Is this correct (or good enough already)

[zener]

I found an amp schematic from GGG. It's a 8 watt amp with TDA2003. I'm going to make one with a 8" 8ohm 18w speaker. I added a Baxandall tone stack.

Do I still need to add a buffer here? If yes, is it before or after the tone stack?

Thanks.

[B Tremblay]

You'll probably want to set it up with an input buffer, then the tonestack, then another buffer, then the power amp.

[Peter Snowberg]

That tone stack wants a high impedance environment. It won't work between a power amp and a speaker.

You might want to think of using something like the Ruby from http://www.runoffgroove.com to drive the tone stack. Even the JFET buffer on the front end may be enough.

Take care,
-Peter

[zener]

Thanks B and Peter

I'm quite still a newbie. I'm still puzzled on what (or how) am I going to build for my practice amp. My wooden chassis will be out tomorrow from the shop.

Anyway, I would list things that I have and things I want (only if possible). These might help you to help me.

THINGS I HAVE:
As I said, I have a 8" 8ohm 18w speaker.
I readily have a Little Gem.
The 386s I can get are those with N-1 only.
I have the TDA2003 and I can readily build the power amp.
I only have two J202 (quite expensive and the store is so FAR :? )
I have BJTs like 2N3904 and 2N5088

THINGS I WANT:
I really want to go as loud as possible as long as my speaker (and anything) can handle it.
I want my amp to have at least a treble and bass control. A switchable drive/clean would be great.

My initial plan is to have a Little Gem as a preamp, the pin 1 and 8 of 386 would be wired as switchable drive/clean. Then, the TDA2003 power amp and a tone stack and buffer/s somewhere, I just don't know exactly where. What do you think about that :roll: ?

I hope I'm not asking for something difficult.

Thanks a lot

[Peter Snowberg]

My initial plan is to have a Little Gem as a preamp, the pin 1 and 8 would be wired as switchable drive/clean. Then, the TDA2003 power amp and a tone stack and buffer/s somewhere, I just don't know exactly where. What do you think about that :roll: ?

That sounds like a good idea to me.

You might want to use one of your JFETs and make a Ruby circuit instead so that you get clear notes on your clean setting. The low impedance of the LM386 (only 50K) will cut the highs off when used without a buffer which is good for distortion but not for clean tones. Place the tone stack between the 386 and the TDA2003 and there you go!

Buffer -> LM386 -> tone stack -> TDA2003 -> speaker

Good luck! That should be a sweet little amp.

Take care,
-Peter

[zener]

You might want to use one of your JFETs and make a Ruby circuit instead so that you get clear notes on your clean setting. The low impedance of the LM386 (only 50K) will cut the highs off when used without a buffer which is good for distortion but not for clean tones. Place the tone stack between the 386 and the TDA2003 and there you go!

Buffer -> LM386 -> tone stack -> TDA2003 -> speaker

Good luck! That should be a sweet little amp.

Take care,
-Peter

Hmmmm... I'm starting to get a clearer picture now .

More on buffer, doesn't the Ruby already have buffer in it? So it would be two buffers in the input :roll: ?

Can I just stick with the Little Gem? How better would it be if I bother to make a Ruby instead? What buffer would I make for the input, FET (I would still have one if I'll make a Ruby) or BJT?

Without buffer will be good for distortion, hmmm... Well, I'm just going to put a bypass swicth there. :wink:

How about the LM386N-1? I read somewhere about issues regarding voltage supply and the suffixes of the chip and from what I can remember, the N-1 is the least favored. The TDA2003 wants 12v and two power transformers inside isn't a good idea at ll . How can i drop the voltage to 9 for the buffer and LM386?

Thanks Peter :wink:

[Peter Snowberg]

You only need a single buffer on the input. I should have been more clear. The Ruby is like a Gem with a buffer added. Any kind of transistor will be fine for the buffer, but I feel JFETs will give the best sound.

The 386-1 will accept 12 volts for a supply, but don't go too far over that or it will fail. If you want to reduce the voltage, you can use either a linear regulator like the 7809, or you can place a small zener in line with the 386 supply. A 3.1 volt zener will chop 3.1 volts off of the supply. Just be sure to add a small capacitor to the power pin of the 386 which is after the diode.

Most 12 volt supplies will give more than 12 volts unless they are regulated. Always look with a meter before you connect any parts that might die with more than 12 volts.

If you are powering your amp with a transformer going into a bridge rectifier which then powers the TDA2003, you could also just add a 7812 to make a clean 12 volts for the buffer and 386 stage.

Take care,
-Peter

[zener]

Wow  Now I know what to do .

BTW, before I go to sleep (it's currently 3:30 morning :shock:  here in the Phil.), are there any sub for the 7812, just in case I can't find one.

