Old Radio Electronics Article? Or am I loosing it?

[lapsteelman]

I have a vague memory of an artical in Radio Electronics magazine in the early 90's about using an audio amplifer IC (I think it was the lm1875) and a low noise op amp in an unusual way. I seem to remember that the power amp was in the feedback loop of the op amp, but it was almost 20 years ago so I could be wrong. I think it would have been around 1993. If anyone has a scan of the original artical or can link me to more information on this circuit technique I would greatly appreciate it.
I was pretty new to electronics at the time, but I think the jist was that using this circuit topology you got all the noise specs and  low THD of the op amp, ( which I believe was a ne5534) and the power amp IC was just acting as the muscle. (again.......20 years ago, forgive me if I got it wrong).

Thanks

[Enzo]

Muscle to do what?   What application would this have?

Maybe it was the other way around, that the op amp was in the feedback loop of the power amp.  I could imagine that somehow massaging the frequency response or reducing distortion.


Or LM386?  That is a lower-power amp IC.

[lapsteelman]

I am pretty sure the LM1875 was the power amp. It could have been the op-amp in the feedback of the power amp. I remember it wasn't the conventional  pre amp/power amp configuration.

The muscle part was it basically gave you the low noise/THD of whatever op-amp amp you were using but with enough drive capacity to move a speaker to 10 to 20 watts.

I used to kill time when my girlfriend (now my wife) was in college by reading the electronics magazines in the library. At the time, I just couldn't get what they were doing in the circuit. I would like to revisit it.

[Enzo]

Well, if it is going to push a speaker, especially 10-20 watts worth, then it is coming from the power amp IC.   I don't recall the article, but then I don;t recall a lot of stuff.

But I would say that you might have better odds at finding someone who knows the article over at the forums on DIYaudio.com, try the solid state section, or possible the chip-amp section.

[Roly]

the jist was that using this circuit topology you got all the noise specs and  low THD of the op amp, (which I believe was a ne5534) and the power amp IC was just acting as the muscle.
...
I just couldn't get what they were doing in the circuit.

It's a cute concept, trouble is, I think it's wrong.  Simply by adding NFB you get a reduction in noise and distortion, and by using a low noise op-amp inside the loop it won't make this any worse, but the idea that a chip amp can inherit the specs of the op-amp in the NFB loop is attractive, but misconceived.

It's true that by adding the op-amp the open loop gain will be increased, and that there should therefore be some reduction in noise and distortion (compared to simple un-amplified NFB) when gain is again reduced to the closed-loop figure, however, the law of diminishing returns means that while N&D will be a bit better they cannot fall to the level of the feedback op-amp alone; that is, there will be some improvement, but it will be marginal.

http://en.wikipedia.org/wiki/Negative_feedback_amplifier

My main concern with such an arrangement would be stability, stopping the whole thing from oscillating.

When I was a teenager I placed absolute trust in every circuit I saw published.  Older and wiser I now know that some real crud slips by editorial staff and into print, and I could give you quite a few examples.


The improvement factor due to NFB is (1 + Beta_FB * A_OL)

Where;

Beta_FB is the proportion of output taken to the input, between 0 and 1
A_OL is the open-loop gain

Let us assume that both the chip amp and op-amp have an open-loop gain of x100,000, and that Beta is 1 (the whole thing is a voltage follower with unity closed loop gain).

For the chip amp alone this will be;
1 + 1 * 10⁵ or 10⁵+1, ~10⁵

For both in a loop it will become 10¹⁰+1, ~10¹⁰

The problem here is that in practical terms a reduction in N&D of ~10⁵ is already so large (0.001%) that making it ~10¹⁰ (0.00000001%) makes no practical difference, and that noise and distortion due to other components outside the loop, such as the source resistance, are very likely to be much greater (e.g. Boltzmann a.k.a Johnson or Nyquist thermal noise in the source resistance, keeping in mind that a good guitar amp will have a 1Meg input resistor).

Mike preamp designs these days are so good that this thermal noise is often the limiting factor, even around 1k input, and to actually measure the improvement in the arrangement above you might need to run it at cryogenic temperatures.

I would suggest you try lashing something like that up on the bench, it's only a couple of chips, if only to satisfy yourself that you can't hear any difference between just the chip amp and the chip-amp plus op-amp - but don't be surprised if it oscillates like crazy at some far off frequency.

[lapsteelman]

I doesn't surprise me that the concept might not be a winner, after all it's been 20 years and it hasn't exactly taken off like wildfire. I would still like to find the original article if possible and reread it, and it might be fun to put one together.

[Enzo]

The whole thing might not get the specs of the op amp, but would it improve the specs compared to the power amp IC alone?   SInce the point of the circuit is a power amp, it would be the goal I would think to improve it as a power amp.

[Roly]

Yes, it would improve the N&D specs of the power amp chip a bit, but a couple of points.

The specs of chip amps today are pretty damn good anyway, and your dominant noise source is normally the first stage of the preamp.  A while back I rebuilt the preamp of my Twin-50 which was all transistor using LM833's and the residual noise is now so low I can't hear it even with all the controls max'ed out.  With the transistor power amp (2x 2N3055's) input shorted it is effectively silent.  As for distortion in a guitar amp ...er... fuzz boxes???

I think the reason this didn't start a bushfire of interest, and I can't recall seeing any commercial amp with gain in the NFB, is that the loop gain would be so high that stability, particularly at supersonic frequencies, would be a real pain.

In its first incarnation my LM833 preamp was vigorously oscillating at 455kHz (now there's a coincidence!) and I had to roll the gain off around 35kHz to get it stable.

Yes, I think it's a fun thing to investigate, lash up, give a whirl, but in the wash up I think the conclusion will be that the extra complexity gives no practical improvement in operation; the noise floor from the preamp will still be vastly greater.  I have a N&D meter on the bench, and it gets used from time to time, but I certainly don't make a practice of using it on guitar amp repairs, or even stereos for that matter.

To me this looks like one of the many (mostly misconceived) ideas that crawl in from the Hyper-Fi world and are best closely examined in the full glare of daylight, then sent packing.  :duh  :trouble

But never let it be said that I discourage experimentalism, all sorts of interesting things turn up during otherwise futile journeys, and who knows what this might throw up; a better fuzzbox maybe. 

So yeah @lapsteelman, go for it, but be sure to let us know what you find, okay?

[lapsteelman]

I kind of want to find/read the article more out of a sense of nostalgia than anything else.

I have never had a problem with the noise or distortion in the chip amps I have played with. ( except when I messed something up  :cheesy:). 

I think the article was geared toward HI FI.

[Enzo]

A quick look found me a number of sites involving the magazine, and I bet library of congress might have them, so could possible scan the tables of contents or otherwise look into the issues from the era.  It would be some work, but what price glory?

[lapsteelman]

I thought there might be some pack rats like me out there who had some back issues. I actually have a pretty good collection myself just not that issue. I'm sure a copy will surface on EBAY or I'll run across it somewhere.

[Enzo]

I have quite a few of them myself, but I don't recall that article.  If we could identify which issue, we might hunt for it.

[lapsteelman]

Pretty sure it was from 93 or 92, and I think the cover was yellow. These are the things that stick in my brain all these years!  :loco

[Enzo]

Oh yeah... the old yellow one...


OK, if I can find those years in my warehouse, I;ll see what I can dig up.  Meanwhile try those other research avenues I mentioned.

[Den.]

Found this:
http://tinyurl.com/b5ydkg5