Circuit Analyses: Why is this stage necessary?

[Ethan]

Is this stage (in grey) from the LA4 necessary? Other than for input gain selection and the overload LED, my noobie eyes can't see a reason for it?

(I'm excited as I think I am actually learning something and able to decipher meaning from a schematic )

Here's what my noobie eyes see:
When the input switch is set to "HI" it's at unity gain (almost), and on
Lo" the gain is 10. The overload LED will start to light with the output at around 9V, and much brighter at 10V and over. There, that's the extent of my knowledge.

 

[gyraf]

Yep, it's a 1x / 10x amplifier. The overload indicator will show precisely when you start to get distortion - mostly because it causes most of it for the first 5-6dB's :razz:

Jakob E.

 

[Ethan]

Thanks for the confirmation Jakob!

Is that stage really necessary though, other than just for the input gain switching and the overload LED? To me it looks like that stage could be eliminated without a problem? :idea:

 

[Ian MacGregor]

The stage has inverting gain, so removing it would cause a polarity flip through the comp. You could swap the COM and +/- input connections to "un-flip" the reversal.

Ian

 

[Kit]

"To me it looks like that stage could be eliminated without a problem?"

If you dont need that 20dB boost, then yes.

 

[Ethan]

Thanks for the help.
Jakob what did you mean by:
[quote author="gyraf"] The overload indicator... causes most of it for the first 5-6dB's :razz: Jakob E.[/quote]

Were you saying that the diode and zener there causes distortion? Could you please explain?

Thank you.

 

[gyraf]

..draws a lot of current at assymetrical spots of the sinewave, off of an already hardworking RC4136 opamp..

 

[Ethan]

Assymetrical spots of the sinewave

Interesting... I think I understand :idea:
Since only the "positive" portion of the signal will pass through the diode it will cause that portion of the waveform to distort badly as the LED begins to light.

Would a bridge rectifier work here? At least that way both sides of the waveform would be used. Would the addition of a current limiting resistor for the LED further help to limit how hard the opamp has to work (and extend the life of the LED) when the LED begins to light?

This is really cool, I almost feel like I'm beginning to understand a few things :shock:

 

[bcarso]

[quote author="Ethan"]

Assymetrical spots of the sinewave

Interesting... I think I understand :idea:
Since only the "positive" portion of the signal will pass through the diode it will cause that portion of the waveform to distort badly as the LED begins to light.

Would a bridge rectifier work here? At least that way both sides of the waveform would be used. Would the addition of a current limiting resistor for the LED further help to limit how hard the opamp has to work (and extend the life of the LED) when the LED begins to light?

This is really cool, I almost feel like I'm beginning to understand a few things :shock:[/quote]

The bridge will make the distortion odd-order (maybe not too bad an idea), and the resistor will soften the threshold (undesirable). I think they must have supposed that there would be so much distortion at the point the LED starts to come on that distortion from the amp would be irrelevant---i.e., if it's flashing, back things down!

 

[Ethan]

Awesome. Thanks for the help guys--I've been going back to look at schematics I've seen before trying to make more sense of them as I learn more. This has been greatly helpful.

Another question if I may
Does the overload indicator portion (Diode-Zener-LED) add distortion when the signal is NOT approaching distortion?

 

[gyraf]

Does the overload indicator portion (Diode-Zener-LED) add distortion when the signal is NOT approaching distortion?

Nope - only when the zener starts to conduct.

 

[Matt Syson]

Hi
If you were thinking of 'acquiring' this circuit, then use another op amp section to drive the LED, separate from the main audio path. Obviously the guys at UA couldn't afford one more op amp section.
Then you could go with a diode bridge and catch both pos and neg peaks.
Mind you this would be adding the fearful cost of 3 more diodes!
Matt S

 

[bcarso]

[quote author="Matt Syson"]Hi
If you were thinking of 'acquiring' this circuit, then use another op amp section to drive the LED, separate from the main audio path. Obviously the guys at UA couldn't afford one more op amp section.
Then you could go with a diode bridge and catch both pos and neg peaks.
Mind you this would be adding the fearful cost of 3 more diodes!
Matt S[/quote]

Those guys were minimalists and in those days opamps were pricey. The very use of the 4136 is an indication of the era this was designed.

I agree---separate buffer for the indicator, or a transistor or two.

 

[jhaible]

I agree with everything that has been said. Just playing the devil's advocate here:
Could it be that the clipping from current limiting (driving the LED) sounds more pleasant than the opamp's getting stuck at its rails?
It's clear that the former happens a bit earlier; you would have a little bit more headroom with a separate LED driver. But if there's a difference in sound (I don't say there is - I just see the possibility!), it may be worth to sacrifice this.
(That said, of course the possibility of clipping in the input stage is a design flaw right from the start, IMO.)

JH.

 

[Matt Syson]

Hi
When you start to look at it there are several aspects that are rather poor. It is just as well that UA did it then, and not a 'cheap' manufacturer doing it now otherwise they would be slated for it.

The input buffers should drop say 3 - 9 dB so that when confronted with a signal from amp with 17 volt rails it does not clip. As there is possibility of gain (-10 position) then DC blocking would be a good idea to remove offsets from incoming equipment.
The lack of any resistance in the LED chain is a bit antisocial to the op amp. I suppose it may help with latch up, I can't remember of these particular opamps will or not. Surely doing it symmetrically would be nicer?
Matt Syson

 

[bcarso]

I still know some of these guys and I can ask (at least those still living!), but from the preponderance of information I've gotten, the overriding variable was parts minimalism and cost to achieve an acceptable performance, in the era when parts and labor were costly. The raw LED etc. load is consistent with a desire for a sharp threshold, taking the eye's persistence of vision into account and the desire to sense short peaks. And as I remarked the limiting is working hard at that point so already distortion is significant anyway.

It's true that they were aware of the coloration sentiment around second-harmonic, but I doubt that entered in. These guys did listen though, even if they were not perfectionists.

IIRC 4136 parts were relatively free from the latchup problems with slightly older opamps. They were considered good parts in their day, even if their day has long passed!