How to build a simple, low power guitar amp

[cbdb]

Never heard of an electric guitar amp without distortion. If all you do is play it clean get an acoustic. Just kidding, sort of. I guess you can add what you want with external,pedals.

 

[filthyone]

... I guess you can add what you want with external,pedals.

Exactly! What i actually want is a "buffer" with some gain (aka an amplifier) between a pedalboard and a speaker. Low power that is...

 

[Gnobuddy]

What i actually want is a "buffer" with some gain (aka an amplifier) between a pedalboard and a speaker.

Ah. That is certainly do-able. You don't have to deal with the high input impedance needed by an electric guitar pickup, so BJTs will do the job.

By the way, my guess is your bad FET experiences have been with MOSFETs. Have you ever tried using JFETs? They are not static sensitive, and don't die as easily; they are at least as robust as a typical small-signal BJT.

(JFETs are endangered species nowadays, though, and seem to be getting rarer by the month.)

If you're open to any discussion on the subject, may I ask what you believe to be the benefit of pure class A all the way to the speaker? As you probably know, there are drastic downsides to class A power amplifier designs. So what's the upside, in your opinion?

-Gnobuddy

 

[filthyone]

Thanks for your time Gnobuddy.
You are right. My bad experiences have been almost exclusively with mosfets, either prematurely dying of static charges or exploding unpredictably, with as much as a mosquito flying over the gate!

As for your class A questions, which i find totally reasonable, i don't think i have convincing answers. But let me try anyway...

Short answer: psychological reasons!

Long answer:
I was thinking that, since the majority of audiophiles think that class A is relatively better sounding than class AB, and since my power demands are really low, then why not follow the class A path?
Since efficiency is not a big issue with such a tiny amp, and i happen to have lots of big aluminum heatsinks collecting dust, why not go the extra mile and get rid of the crossover distortion that usually attacks the high-order harmonics?
I know that there seems to be some controversy over this issue, and to be honnest, i'm sure i would repeatedly fail an a/b blindfolded test since i don't possess a golden ear. But since i get in the trouble of building something, why not build it to the highest reasonable standards, and then be proud(er) of my creation afterwards (even though the benefits might not be there for real, and/or i'm not able to appreciate them).

I WANT TO BELIEVE!

 

[Gnobuddy]

Well, there's no known cure for wanting to believe.

Why not do class A? Good question, and I have no answer that applies in your situation! Personally, I tend to think like an engineer (though I'm not actually one), so for me there is a tendency to optimise a design, which also includes sanity-checks. If there isn't an objective, repeatable, indubitable benefit to a design feature, I won't do it, particularly if it comes with negatives instead of positives - adds complexity, heat, weight, reduced component lifetime, etc.

But a more artistic way to look at a design is "Why the heck not?", as you are doing. And since it is your personal project, for your personal enjoyment, you are absolutely right!

I do have a personal reason for disliking class A, and it does have to do with heat dissipation. It takes a ridiculously large heatsink and/or forced air cooling to keep a semiconductor reasonably cool when it's dissipating lots of power. I have bad memories of trying to build linear voltage regulators when I was a teenager - I could never manage to get enough heatsinking onto the blasted things!

Getting 5 W RMS of clean audio out of a class A output stage will probably require getting rid of 7 - 10 watts of heat, worst case, i.e., when the amp is idle.

This is actually quite a lot of heat to deal with. I have used soldering irons with less wattage than that, and they got hot enough to melt metal!

If you already have power supplies and heatsinks lying around, it may be worth mounting a couple of power resistors to the heatsink, running 5 to 10 watts of (DC) power into them for several minutes, and finding out just how much heatsinking it takes to keep temperatures below, say, 80 degrees Celsius, maximum. (I usually targeted 60 - 70 deg C case temperature, I don't trust hot semiconductors.)

You can, of course, calculate the size of the heatsink you need, if you have all the right data. I always found the quick experiment with the power resistors more trustworthy, though, because of the uncertainties in the calculations (such as available airflow).

