Power dissipation and resistors...noob here

[Marcel_D]

Well this is the amp I've decided to take a whack at building. What better way to learn than by doing eh? Anyways the only thing that has me stumped is how many watts of power the resistors should be able to dissipate. This is just my estimation, but I believe all the resistors over 220R ohm needn't be more than 1W. It's the 10, 0.25 and 0.1 ohm resistors I'm unsure about.

The power supply is 20V and the amp itself puts out 2.6A with an 8 ohm load. If someone could possibly show me how to calculate the proper values I would be very grateful.

[CBS240]

Hi

Well, with 3 mA divided into each leg of the input diff, dissipation isn't higher than 60mW. The VAS resistors are burdened with less than a 16th of a Watt each. Good ol' Mr. Ohm and his law The 10 Ohm is the load for the Zobel filter. This is to sink high frequency RF junk that might get through to the feedback loop and cause instability. I'd go for at least 1/2W here. You are missing the output inductor||10 Ohm (1/2W) resistor in series with the speaker at the output. This can be made simply by wrapping some 22 gauge magnet wire around that 1/2W resistor and solder it to each end. It helps to shield the FB loop from RF picked up by the antenna, err I mean speaker/wire load. The world is a noisy place. As for the .1 Ohm resistors, at least 1 W would be better but physically, you could get away with less. A larger power resistor typically has less noise for the same amount of Pd than a small one. The type of resistor is important wrt this issue as well. Larger resistors can impede your layout so you would want to use smaller sizes for the smaller power input, VAS, and driver. I don't like to do more than 40% Pd, one man's opinion. Metal film are quite nice for the price.

I wonder if this amp could have a single end supply and be capacitive coupled. Might be a good way to test if it is hooked up right initially, might prevent an accident.







The 2 transistor CCS could benefit from a small cap from collector to emitter of the BC549. They can sometimes oscillate.

[Marcel_D]

Thanks for the insight and explanation. I never thought of the speaker as being an antenna but it makes sense. I suppose anything picked up by it would go straight to the feedback loop and amplified to audible levels.

I'm ordering parts for this at Digi-Key. They seems to have decent prices and a bloody massive selection of parts. Right now I'm stuck at the transformer. What makes a fist sized piece of iron and copper $150?! Ugh I'm a student, not some rich fool. I may have to go dumpster diving for parts...maybe I can use an old computer PSU or something...

I know enough about electronics to design basic circuits and grasp a bunch of concepts but I haven't done enough mathematical or practical work yet. It's all so interesting to me though so I have no plans on stopping any time soon.

[woody]

You might check out apexjr.com . He has a lot of cheaper surplus
transformers and caps. By the way that amp will need rather
large heatsinks and they are expensive also.

[AndrewT]

if you have a tight/small budget, then a ClassA single ended power amplifier is not the way to waste money.

Re-consider whether you want (can afford) to go push/pull ClassA or push/pull ClassAB or single ended ClassA or single ended ClassAB.

[Marcel_D]

Originally posted by AndrewT
if you have a tight/small budget, then a ClassA single ended power amplifier is not the way to waste money.

Re-consider whether you want (can afford) to go push/pull ClassA or push/pull ClassAB or single ended ClassA or single ended ClassAB.

Heh yeah it seems to be that way...but the appeal is too strong. All I own are Class AB and D and I've never heard a Class A in person (except for single transistor bench tests) so I really wanted to build one to see what it's all about.

Also that Apexjr site is pretty neat. Lots of cheap interesting parts to be bought there.

Also as far as thermal management is concerned, I'm planning on using something like these .If it can manage a CPU putting out 100W of heat I'm sure a pair of NPN bipolar transistors is no match for its cooling abilities. I should also be able to halve the fan speed making it relatively silent.

[eyoung]

Hola Y,all...

I am sure Mr Elliot could shead some light on your questions as that is his schematic. He also sells kits and boards as well as a bunch of info on his designs are available on his web site..

Regards, Elwood

[AndrewT]

2.6A and +-20Vdc is a dissipation of 104W.
That CPU cooler has no hope of keeping the Tc of the output devices below 50degC.

I wonder if 4 CPU coolers, fitted one to each device, could cool this ClassA stage adequately with the fan speed turned down.

[Marcel_D]

Originally posted by AndrewT
2.6A and +-20Vdc is a dissipation of 104W.
That CPU cooler has no hope of keeping the Tc of the output devices below 50degC.

I wonder if 4 CPU coolers, fitted one to each device, could cool this ClassA stage adequately with the fan speed turned down.

I was thinking one cooler per 2 transistors. Bridge them with a chunk of aluminum and stick the cooler on top. This actually leads me to the question will I be able to get away with this since the aluminum will create a path between the collectors of the transistors, correct?

[CBS240]

Yes, so you have to use mica or some other insulator and heat sink compound. Pretty standard.