The Worst Distortion

[JCM]

There is one inherent advantage in class A amp. Please remember the horrible sound - to me the Worst kind of distortion there is - in many class AB amps - the kind you find in hi-fi shops. When you turn the volume to 4 or even 3.

Every amp I had sounded the same. My old 50 Watt Sansui with feedforward, had a little 25 W Marantz, quadraphonic JVC, my clunky old Radio Shack 50 Watter, the 70 (?) watt Curtes Mathes amp I just repaired my little Sherwood... not much difference.

They all have (had) skimpy power supplies. The power amp reservoir caps are smaller probably than a diyer would use in a Preamp. Because every transformer has internal resistance, this is a part of the REAL schematic not shown. You never see this distortion on a scope running a sine wave at a constant level. You would have to have some kind of burst generator, etc.

Class A amps, running at a constant current drain, don't have this distortion. They can't. The signal never demands any sudden current - it averages out to a net zero change (theoretically for a 50% duty cycle waveform - if I understand rightly). For that reason I don't see a problem in using some resistors in the power supply for filtering. That would be a disaster in a Class AB design. In Class A bias it merely reduces the supply voltage and reduces hum. I don't believe this is an example of clipping, though it causes that.

In my experiments, this bears out. Deliberately driving my experimental class A amp into distortion, I noticed the sound only got distorted - sounding ugly, of course, but that effect I Loathe was absent - that dynamic range compression which makes the music sound like it's getting buried under an avalanche is not present. I never want to build an amp capable of that ugly sound.

What is the name of this distortion ? Does it have a name ? I haven't read anything I can remember about this anywhere, yet it obviously Must have been addressed.

If this type of distortion were to be listed along with the other specs, a whole lot of amps would be looking pretty sorry. 0.05 % distortion ? The truth is, it may be more like 20 % I think, under REAL conditions. This is not make-believe, it could be measured. I do wonder about how pernicious this effect may be.

Time and time again I have turned up the volume (even only driving headphones !!) and had the music suddenly turn sour. Even at a setting as low as 2 (which makes one suspect it's something else in that case; whatever it is, it's bad).

This is one of the reasons I like class A bias better. I know that crossover distortion is supposed to be hideous - but in my experience it isn't as bad as the other - and I wonder how much is Really audible, especially in some of these designs you see where the designer can't even measure it. There was one amp I recollect - I Know it was AB - and it was maybe the best I ever heard - it also weighed and cost a lot. I believe it had a real power supply inside.

It was natural enough that I found Nelson Pass's site and looked at some designs. I built a kind of Zen amp, the performance of which I want to describe next.

I am beginning to explore hi fi.

What I have said here is, I think, somewhat obvious, seeing all the diyer amps with gigantic cans in the power supply. As long as you put those in your class AB amps, and hopefully a stout tranny, none of this will even be on your radar screen. It's what you might experience when you buy some hi-fi type amp, very real and ugly sounding. It's one big reason a diyer can create a superior amp.

[Tweeker]

Some people like some R in AB powersupplies, but then they are using diy amounts of capacitance- 35k uF+ per rail.

One approach in AB tube amps is the use of a swinging choke LC filter.

Alot of guitar amps have deliberately droopy rails, but this is instrument use and not hifi.

[Tweeker]

Im sure a class A amp with a similarly skimpy PSU would sound wretched too, its just that they cant try to get away with as skimping much there. For reliability reasons if nothing else. In the AB amp they have more PSRR to try to lean on.

[moe29]

... i think you're preaching to the choir.

[Nelson Pass]

Originally posted by JCM
I am beginning to explore hi fi.

Now that's what we like to hear about.

[steenoe]

Now that's what we like to hear about.

Just what I thought, replying to JCM's other thread

Steen

[JCM]

Thanks for all responses :-)

Yes, Tweeker, I have noticed that, in some guitar amps. An example : A little 12 watt Marshall. The schematic showed a bipolar supply, two 1000 uF caps, and the preamp op amp fed from a 1 k resistor and decoupling cap. The whole amp was very simple.

It sounded MUCH better than the Peavey Decade, which used a lot more capacitance in the supply, and more complex.

Maybe a good reason for resistance in the supply of an AB or any amp is to try to reduce something that preys on my mind, which is the thought of harmonics being sprayed into the supply rails b/c of the bridge diodes (?) On both my pwr amp & pre amps I inserted a power resistor between the bridge and caps, in the hope of reducing it. I couldn't tell a difference, tho, I suspect it does nothing. I can imagine it may reduce clipping harshness on AB amps.

[Brian Donaldson]

Talk about hideous big box POS. I have a Sony HT reciever that says 100W x 5 on the front, but says energy consumption 370 watts on the back. Only sound OK when it's switched off.

There used to be FTC laws against this kind of crap.

[DSP_Geek]

Originally posted by JCM
Maybe a good reason for resistance in the supply of an AB or any amp is to try to reduce something that preys on my mind, which is the thought of harmonics being sprayed into the supply rails b/c of the bridge diodes (?) On both my pwr amp & pre amps I inserted a power resistor between the bridge and caps, in the hope of reducing it. I couldn't tell a difference, tho, I suspect it does nothing. I can imagine it may reduce clipping harshness on AB amps.

That resistor isn't really in the right place, because you're still charging/discharging the cap with pulses from the rectifier. Try an RC network after the first capacitor, and bypass the second cap with a good-sized polypropylene.

[JCM]

>Talk about hideous big box POS. I have a Sony HT receiver that >says 100W x 5 on the front, but says energy consumption 370 >watts on the back. Only sound OK when it's switched off.

>There used to be FTC laws against this kind of ****.

That reminds me - bought a Radio Shack 50 Watt / ch receiver, the Optimus STA 5500. (I'm not sure what I was thinking then.) The power consumption : 95 W. How that amp manages to break the laws of physics, I'm not sure.

I've noticed, along with bell bottoms and sideburns and all, another fad of the 70s : a lot of "100 Watt" boom boxes all over the place. They pulled that in the 70s, and yep, they're at it again, even the "hi fi" products.

>That resistor isn't really in the right place, because you're still >charging/discharging the cap with pulses from the rectifier. Try >an RC network after the first capacitor, and bypass the second >cap with a good-sized polypropylene.

I tried different small caps, but couldn't see any difference on the scope. Mainly it's broadband noise now (there's an 8 Henry inductor in the preamp supply). My thought was that while the pulses remain, the amperage would be reduced. For both supplies, I used so many capacitors, it's almost the whole thing, & 4 resistors in the pwr amp supply, so there's 4 RC networks.