On the subject of audiophiles: Here Be Dragons is written on that part of the map, and I fear going there. However, for the safety of your sanity, intellect, wallet (and the preservation of your personal relationships!), I recommend a strong protective suit of armour when going anywhere near contemporary audiophiles; said armour consisting of a very large dose of skepticism, backed up by as much hard-core, thoroughly tested, objective engineering knowledge as you can get - stuff done using actual research methodology by qualified people, for example, the many decades of work done by Bell Labs researchers.

If you haven't seen the following video already, this may be a good start to growing that suit of protective armour. You will never trust your ears again once you've watched it, and that is exactly what you should do: NOT trust your ears, because they are as easily fooled as our eyes are (just Google "optical illusions" for plenty of proof!)

Here's the video: https://www.youtube.com/watch?v=BYTlN6wjcvQ

-Gnobuddy

 

[JMFahey]

In general avoid Red Circuits.

Experienced hobbyists or Pros avoid them like the plague they are, but magnetically attract beginners, they * look* simple for what they promise , problem is they don't deliver or plain don't work.

Yes, they are that bad.

 

[Mooly]

If you want Class A and a simple design then why not look at the original JLH69. Its pretty much as good as it gets, a four transistor bjt design that set (and still does in many ways) the audiophile standard for decades. There is a mass of info about it if you search.

 

[Printer2]

In general avoid Red Circuits.

Experienced hobbyists or Pros avoid them like the plague they are, but magnetically attract beginners, they * look* simple for what they promise , problem is they don't deliver or plain don't work.

Yes, they are that bad.

In that case, have built and abused this one. Needs a low output impedance to feed it though. Also not Class A but you can adjust the bias to run hotter if you heat sink it well enough.

An externally hosted image should be here but it was not working when we last tested it.

http://www.swtpc.com/mholley/PopularElectronics/Dec1967/PE_Dec_1967_pg30.jpg

As far as the JLH69, I think I have all the parts for one, just other projects that have greater priority.

 

[filthyone]

@JMFahey
Thanks for the tip man!

@Mooly
I have already tried to build this amplifier, actually four of them, two of the '69 original and two of the '96 improved versions with symmetrical sypply, but not any of them worked. All i could hear was a loud buzzing at the output (with regulated supplies) and the trimpots did not make much of a difference... I struggled for some days but didn't get anywhere, so i was discouraged and abandoned the project. Most probably, my transistor substitutions were not optimal (as i recall, i used bc560 inputs, bd139 splitter/driver and TIP3055 outputs). Had i succeded with that build, i would probably build another one for the present project. It would probably work fine for that purpose.


@Gnobuddy
I'm 100% with you! I'm usually really sceptical and don't accept "facts" without HARD evidence. But, as you say, this project was more a desire of my artistic side, where the scientific facts come second. Or in other words, just a scientific, unscientific experiment!
I watched the video, which i admit was totally eye-opening for me! I mean, i knew that there's too much ******** floating around the audio technology/industry, but i was missing a lot of the specific aspects of it. Thank you for that!!

After all of this, i think i will just build a simple chipamp with my spare tda2030s, and save all lot of time and trouble... In any case, i'm sure i won't be able to tell the difference!

Thank you all guys!

 

[Mooly]

The JLH is a pretty tolerant design and can even be built with point to point wiring. Don't know what to say over your experiences with it and those transistor types should be just fine.

Why not have another go, initially with a straightforward PSU. There's loads of help on here if you ran into trouble.

 

[filthyone]

About the Jlh, i think it's time i revisit the project. I remember thoroughly checking everything, including voltage/current and even pcb checks for topology mistakes, but everything seemed ok, except for the no audio at the output part. But i think it's high time i give the original 69 version another go, maybe start from scratch with a breadboard build.
I know that people have build it and everyone seems happy with the results (except for the PSRR part, where these kinds of circuits leave a lot to be desired). But that's where the carefull psu design techniques come at play...

 

[Gnobuddy]

I remember thoroughly checking everything, including voltage/current and even pcb checks for topology mistakes, but everything seemed ok, except for the no audio at the output part.

Solid-state audio amps have, for many decades now, consisted of several stages, all DC-coupled to each other. This is wonderful for performance, but there is a downside: if you make a mistake in construction, the whole circuit often goes out of kilter. Wrong voltages everywhere, wrong currents everywhere, and, quite possibly, overheated and destroyed output devices to add to the general state of annoyance.

I think it's really helpful to develop one's construction and fault-finding on simpler circuits first, things that don't involve multiple active devices all directly coupled to each other.

In your case, if all the voltages in the JLH amps were correct, I wonder why it was humming so much? A lot of hum, in this case, could be a sign of too much current being drawn from a power supply, causing ripple voltage to shoot through the roof. I've seen that happen before.

A signal-generator (even a crude one, like your phone playing a suitable MP3 file) and an oscilloscope are wonderfully useful tools to chase down mysterious problems like this. Not everyone has a 'scope, though, and if you don't have one, it is really helpful to have a signal tracer: a low-powered, sensitive audio amp with it's own speaker and power supply.

Wire a test probe to the input of the the low-powered signal-tracer amp. Also clip-lead the ground of the signal tester to the ground line of the circuit under test.

To use the signal tracer, you inject your audio test signal into the input of your misbehaving circuit, and probe with the tracer at various points in your circuit, starting at the input, and progressing towards the output. If you have a signal present at the point you're probing, you will hear it through the signal tracer's speaker.

So, if the tracer picks up your test signal at point A, but not at point B, you now know the signal died somewhere between those two points. Now you have localized the problem, and that will help you pin-point it using your other test equipment (voltage checks, etc).

Once again, entirely direct-coupled circuits can be hard to test, even with a signal tracer; if all the DC operating points are out of whack, the signal may just die right at the input, even though the actual problem is much further downstream.

Building your own signal tracer could be a useful and fun little project, actually. Something like a little LM386 chip, a small 2" or so speaker, a few AA cells or a 9V flat battery, a couple of resistors and caps, and you have a useful piece of old-school test equipment.

Please don't give up, and keep working on those trouble-shooting skills. It's no fun throwing out an entire completed project, and starting over from scratch just because there's one tiny mistake somewhere. It's much more satisfying to find that one tiny mistake, fix it, and watch your creation burst into life!

-Gnobuddy

 

[Mooly]

About the Jlh, i think it's time i revisit the project. .

Yes, I'd say it was worth having another go. You can keep R1 and R2 high (initially) to keep the current low. It will still all work correctly like that except of course the maximum output is reduced. Even values such as 470 ohm and 2k2 would be fine for testing, and these would see the current down in the one to two hundred milliamp region.

 

[Gratewhitehuntr]

Have you ever tried using JFETs? They are not static sensitive, and don't die as easily; they are at least as robust as a typical small-signal BJT.

(JFETs are endangered species nowadays, though, and seem to be getting rarer by the month.)

-Gnobuddy

Why are they becoming endangered? The demise of analog? (assuming that what I just read about analog switching duty is accurate or universal)

Where would I look to salvage something desirable?

EDIT I see the complaints of thru hole giving way to SMT, LOL I've been soldering tiny bits on surf boards for 10 years now and was floored by the space in most point to point devices.
Also, I would like to thank Mooly for his sig line. I'm reading it right nnnnnnnnow......

 

[Gnobuddy]

Why are they becoming endangered?

I don't know for sure, but I think MOSFETs matched or exceeded JFETs in just about every way, except for low noise, where JFETs are better. MOSFETs are cheaper to make as well (fewer steps in the manufacturing process).

The JFETs that I know about, typically have transconductance and voltage gain in the same general range as ye olde triode valves, so BJTs beat the pants off them in these areas.

I have the impression JFETs were never more than minor players when it came to commercial audio products.

Last time I checked, Mouser still shows some through-hole JFETs available, but all of them were listed as "end of life".

I've been soldering tiny bits on surf boards for 10 years now and was floored by the space in most point to point devices.

When I was about 10-12 years old I happily soldered devices with 0.1" (2.54mm) lead pitch into incredibly cramped circuits of my own devising.

Now, with middle-aged eyes, working with even those old through-hole parts would be a challenge - never mind SMD bits and pieces!

I got back into music a few years ago, and it led me, eventually, back to electronics. Valves and 1/2 watt resistors, it turns out, not only are big enough to see clearly, they also can sound great with electric guitars.

I draw the line at octal valves though - those things are just too enormous for my sense of propriety!

Also, I would like to thank Mooly for his sig line. I'm reading it right nnnnnnnnow......

I'll add my thanks to yours. I learned a lot about using LTSpice from Mooly's tutorials.

-Gnobuddy

 

[Gratewhitehuntr]

...and now I might have jinxed myself and there is a distinct possibility that the TAS5630 in my favorite instrument amp is kaput .... like soldering a millipede... what a pit .....

 

[Gnobuddy]

400 watts in a little surface-mount chip? Wow. It's amazing what these class-D chip amps can do now!

-Gnobuddy

 

[filthyone]

Yes, I'd say it was worth having another go. You can keep R1 and R2 high (initially) to keep the current low. It will still all work correctly like that except of course the maximum output is reduced. Even values such as 470 ohm and 2k2 would be fine for testing, and these would see the current down in the one to two hundred milliamp region.

Thanks for the tips Mooly,
i finally found my old prototype boards of shame (hidden in a box, far, far away). I will study the original circuit again, but i think i will begin from scratch, with fresh transistors on a breadboard. I cannot disassemble the existing transistors because i bent the leads under, parallel to the board and used like half a kilogram of solder (i was young, and foolish, probably still am...)

@Gnobuddy
The audio signal tracer is a fantastic idea, and a really useful tool, that i will definately make, since i still don't have an oscilloscope. Thanks for the idea!!!

 

[Mooly]

^ OK that sounds like a plan.

And thanks for the comments on LTspice guys.

 

[Gnobuddy]

I cannot disassemble the existing transistors because i bent the leads under, parallel to the board and used like half a kilogram of solder (i was young, and foolish, probably still am...)

Mistakes are part of learning - if you're not making mistakes, you already know what you're doing, so you're not learning anything new!

When this business of "shame" about mistakes comes up, I try to remind myself about babies learning to stand up and walk for the first time. They fall, over and over and over again. They look silly, they get angry, they get sad, they get probably get sore or hurt a bit. But they keep trying, and so eventually they succeed.

If they felt ashamed, and gave up, as we silly adults do, they would stop trying. If that happened, all of us adults would still be crawling around on all fours, or scooting around on our bottoms!

Now, do you have a de-soldering tool / solder sucker? The small ones are very inexpensive ( http://www.amazon.com/Elenco-060820-Solder-Sucker-desoldering/dp/B0002KRAAG ). You heat up your big ugly blob of solder till it melts, and use the solder sucker to remove as much solder as you can.

If you used far too much solder initially, this job actually becomes easier. A big blob of solder wants to fall off when heated, so even a little suction will probably get most of it off.

If there are still traces of solder which the sucker cannot remove, you can use copper desoldering braid: Desoldering Braid - Power Soldering Accessories - Amazon.com

You heat up the solder with the tip of the copper braid pressed against it, and it will wick up the solder, removing it from the circuit board.

Once the solder is gone, you can pry the bent transistor legs straight (if necessary, do this while the iron keeps things hot and the solder liquid).

I think it's worth making the attempt, because it's really such a valuable skill to to learn how to recover from mistakes, whether in electronics, or in life itself. Mistakes are part of learning, they're inevitable. But to proceed, we need to recover from them, and move on. (On to the next new mistake, of course! )

What's the worst that could happen? The transistors get damaged by the heat, or the PCB itself does. Okay, that makes them useless, but they are already useless just sitting there. So you're no worse off than before, and in the meantime, you developed some new desoldering and component removal skills which will help you with your next project!

-Gnobuddy