Looking to build a cheep 100 watt or so power amp

[deadspeaker]

Hey, What do you guys think of the Jon Tirone's amp? http://pluto.spaceports.com/~fishbake/amp/ca100.htm

Do you think I could use this transformer?  http://www.mouser.com/search/productdetail.aspx?R=166K35virtualkey54600000virtualkey546-166K35

If you guys know of another 100watt or so power amp that I can make for cheap that would be cool too. I plan to use a Sans Amp with it.

thanks for the help

[teemuk]

Mildly put: that amp is horrible.

If you ever get that design to run at 100 watts, let alone 150, it is at its dying breath. The output devices have a maximum rating of 120 watts at 25 degree Celsius – you'd need to at least double them (for example, I used two pairs for a 60 watt amp) but don't bother; there are other issues as well:

TIP41 is likely too slow for VAS. Worse than that, the Miller compensation is omitted, which means the amplifier will oscillate and die in a matter of seconds. Not all amps require a Miller cap at VAS but as a rule all amps with differential input need one. The problem is worsened by the lack of Zobel network in the output.

Thermal compensation is "handled" only with a simple diode string. This indeed works in many low power amplifiers or in amplifiers that use Sziklai output instead of Darlington. However, in this case I very much suspect that the amp will go into thermal runaway and destroy itself. A fact helps that only two diodes are used for compensating the forward voltage drop of two Darlingtons (four diode junctions): This puts the amp to work in class-B but at the expense of plenty of crossover distortion. That design would really need an adequate biasing circuit.

The emitter current of the differential pair should stay as constant as possible. You need more than one resistor to achieve that. Most amps use at least an RC filter in series with that resistor. While that solution is far from excellent it is still better than nothing.

There are no current limiting schemes so if you happen to short the load (or simply load the amplifier too hard) it will be toasted. Protecting an amplifier from a short circuit requires a thoroughly planned and fairly complex design that usually limits both VAS and output stage current. Sufficient schemes in guitar amps are a rarity.


In crowded forums this circuit comes up once a year and the same issues are always pointed out. There has been a lot of discussion about this amp in DIYaudio forum.

http://www.diyaudio.com/forums/showthread.php?s=&threadid=103632&highlight=
http://www.diyaudio.com/forums/showthread.php?s=&threadid=100618&highlight=

I think someone asked something about this amp here as well. John Fisher would really do a favour for many if he would remove the concerned website or give it a major update. If you are lucky you can tweak that design to run sufficiently from lower rail voltages. Reputedly someone got it to run at about 30 – 40 watts. Yet, it would still suffer from many issues presented. Again, why bother since there are a lot of better designs out there.

You want to build a 100 W solid-state amp: Why don't you look for an example of a commercial guitar or PA amp design in that power range? There sure are plenty. You will see that while the main topology is pretty much similar there are some important details, which separate poor, unreliable amps from good reliable ones. Take a look at Lab Series L5 power amp (in lower right corner)
http://www.netads.com/~meo/Guitar/Amps/Lab/Images/Lab_Series_L5_L7_L9_L11_scheme.jpeg
It is a hundred watt amp. See anything different? Thought so.

Rod Elliott's power amp in his guitar amp MkII project (http://sound.westhost.com/project27.htm) also looks sufficient. Note that while it has at least a good attempt to limit output current it would still die under sustained overload.

If you look for schematics tailored for certain output devices, then at least Fender used the TIP142/147 pair in some of their amplifiers. You can likely find examples close to following:
http://www.mif.pg.gda.pl/homepages/tom/files/65W.gif

I can imagine that the Fisher's design looks tempting because of its simplicity, yet that is also the cause for most of its issues. Why not build a chip amp? While I am not personally very interested about them they are about the best option for a first project when you want adequate amount of power. They are simple and easy to build and even better: They pack a bunch of protective features inside so they are quite reliable as well.


In my opinion, with that SansAmp pre you might wish to avoid typical guitar amp designs: The SansAmp already processes the sound pretty heavily, providing a "ready-to-record" tone, so you likely have more luck with "Sterile" and rather linear designs that do not colour the sound any further. The SansAmp already mimics a guitar amp, its speaker, cabinet and miking so you really don't need a signal chain that does it again. Since guitar amps have a habit of emphasizing these things and, in principle, processing the signal a lot (some even in the power amp section) a PA amp with a speaker cabinet tailored for linear response might be a better option.


As a final point, although the Fisher amp's power supply schematic does not show them, do not build anything like this without fuses!

Edit: The transformer under the link is no way suitable for a 100-watt amplifier: Check out the VA rating: 35V x 1.5A = 52.5 W. This is about half of hundred watts while you typically need a VA rating that is at least two times the required output power.

35V CT typically means that you get bipolar supply voltages of about ± 23 VDC (35 / 2 * 1.414 - diode drops from rectifier). The 35V (RMS) stands for the reading across the entire secondary. The center tap then splits this. Rectifier configuration matters as well. Anyway, the 23 VDC is about the optimum you get from that but for building a high power amp this is too low.

The current rating of 1.5 A is too low as well: For example, 100-watts to 4 ohms requires at least 5 amperes of current delivering capacity. You need both sufficient current and voltage or the supply will sag under sustained loading and the amp will begin to exhibit sooner clipping and loss of output power.

For example, a pretty good transformer would provide 5A x 48V (RMS) = 240W (VA). You see now why the need for a rather high VA rating? Plus we have a fact that realistic amps are never 100% efficient: Assume 60% - 70% max. for class-AB. So, in conclusion: The transformer is not sufficient but likely you now have a better idea about the properties required.

[deadspeaker]

Wow thanks for the great reply. It might be the best reply I have ever gotten. I'm defiantly not going to build that amp now. The link to that large schematic for a power amp looked complicated, but not impossible. I don't know if I can afford to make that one yet, but I guess I'll book mark that and search for other power amps for now. What kind of transformer would be good for that particular amp. And is this a popular build? Just asking in case I need help with it. Thanks again for the great reply.