leads so they're firm as a 3D sculpture, but you can also add a two position terminal strip for the connections to the 100K. I think the terminal strip is more trouble than it's worth, but either way is fine.

Draw it on paper first and then dig in.

Take care (and don't neglect the family too much),
-Peter

PS: If you have a kid that's into guitar, consider building one of these things with them.

[javacody]

Peter,
    I was thinking about classic Fender amps, and don't they have at least two  or more preamp stages?  Would adding more buffers simulate this more closely?

I was looking at v1 on the Sovtek Mig50 ( supposedly a Bassman clone) and noticed some similarities to this buffer.

http://heysam.com/schem/sovmig50.pdf

I noticed it has 3 preamp tubes, not sure what exactly v1 is for (inital buffering?), but v2 looks like it is driving the tone stack? And what does v3 do? Any good sites that will explain this stuff to me in fairly plain language?

[javacody]

OK, well taking what I've learned from this process, it looks like v1 is split between the two inputs and there is a pre gain pot attached to them? This is then routed through each half of v2, which then goes to the tone stack, I'm not sure I understand this? Would it be worthwhile for me to duplicate this in my little amp?

[Peter Snowberg]

That's a pretty classic design; now referred to as a Marshall.

Look at the different cathode resistor and cap differences between the two input stages. That's called voicing. The top one will have more bass because of the larger cap. Each one has the tube biased a little differently.

The third stage the signal hits is called a cathode follower. It drops the impedance to drive the stack. A 12AX7 has a much high impedance than your JFET setup.

The third tube is a phase inverter which makes two signals to drive the output tubes. One is in phase with the input and the other is inverted. One signal is used to push while the other pulls. This is all taken care of internally by the 386.

Sorry, I don't have any sites off the top of my head and I'm in a bit of a rush now. Try a google search, I know there's a lot of into out there.

If you want to simulate the tube amp more closely, you'll also need to loose the 386 and that gets into a whole new territory.

Take care,
-Peter

[javacody]

Wow Peter. I could pick your brain on this stuff for days (I already have). I'll do some searching. Could you give me a little hint about losing the 386? Or is that what you were talking about having in your head for the last 3 years?

[Peter Snowberg]

Your description is correct.

If you want monster distortion you could build the same sort of thing, adding two additional JFET stages, but there is no need. That's a very different amp all together.

All (most) of your gain is provided by the 386. The JFET is there to give you a high impedance input. In the tube amp there are two high impedance gain stages followed by an impedance dropping cathode follower and then the stack, phase inverter and output stage. Totally different animal with different parts and needs.

Take care,
-Peter

[eliktronik]

Hey, nice looking ckt. I just put it together and it sounds really good. I was wondering whether anyone has put together the MkII little gem w/ 2 lm386's. I recently tried it (to get some more power) and it sounded awfull. Anyone else try it? I might put together a lm286 ckt w/ a tda2030a (pulled it out of my old ss amp) after it to get some more power...

[Peter Snowberg]

Wow Peter. I could pick your brain on this stuff for days (I already have). I'll do some searching. Could you give me a little hint about losing the 386? Or is that what you were talking about having in your head for the last 3 years?

The neat thing about the 386 is that it takes care of a building block that gets rather complex and painful to do with individual parts. Tubes have the inherent advantage of biasing much easier. I won't even touch discrete transistor driver stages for the most part. The commercial stuff I have is based on the JRC386BD and other "single chip" power amplifiers.

If you want to build the whole thing from scratch, it starts to look like this:

Full size image here: http://www.lcaudio.dk/ulcamp1.gif
More stuff in this zoo: http://www.diyaudio.com/index.php

I'm almost out of my territory at that point. I just don't have the time to re-invent the wheel when the integrated chips do a better job than I could, plus they generally add things like thermal shutdown.

There are lots and lots of power amp chips around. Unfortunately the Philips site is very poorly indexed for searching, but try this link:
http://www.semiconductors.philips.com/markets/mms/products/analog/product_catalog/signal/index.html

I hope that helps,
-Peter

[javacody]

Sorry, I got ahead of myself there. I'm really just trying to understand the principles behind each of the three components that I'm working on thusfar.

Also, wired up the tone stack, but I think something is wrong. Only the treble is affecting the tone at all. It seems to act as a volume/treble bleed. Any ideas what I did wrong? I'm positive I followed the schem above. Did I make a mistake with the schem? Is the signal only supposed to go to the amp from the wiper of the treble pot?

[javacody]

Found the solution, see http://www.diystompboxes.com/sboxforum/viewtopic.php?p=93639#93639