A million thanks and goodbye Peter :wink:                    
(yawn)

[Peter Snowberg]

There are many regulator chips in the world. The 7812 is the most generic part number I know for such a chip. The next most generic number would be an LM317, but that is an adjustable regulator so you need to wire up two resistors along with it to make it put out 12 volts. One resistor will be 240 ohms and the other can be a 5K linear pot to adjust the output voltage. Look up the LM317 datasheet for more info there.

Take care,
-Peter

[zener]

Can I put a gain control somewhere in the power amp, so there would be a pre-gain and a post-gain controls?

Or am I just posting another stupid question and asking for "the thing that should not be" :roll: ?

[Peter Snowberg]

I think that is a very good idea.

I would add a master volume control using a pot after the tone stack and just before the TDA2003.

For a pre-gain, I would use a volume contfol between the buffer and the 386.

Finally, for an overdrive "channel", I would put the Ruby's "gain" control (pot from pin 1 to pin 8) on a switch. I would also place a 10uF or 22uF cap in series with pot and use a 5K pot instead of the 1K used by the 'Groove engineering department.

Here is the layout as I see it:

Code Select Expand

in -> buffer -> Preamp -> LM386 -> Tone -> Master -> TDA2003 -> Speaker
               Volume      |               Volume
                         Channel
                         Switch
                           |
                         Gain pot
It looks like a cool little amp to me.

Take care,
-Peter

[zener]

Whoooaa  My amp is starting to get better .

What could be the value for the pot between the buffer and the 386? 10k? How about the master volume? 1M?

The suggested 10-22uf cap can be either the pin1 or pin8, right?

Just got home with the wood chassis from the shop. Though it's just on a 1/4" board, it looks very sturdy. How I wish I have at least a webcam to post pics. I can just settle for the a film camera and have the photos scanned.


Thanks Pete :wink:

[Peter Snowberg]

Whoooaa  My amp is starting to get better .

What could be the value for the pot between the buffer and the 386? 10k? How about the master volume? 1M?

The suggested 10-22uf cap can be either the pin1 or pin8, right?

I think you're going to have a really cool little amp.

For the pre-volume, 10K is fine. You could probably use anything from 10K to 100K without noticing any change.

For the master volume, you might want something more like 100K. You have a mismatch here.... the Baxandall stack wants a high output impedance like 1M to work properly. The TDA2003 has about 70K input impedance. If you use a 1M pot, the volume control and the tone controls will interact quite a bit, making it hard to tube possibly. You will also get a midrange boost that may be hard to get rid of.  

A better way to do this stage would be to add another buffer after a 1M master volume pot, just in front of the TDA2003.

Think of it as two Rubys in series. The first one has the overdrive controls and feeds the tone stack. The second one makes the power for the speaker. The second buffer does not need to provide any gain. The TDA2003 will be very happy with 50mV so you could actually overdrive it with just your guitar and no buffer, the problem is matching the impedances.

Here is an updated block diagram:

Code Select Expand


in -> buffer -> Preamp -> LM386 -> Tone -> Buffer -> Master -> TDA2003 -> Speaker
               Volume      |                        Volume
                         Channel
                         Switch
                           |
                         Gain pot
The design is getting more complex, but the result is getting better too.

That little amp should be very versatile.

Oh, and on the 386 boost cap question…. Yes, the cap can be on either pin 1 or 8.

Code Select Expand


386 pin 1 -> 22uF cap -> gain pot (5K) -> channel switch -> 386 pin 8
I haven’t done much testing without the cap in the gain boost loop. I see many schematics where people leave it off, but the 386 datasheets all show it and I don’t want to upset the DC bias of the 386 input stage so I always include it. From what I remember of testing, I didn’t like the overdrive as much without the cap.

Take care,
-Peter

[zener]

I finally build only the power amp few hours ago. It failed . I built it having the negative 10uf towards the input. I plug the power and after five minutes of hooking up all the wires, I accidentally touched the back of my IC and it was so hot :x , it burned the tip of my index finger. It was like holding a soldering iron. I forgot to put the heatsink.

Nothing. I used the 8" 8ohm 18w speaker. I reversed the 10uf to have the positive towards the input and still nothing.

There are two LM383-based amps here http://www.generalguitargadgets.com/v2/richardo/practiceamp/lm383.gif  and  http://www.generalguitargadgets.com/v2/richardo/practiceamp/craig.gif which input caps are suspicious (or definitely wrong)

The layout I used. I made sure that there were no cold joints or shorts especially on pins 3, 4 and 5. I don't have 200nf caps, I just put two 100nf caps in parallel.

Is that IC a heat monster when you don't put a heatsink. I think it heats up as fast or faster than my soldering iron. Could I have fried it?

As soon as I've posted this, I'm already googling for other amps that I can sub for what Peter has come up. However, If it will be proven that my IC was fried, I'll just buy another and go on with it.

Thanks. :wink: