Amp guys (Aaron, Neil, Pete, etc) I need your help

[Paul Carmody]

It's time for me to have a summer project, aside from the steady backlog of loudspeaker designs I'm working on, and will be working on. This is something that I've wanted to understand for a dozen years--much like speaker design, which I'm actually starting to understand (so there is hope for me).

I want to understand how amplifier design works. Sure, there are amp kits out there--at least one of which I'm going to be building this summer. But really, building a kit is nothing new; I built amplifier and preamp kits in college. What I really want to know is what the parts DO in an amplifier.

I also would like to understand what the parts do in a power supply. Yes, I understand that a transformer steps down voltage, and what a bridge rectifier does, but I'm having a hard time turning that abstract knowledge into something useful. Much like, how 2 years ago I understood what a coil, a cap, and a resistor did in a speaker, but not how to actually use them to make exciting, tangible results.

So do you have any recommendations for good reading, or methods of tinkering-to-learn? The only electronics book I own is "Getting Started in Electronics" by Forrest Mims. I've read it cover to cover many times, but couldn't do a damn thing to prove that I understand it.

Any and all help is greatly appreciated! Oh, and by the way, I wasn't actually expecting to be able to be an amp expert in a matter of a few months. I expect that, like speaker design, this is something I will understand more and more over the course of years, through actual practice and implementation of ideas. I just wanted to start this summer;)

[Æ]

Re: Amp guys (Aaron, Neil, Pete, etc) I need your help

Every designer has his own, sometimes unique philosophy.
If you really want to know something about amplifier design, you should talk to someone like Nelson Pass.
I had the opportunity to work with Henry Wolcott, a world class amplifier engineer and holder of 14 patents. Real inventions, not bogus claims. He designed headphone amplifiers for the NASA Space Shuttle and a wheel on the track dynamo powered amplifier for the Indiana Jones ride at Disneyland. His audio amps were reviewed very favorably by big name audio critics.

http://www.enjoythemusic.com/magazine/equipment/0700/wolcottaudio.htm

He used a similar approach to that of Thomas Edison, 1 percent inspiration and 99 percent perspiration. Many, many hours of hard work.
If you want to make a good amp, just stick to a proven design. No need to try and reinvent the wheel.
You're asking a really open and vague question, you need to be more specific.

[Wolf]

"High Power Amplifier Design"....

I think it's what it's called off the top of my head, written by Randy Slone. PE used to carry it, but I'm uncertain if they still do.

Aaron recommended this book to me, and I still haven't finished it's infinite number of pages. Aaron said that the guy prefers Jfets, but that he really sees no reason to not use MOSFETs or other type transistors.

It requires some standard knowledge of PSU's, but everything else is supposedly there. He breaks it down into sections of the circuitry, and goes further. I think there are some you can build in there as well.

I would also read up at ESP, Rod Elliott's webspace. It has lots of info there that is useful.

http://sound.westhost.com/projects.htm

That's where I've come so far in my electronics endeavors...
I think I'll build proven designs, planned designs, or kits,
Wolf

[Paul Carmody]

Re: Amp guys (Aaron, Neil, Pete, etc) I need your help

If you want to make a good amp, just stick to a proven design. No need to try and reinvent the wheel.
You're asking a really open and vague question, you need to be more specific.

I'm not trying to re-invent anything. There are so many good amps out there, that come at a vast array of pricing. But how on earth do they work? That's my question.

To be more specific about what I want, I'll use an analogy:
Two years ago, when I looked at the schematic for a crossover network, it looked like a bunch of hieroglyphs. Nothing made any sense. Now, for the vast majority of networks out there, I can look at it and pretty much right away know what everything is for, and how it impacts the sound. I can even know how the sound will change if I change a component.
When I look at the schematic for an amplifier, it still looks like a bunch of hieroglyphs. I see a bunch of parts, and have no idea why they are placed where they are, and why they have the values they do. I want to be able to have a better understanding of "what's going on and why" with amplifiers. Not because I am looking to start my own company, or even outdo anyone's design. I have just always been fascinated by them, and want a much deeper understanding of them... so that, maybe someday I could even try making my own. Even if it sounds like crap, it would be very gratifying.

[Deward Hastings]

Re: Amp guys (Aaron, Neil, Pete, etc) I need your help

how on earth do they work? That's my question.

"Audio Power Amplifier Design Handbook", Douglas Self

http://www.amazon.com/Audio-Amplifier-Design-Handbook-Fourth/dp/0750680725/ref=pd_bbs_1?ie=UTF8&s=books&qid=1213343001&sr=8-1

This book is not about the nuts-and-bolts of *building* an amplifier (Sloan is better for that), it's all about *designing* an amplifier . . . what works, what doesn't work, and why. Especially "and why".

A summary from the author's web site is a "must read" all by itself:

http://www.dself.dsl.pipex.com/ampins/dipa/dipa.htm

[Deward Hastings]

Oops . . .

Make that "Slone" . . .

[philiparcario]

Re: Amp guys (Aaron, Neil, Pete, etc) I need your help

I'm not trying to re-invent anything. There are so many good amps out there, that come at a vast array of pricing. But how on earth do they work? That's my question.

To be more specific about what I want, I'll use an analogy:
Two years ago, when I looked at the schematic for a crossover network, it looked like a bunch of hieroglyphs. Nothing made any sense. Now, for the vast majority of networks out there, I can look at it and pretty much right away know what everything is for, and how it impacts the sound. I can even know how the sound will change if I change a component.
When I look at the schematic for an amplifier, it still looks like a bunch of hieroglyphs. I see a bunch of parts, and have no idea why they are placed where they are, and why they have the values they do. I want to be able to have a better understanding of "what's going on and why" with amplifiers. Not because I am looking to start my own company, or even outdo anyone's design. I have just always been fascinated by them, and want a much deeper understanding of them... so that, maybe someday I could even try making my own. Even if it sounds like crap, it would be very gratifying.

Look at a simple proven design Gainclones don't have a lot of parts play around with a few of them. There is a link for the 3886 chip and its aplications. You want to look at a part in an amp and say of I change that value this will happen.. I have a few gainclones also go to ska-audio.com the builder designer is greg ball nice guy an aussie he is based in austraila read the forums on his site. the ska stands for simple killer amp I own 2 stereo versions and a pair of monoblocks . I put the monoblocks together for around 700. other then power limits 45watts they are magic . I still don't fully understand but like you I am working at it.

[neildavis]

Re: Amp guys (Aaron, Neil, Pete, etc) I need your help

...What I really want to know is what the parts DO in an amplifier.
I also would like to understand what the parts do in a power supply. I just wanted to start this summer;)

Amps are built from many building blocks that are covered in a good circuits course. You should know how to recognize the basic gain stages, emitter follower stages, current mirrors, inverters, and you should know how to calculate the gain, and input and output impedance for each of these building blocks. Also, you should have a basic understanding of how a semiconductor works and why a diode behaves the way it does (the "diode equation"). So if you really want to know how an amp circuit works, you should take a circuits class at a local school--that will give you the foundation to understand all of the piece-parts that make up an amp. This course should be mostly concepts but enough algebra to de-mystify the gain and impedance calculations.

Once you understand the "static" stuff you can start on frequency response, phase and stability--this is where you start looking at real circuits instead of "ideal" amplifiers. But that course is usually taught using LaPlace transforms and unless you already have the math background it can be rough.

So unless you just want to treat the amp as a "black box" you should plan on taking the equivalent of two engineering courses. There are probably some good books that will walk you through this basic material, but taking a class forces you to do the homework to make sure you really, really understand it.

If you are content to deal with the amp as a "black box", just make a project or two with one of the "chip amps" or build from a circuit that someone has already designed--that's a lot easier. This would still be a good learning experience, because unless you are really lucky and get it right the first time, you will learn a lot about grounding and decoupling (circuit layout issues). But it's not going to de-mystify the circuitry, and it is a hobby-shop approach that won't yield the insight that I think you are looking for.

However, I would encourage you to redirect your interests to focus on Class D amps and get your feet wet with digital signal processing. The push to go "green" and the price of aluminum heatsinks and copper wire in transformers is making efficient amps more cost-effective. There are a lot of nice Class D amps out now, and there are some really nice DSP components that make crossovers and EQ quite simple. Even better, there are a lot of Class D amps with built-in DSP. I'll post some more on this topic later.

[Pete Schumacher]

I know how you feel

It's time for me to have a summer project, aside from the steady backlog of loudspeaker designs I'm working on, and will be working on. This is something that I've wanted to understand for a dozen years--much like speaker design, which I'm actually starting to understand (so there is hope for me).

I want to understand how amplifier design works. Sure, there are amp kits out there--at least one of which I'm going to be building this summer. But really, building a kit is nothing new; I built amplifier and preamp kits in college. What I really want to know is what the parts DO in an amplifier.

I also would like to understand what the parts do in a power supply. Yes, I understand that a transformer steps down voltage, and what a bridge rectifier does, but I'm having a hard time turning that abstract knowledge into something useful. Much like, how 2 years ago I understood what a coil, a cap, and a resistor did in a speaker, but not how to actually use them to make exciting, tangible results.

So do you have any recommendations for good reading, or methods of tinkering-to-learn? The only electronics book I own is "Getting Started in Electronics" by Forrest Mims. I've read it cover to cover many times, but couldn't do a damn thing to prove that I understand it.

Any and all help is greatly appreciated! Oh, and by the way, I wasn't actually expecting to be able to be an amp expert in a matter of a few months. I expect that, like speaker design, this is something I will understand more and more over the course of years, through actual practice and implementation of ideas. I just wanted to start this summer;)

I took 4 years of EE courses at UMR. After graduation, I still couldn't design an audio amp of anything more than a couple watts.

After 3 years in a lab building servo controllers, switching power supplies and other feedback stabilized circuits, I finally attempted my own design. I used an op-amp gain stage followed by a cascode voltage gain stage and source follower current buffer. The result was outstanding with all types of distortion coming in below the noise floor of the HP analyzer, more than 80dB down.

