DIY amplifier - Page 2

for most diy stuff like this, i think a better forum would be diyaudio.com

diyaudio.com, a simple project that thousands of people have built would be a gain clone its a simple chip amp that does about 70 watts . u could also try a f5 which is class a 25 watt amp and then there are hundreds of ab amps one could build including krell clones and such. then there is class d i use these http://www.ebay.com/itm/120944373938?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649 they are cheap and easy to build used them on subwoofers i have given to friends without a problem i also use them for my mains

Agreed.

Hi Wag, Well, yes, if you wanted to keep that cap.

The author says it is to prevent "pops" when you turn the amp on, but that is not my experience. Typically, a cap placed at that location is there to protect the switch-contacts from the inductive arc that occurs when you turn the amp off (the transformer is an inductive load).

The issue I have with a cap across a switch with AC is that it will allow a small amount of current to flow when the amp is off. Not much current, but not zero. I would leave the cap off, and if you do indeed hear a "pop" when you power-up the amp, then you can add the cap later.

One book I found when I was thinking about building an audio amp was Designing Audio Power Amplifiers By Bob Cordell. It's a very thorough discussion of amplifier building blocks and amplifier design specifically for audio applications. It's a 600 page book too so there is plenty of detail.

vu meter

Hi Marty, Sorry, I can't recommend a good book, but I can give some insight at what is involved.

It takes an awful lot to design an amp like this one, but it is not too difficult to build, although it is time-consuming. You also may need special software and tools. I would not bread-board it, but would create a PCB, using mainly surface-mount components. I would then have the bare boards made by a low-cost prototyping house (like Sunstone) and have a paste-mask made at the same time. The paste-mask allows me to "screen" the solder-paste onto the PCB in one swell-foop, and I then place the parts into the paste. Five minutes in a toaster over to reflow the solder-paste and you're done.

If I'm doing more than one or two, I have a prototyping shop do the assembly.

With surface-mount, it's difficult to swap components to experiment. That's never been a problem for me, as I do all of my experimenting on the computer with a Spice simulator. Swapping components in circuit won't tell you the important things, like whether a value-change is causing stress on other components, or whether a transistor is getting too hot internally. Spice also helps you to determine the best tolerances for your components, which is close to impossible with experimenting.

If you really want to experiment with amplifiers, I would suggest starting with a free version of Spice, start drawing schematics and simulating them on the computer. It may not be as rewarding as actually hearing the results, but it's the best place to start, and I suspect your ears would be pleased when you're done.

Hi Wag, There are two schools of thought on this. One is to measure the incoming signal, through a buffer amp, and display that on the VU meter. Because of the possible deviations in amplifier gain, I would consider this just an estimation of output power. But I suspect that it effects the output the least.

More common in high-end amps, like that McIntosh, is to place the meter directly on the output. My Phase Linear 400 puts the meter, a diode and a 12k resistor (all in series) across the speaker output. My Sansui BA-2000s does the same, but adds a buffer-amplifier between the speaker output and the VU meter. I suspect they both have minimal effect on the output, but the Sansui should have almost no effect with the buffer-amp. Both amps are vintage, late 70s designs.

Notice that in all designs, the meter is actually reading voltage, not power. So the power-markings on the meters are non-linear, and assume an 8 ohm load.

Hmmm . . . I'm puzzled by the numbers on both of those.

The first one is decibels only, so an appropriate sized series resistor, as well as a diode, could give you a decibel reading with 0db set to your full-power output (or anything else you choose). But determining that resistor may be a trial-and-error exercise, as I'm not sure their numbers mean anything. The one number you might try is the 500 microamps for full-scale.

The second meter is designed for 0db to be at 700 watts. That would be great if you were building a 700 watt amplifier. Since 700 watts is about 75 volts into 8 ohms, then you would want 75 volts to produce 500 uAmps through the meter. That would need a total resistance of 150 kohm, so a series resistor would indeed be called for.

One other component you need is a diode. Since the output swings in both positive and negative directions, and the meter only swings in one direction, you need a diode is series with the meter and resistor to rectify the output as well. It is the dis-symmetry that the diode adds that causes some to say that the meter "distorts" the output. But if you compare 8 ohms with 150 kohms, you can predict that any distortion would be rather small. However, you can satisfy those that won't do the math by using a bridge instead of a single diode, and thereby driving the meter with both the positive and negative halfs of the output.

A simple 1N4002 would do. Almost any diode that can handle 60 volts would do, but the lower the forward voltage-drop, the more accurate the meter would be. Myself, I wouldn't use a bridge, as you would be adding two diode-drops. Schottky diodes typically have lower voltage-drops. Any diode would be able to handle the 500 microamps of current that the meter might draw.

Random order.

It won't matter what order they are in, but the polarity of the diode with respect to the meter needs to be correct. You can tell if its incorrect, as the needle will move in the wrong direction. If that happens, simply swap the wires on either the diode or the meter.

people in diyaudio sure ain't as friendly as in here.
in replying i mean. especially to beginners question like these.

post there and bet you'll get 'go google' hehe

An excellent book, which I would also recommend (I own it).
Another worth reading simply for the excellent basic knowledge on circuit design is Morgan Jones' Valve Amplifiers. If you're not interested in tubes, might not be worth buying it, but definitely read it if you can borrow it.

lol smd uhh u could attach little wires to each side next time make sure u are buying radial or axial caps and resistors

SMD isn't too bad to do DIY as long as there are only 3 of them. Assuming you're not going to convert a toaster oven to bake 'em, all you need is a candle/coffee warmer and a heat embossing gun.

Get some solder paste from zeph.com.

Armed with those keywords you should be able to google up some good howto's

Tim

Paste is for SMD. Use the standard wire solder for your other components. What iron do you have? I you can get a really fine tip for it, you may be able to use it on the SMD, but more than likely something will get screwed up, especially if you're new to soldering and don't have an adjustable iron.

Tim

I think solder paste is a waste unless you have alot of components and a reflow method. If you are just soldering a couple SMD caps you can just use a regular soldering iron. Assuming you are having a PCB made just put the cap down on the pads and hold it there with tweezers while your bring your soldering iron (with a little solder on the tip already) over to one end of the cap and let the solder flow to the board and component. Sometimes it doesn't come out great but you just need to tac one side down so you can solder the other like normal. After both sides are soldered you can go back and touch either side up one at a time.

There are a bunch of soldering videos out there to look at if you haven't a clue. Plus most parts are cheap as dirt so even if you totally ruin them, you can always get another.