High Power Amps - guitar, car, PA, etc.

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
Post Reply
gjohnson
Posts: 7
Joined: Wed Dec 17, 2003 1:01 am
Location: Chico, CA
Contact:

High Power Amps - guitar, car, PA, etc.

Post by gjohnson » Wed Dec 17, 2003 12:49 pm

I'm an EE student and I'm currently taking a course that covers power amps. I live with two guitar players who would like to have me build them a good amplifier. I'm shooting for something in the 200-400W range for now, with the overall goal being to move into the 1000W neighboorhood. Think 'Back to the Future'. :D <p>Ideally this would be a Class A amp, but A/B would suffice. I'll play around with a Class C, but I haven't had the DSP coursework yet, so that's an on-the-horizon thing for now. <p>My question to you guys is two-fold. First, where do I go to find components that can dissipate this kind of heat? I looked at all the big names, Jameco, Digi-Key, etc, but I can't find a transistor that handles more than just a few watts, maybe 100 or so. Tubes would work, but I can't find any good documentation on how to design with them. Are there high-power bipolar dealers out there somewhere or do I have to use FETs? Also, where do I look for the high-performance transformers my instructors tell me I'll need? Will any over-the-counter unit work, or do I need a prescription?<p>The second part is an open-ended request for any tips, basic schematics, or anything else you guys might think would be useful to a fledgling amp designer. I know the basics of A, A/B and C class amps and some of the calculations involved therin, but I have yet to actually design one from scratch. Any tips?<p>Thanks a million!
Glenn Johnson
Electrical Engineer in Training
CSU, Chico<p>-- I am Kaiser Soze.

dyarker
Posts: 1753
Joined: Fri Aug 22, 2003 1:01 am
Location: Izmir, Turkiye; from Rochester, NY
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by dyarker » Wed Dec 17, 2003 7:18 pm

Ganged A/B amplifier sections, with outputs isolated/balanced by low Ohm, high power, resistors. (The output of each A/B amp is connected to a resistor, the other ends of the resistors are connected in parallel to feed the speaker(s). Transistors have variations in gain, the resistors help prevent a strong section from driving power into a weaker section.) <p>I haven't done the math (you said you could), but something like 0.5 Ohm, 15 Watt, aluminum case, power resistors comes to mind.<p>For safety, the heatsinks should be grounded (green wire). That means mounting the power transistors with thermal conductive/electric insulating washers/gaskets, non-conductive shoulder washers on screws, etc.<p>tip- thermally connect the bias diodes to their power transistors.(So as the transistors warm-up and gain and forward base emitter voltage change, the diodes warm-up too.)<p>Are the transformers mentioned for the power supply, or speaker output?
Dale Y

rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by rshayes » Wed Dec 17, 2003 8:10 pm

The parts sold by Digikey and other distributors are representative of those in high volume production and available at reasonable cost. Higher power devices are available, but at substantially higher cost. In the semiconductor business, 100 watts dissipation is not a "few" watts. Past this point, removing heat from a device becomes much more complicated, such as forced air cooling, and ultimately resulting in things like water cooled heat sinks. The usual solution is to use devices in parallel, which places the thermal paths in parallel, and allows higher overall dissipation.<p>Class A is 50% efficient at best. The power drawn from the power supply is constant, even with no signal out. The result is that the amplifying devices must be capable of dissipating twice the maximum power output on a continuous basis. This would be 200 watts for a 100 watt amplifier.<p>Class B is about 67% efficient at full power (with sine waves). Current drain and dissipation is less with lower power. The dissipation is lower if the signals are not continuously at full power, and some audio amplifiers take advantage of this fact. Audio signals tend to have high peak power with lower average power. There are differences of opinion as to how to rate audio amplifiers, ie. peak power, music power, or continuous power.<p>Class C is usually used for RF amplifiers. The output level depends mainly on the power supply voltage rather than the input signal. Efficiency is quite high, up to 90%.<p>Switching amplifiers have high efficiency, but are difficult to make work as audio amplifiers. The switching frequency components have to be removed from the output, which may be difficult since they may be as large as the maximum signal. Switching times have to be short, which requires high frequency switching devices. These amplifiers can be built, but I wouldn't recommend them as an initial project.<p>The impedance levels of loudspeakers are more compatible with transistors (either bipolar or FET) than with vacuum tubes. Normally, an output transformer is used with vacuum tubes to match the impedance levels. A high quality output transformer is a design project in itself. One of the classic designs was the Mackintosh output transformer. These are no longer made, and used Mackintosh amplifiers (30 or 40 years old) can cost several hundred dollars. This is for a 50 to 75 watt amplifier. Power supply voltages are high with vacuum tubes, a 2 or 3 hundred watt amplifier might use power supplies in the 1 to 2 kilovolt range. Power dissipation is fairly easy to obtain with vacuum tubes, transmitting tubes with dissipation ratings of several hundred watts are not uncommon (for example, the 4CX150A and similar tubes).<p>The conventional approach is to use paralleled semiconductior devices in a class B output stage. As you can see, this is basically the lesser of several evils. Amplifiers have been built with the other approaches, but they tend to be very expensive and/or difficult to get parts for.

rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by rshayes » Wed Dec 17, 2003 8:24 pm

A somwhat unconventional approach, after reading Dale's posting, would be to build modules of about 100 to 200 watt capacity and combine their outputs with audio versions of the hybrid power combiners used for RF amplifiers. The transformers needed for this would be low impedance designs with relatively few turns required. These could be wound by hand if necessary. Since voltage levels would be low, bifilar windings could be used to reduce leakage inductances and maintain high frequency response.

User avatar
Edd
Posts: 885
Joined: Sun Jan 27, 2002 1:01 am
Location: Dallas Tx
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by Edd » Thu Dec 18, 2003 12:39 am

Glen:
<<The second part is an open-ended request for any tips, basic schematics, or anything else you guys might think would be useful to a fledgling amp designer. I know the basics of A, A/B and C class amps and some of the calculations involved therin, but I have yet to actually design one from scratch. Any tips?>>
Sounds like you potentially might have a hand full of outputs to crunch in milliseconds ;)

gjohnson
Posts: 7
Joined: Wed Dec 17, 2003 1:01 am
Location: Chico, CA
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by gjohnson » Thu Dec 18, 2003 4:08 pm

Thanks so much for all the input - I'll have to sit down and digest it all over a Guinness or something, but one thing that caught my eye was the mention of parallel output stages.<p>I was never told of this idea in my classes, probably because of the foundational nature of the intro material. It would appear from a cursory glance at the diagrams posted that I could design smaller A or A/B stages and then just put them in parallel, adding the output power from each...? Is this correct? A phenomenal idea but you'll have to forgive me, it almost sounds too easy...<p>As for the transformer, I was thinking of one for the output stage (to boost the efficiency), although i will probably need one for input voltage rectification as well. <p>Thanks again, all your ideas are greatly appreciated!
Glenn Johnson
Electrical Engineer in Training
CSU, Chico<p>-- I am Kaiser Soze.

rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by rshayes » Fri Dec 19, 2003 5:34 am

Directly paralleling low impedance outputs is too easy. If the output voltages of the amplifiers are slightly different, then the amplifier with the highest output will supply additional current to increase the output voltage. The lower output amplifiers will sink current in an attempt to reduce the output voltage.<p>High power RF amplifiers are impractical at high frequencies. The transistors have to be physically small, which limits their power output. Power combiners are used to parallel the outputs of several low power amplifiers to get higher power. Motorola used to have ap notes showing this type of design. I don't know if the RF devices are still made by Motorola or by On Semiconductor. Try using a search engine to locate their web sites.

rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by rshayes » Fri Dec 19, 2003 5:35 am

Directly paralleling low impedance outputs is too easy. If the output voltages of the amplifiers are slightly different, then the amplifier with the highest output will supply additional current to increase the output voltage. The lower output amplifiers will sink current in an attempt to reduce the output voltage.<p>High power RF amplifiers are impractical at high frequencies. The transistors have to be physically small, which limits their power output. Power combiners are used to parallel the outputs of several low power amplifiers to get higher power. Motorola used to have ap notes showing this type of design. I don't know if the RF devices are still made by Motorola or by On Semiconductor. Try using a search engine to locate their web sites.

dyarker
Posts: 1753
Joined: Fri Aug 22, 2003 1:01 am
Location: Izmir, Turkiye; from Rochester, NY
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by dyarker » Fri Dec 19, 2003 8:31 pm

Don't directly parallel the outputs. The "weaker" transistor will not handle any current, the "stronger" transistor will pass all the current till it blows. A resistive combiner lets each section handle it's share of the load.<p>For example, an 8 Ohm speaker and the strongest amp puts out 10% more voltage than the weakest, tested individually under load. Each output has a 0.5 Ohm resistor, so the path between amps is 1 Ohm. 1 Ohm is 12.5% of 8 Ohms, so it should work reliably. The better the match of amp sections, the lowwer the value of the resistors can be, less power wasted heating the room. From the speaker's "point of view" the resistors are parallel. So for 4 sections, 0.5 Ohm resistors would add only 0.125 Ohms to the speaker circuit!<p>Use fast blow fuses in the output to protect the amps against shorts.<p>C U L -
Dale Y

gjohnson
Posts: 7
Joined: Wed Dec 17, 2003 1:01 am
Location: Chico, CA
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by gjohnson » Wed Dec 24, 2003 7:18 pm

Thanks again to all who replied. I'm looking at the schematics that were posted and I think I've almost got all the information I need. Dale, thanks for the elaboration on the output resistance coupling between amps and speaker. I noticed something like that in the aforementioned schematics, but didn't really know why it was there. <p>Off I go to blackout a city or two in testing! :D
Glenn Johnson
Electrical Engineer in Training
CSU, Chico<p>-- I am Kaiser Soze.

Bernius1
Posts: 349
Joined: Sat Sep 21, 2002 1:01 am
Location: NY
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by Bernius1 » Fri Dec 26, 2003 5:33 am

$.02 ; Crown brand PA amps have 4x(6) TO-3 output transistors, driven by 2x(2) TO-220A drivers per bank. Looks to me like a current-mirror arrangement, which would explain why the only load resistance on the TO-3's is the speaker. But the current hogging makes problematic sense.
Someone (OK to take bow if you come forward) on an earlier power-supply post made a current limiter by paralleling a resistor and the E-B junction of a transistor, so that as Vdrop over the R passed .07V, drive current would start shunting to ground.
Granted, with an AC input you don't want to clip SIGNAL, but maybe a differential comparator sensing Vce on each of the TO-3's. Lowering current at the driver's base would limit hi-side current. But then each would need its own driver. Mo' spensive.
I've also seen something where an AGC circuit uses a 'static' charge across a capacitor as a voltage level for gain control. As voltage on the cap changes, so does the operating point of the amp. The advantage to this is that one comparator can sample each TO-3's Vce ( or DC Vbe ?), report it to a PIC, which can output a PWM'd pulse to the cap (+ or -) to vary the gain on the TO-3 with the greatest Vdrop. 'Smart' AGC. And at 1Mhz, variations between transistors would be minimal even with signal applied.
Can't we end all posts with a comical quip?

chessman
Posts: 292
Joined: Tue Jan 14, 2003 1:01 am
Location: Issaquah, WA
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by chessman » Sat Jan 03, 2004 10:17 am

Just thought I'd add that a true Class A amplifier will realisticly be much less efficient than 50%.<p>I know people that have built some class A amps that are lucky to get 10% efficieny.

gjohnson
Posts: 7
Joined: Wed Dec 17, 2003 1:01 am
Location: Chico, CA
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by gjohnson » Sun Jan 04, 2004 5:39 pm

True enough - 50% is only a theoretical maximum, and even then only if an output transformer is used. The schematics posted above look to be A/B amps though, and if biased correctly I know these can nearly eliminate the dreaded crossover distortion that seems to be the only attractive thing with a pure Class A. <p>Once again I sincerely appreciate all your responses in this thread. I now know virtually all I need to know now to start putting one together. Special thanks to those of you who posted the schematics, as this will significantly aid in my learning in this subject. I wish everyone here a very happy new year and send out my thanks once again!
Glenn Johnson
Electrical Engineer in Training
CSU, Chico<p>-- I am Kaiser Soze.

chessman
Posts: 292
Joined: Tue Jan 14, 2003 1:01 am
Location: Issaquah, WA
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by chessman » Mon Jan 05, 2004 12:34 pm

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by Glenn Johnson:
the dreaded crossover distortion that seems to be the only attractive thing with a pure Class A. <p><hr></blockquote><p>Also remember that class A amps are mainly used by audiophilis when everything has to be perfect. Another advantage of class A is there is only one "side" of the amp, if you will. It isn't two identical sections - one for negative and one for positive. With todays component tolerances, the two sides of the amp could be off in frequency response, distortion, gain, impedence, and phase response. That can lead to a sound that isn't as clear as class A.

Koldby
Posts: 1
Joined: Mon Jan 19, 2004 1:01 am
Location: Denmark
Contact:

Re: High Power Amps - guitar, car, PA, etc.

Post by Koldby » Mon Jan 19, 2004 11:25 pm

Chessman<p>Not quite true:<p>When you are talking abount a "one sided amp." you are reffering to singel ended designs - by nessesity a class A design - but the lot of class A designs are indeed complementary push-pull designs biased into class A operation.

Post Reply

Who is online

Users browsing this forum: No registered users and 34 guests