Schematics are an art in themselves. I usually found myself analyzing them for quite a while, only to re-draw them so that the individual sections were more easily identified, such as the current mirrors and output sections. Otherwise, the hieroglyph analogy held sway. Amp design is an order of magnitude or more beyond the design of passive XO networks from an electrical understanding standpoint. It's not something that most folks could pick up as quickly. Active circuitry is just much more involved. You may have just found yourself sitting down to a dinner of elephant. And if you're intent on eating the whole thing, remember, the best way to finish it is one bite at a time.

[Deward Hastings]

Re: I know how you feel

Schematics are an art in themselves. I usually found myself analyzing them for quite a while, only to re-draw them so that the individual sections were more easily identified, such as the current mirrors and output sections. Otherwise, the hieroglyph analogy held sway. Amp design is an order of magnitude or more beyond the design of passive XO networks from an electrical understanding standpoint. It's not something that most folks could pick up as quickly. Active circuitry is just much more involved.

What Pete said . . .

One of the strengths of Douglas Self’s book (and approach) is that he does just that, analyzing the distortion contributions of each stage along the way, and then how they interact in real circuits. It was, for a while anyway, accepted as gospel that complementary pairs was the “way to go”, especially for output stages. It turns out to be one of those “theory and practice” things (since truly complementary pairs turn out to be almost impossible to produce), and in the “real world” circuits like the output stage of the LM3886 turn out to be “good enough”. Better than “good enough” in fact, since up to the level where its protection circuits kick in the 3886 produces no audible distortion . . . most notably not suffering the “crossover distortion” that used to plague solid state amps. There are a number of circuit topologies now published for low to medium power amps which can be built for $20-30 in parts (not counting power supply, heat sinks and box) and which for practical purpose meet Self’s “blameless” criteria, the 3886 among them. It really does not matter which among them you chose to build.

http://www.national.com/ds/LM/LM3886.pdf

http://www.national.com/an/AN/AN-898.pdf

You might consider as a place to begin your study not power amplifiers but active (analog) crossovers instead. All the “semiconductor electronics” is hidden in the op amp “black box”, so you can focus on frequency response contouring and the effects of feedback, and there will be a clear bridge to concepts already mastered in passive crossover design. You get the added benefit of easily built devices that immediately improve the performance of your speaker designs (and a relaxed output power requirement for the amplifiers you will eventually build).

[envisionelec]

Re: "High Power Amplifier Design"....

Double Post.

[envisionelec]

Re: "High Power Amplifier Design"....

I Aaron said that the guy prefers Jfets, but that he really sees no reason to not use MOSFETs or other type transistors.

For the record, I didn't say that. I said that Slone has a preference for Lateral Fets.

My personal preference is Class D. I've given up with retro-tech analog amps.

[neildavis]

Analog stinks

...You might consider as a place to begin your study not power amplifiers but active (analog) crossovers instead.

Just kidding about the title, but take a look at the program in the link and you can understand why I'm much more excited by a DIY digital alternative to active crossovers. This program is a very early version of the Biquad Crossover Designer, which allows you to replace all that old analog stuff with low-cost digital biquads. It is similar to PCD, in that you can open up the same FRD response files and specify driver offsets. The program allows you to use the biquads to implement any standard crossover type and then tweak the response with shelf, peaking or notch filters--just by typing in the Fc, Q and Gain for each filter. You can play with this program to get a feel for active crossovers while using a modern digital approach.

The current version of this program works with the TAS3004 chip and I've got it partially working with a 6-channel STA308-based amp (one of those Spherex 5.1 Xbox amps). Eventually I hope to have this program work with a wide range of digital amplifiers such as the TI Purepath series, the ST DDX amps or similar products from Analog Devices and Cirrus. The biquads in these amps can replace as many as 100 reactive components (capacitor or inductor), yet the chips are in the $5 to $12 range. Cool stuff.

You may need to download Microsoft .NET for this program, and if you don't have a fast machine it may be slow (there is a LOT of number-crunching). I haven't tried the Mono program that converts the .NET Common Language Runtime (CLR) calls to different platforms, but if Mono works as advertised then this program should work on a Mac or Linux. The Mono download is at: http://www.go-mono.com/

An early version of BCD.exe that you can play with is at this link:
http://www.audiodevelopers.com/plateamps/BCD.zip

[Wolf]

Re: "High Power Amplifier Design"....

For the record, I didn't say that. I said that Slone has a preference for Lateral Fets.

My personal preference is Class D. I've given up with retro-tech analog amps.

L? J? Bunch of "they aren't the same?" to me- you know that. I thought I had it right, so I apologize and stand corrected.

Class D? I didn't know you went that direction.....;)
Later,
Wolf

[nick29498141]

Re: Amp guys (Aaron, Neil, Pete, etc) I need your help

Class D... ugh... :p

(remember, I'm the tubes and vinyl guy) :eek: