More questions about the Flatline

[David]

I just stocked up on parts yesterday to build a Flatline.  The schematic calls for a TL072 as the op-amp.  I saw something on the Web yesterday that indicated that this chip was not a good choice for a comparator.  This greatly concerns me because it appears that this is exactly what one of the op-amps is doing.  Could this be why the Flatlline has been called "problematic"?  It just so happens that I was planning to use a LF353 op-amp instead because I already had it.  Does anyone know if this is subject to the same problems as the 072 in functioning as a comparator?

Thanks in advance...

[Mark Hammer]

Neither of the op-amps in the Flatline are being used as a comparator from what I can see.  One is a basic non-inverting gain stage, and the other is a rectifier.  In the original drawing, John specifies a TL072, so I think you should be okay.  John's designs are of the minimalist type that can sometimes depend on the quirks of the recommended components to achieve functionality.  So, if he scribbles down TL072 then I would imagine it safe to use one.  There may be other duals that sound or perform "better" (whatever that is), but the TL072 should work.

[David]

A rectifier?  I sit corrected, Mark.  My apologies...    :oops:
If I might humbly ask:  how does a rectifier cause compression?  I saw the Vbias connection to op-amp B and I guess I assumed voltage comparison was going on.

[Ansil]

A rectifier?  I sit corrected, Mark.  My apologies...    :oops:
If I might humbly ask:  how does a rectifier cause compression?  I saw the Vbias connection to op-amp B and I guess I assumed voltage comparison was going on.

basically from what i understand from the circuit.  the input signal goes through the first stage which is in fact a variable gainstage, due in part to the pot and the ldr there.   when the signal coming out of the first gainstage hits the rectifier the amount of light that hits the ldr detrmines the feedback loop resistance which drops the gain to said level.  for instance if you hit it really hard and the led lights up enough to drop resistance to 5k hypothetically.  which basically would make the gain of that circuit, unity.

if i am wrong let me know mark

[Ansil]

damn double post.  heres something more interesting to look at than a damn dirty double post.

[Mark Hammer]

As Ed McMahon would say to Johnny Carson "You are correct, sir!"

[Ansil]

As Ed McMahon would say to Johnny Carson "You are correct, sir!"

wooooo hooo i was right..  hot damn

[David]

GAAHHH!  I just realized that the schematic calls for germanium diodes!!  Will it not work if silicon are substituted?  Why - leakage?  I've got some silicon diodes, but no germaniums (yet) - those babies are expensive!

[Mark Hammer]

I wouldn't predict that it will *not* work, but since you are going to lose some envelope voltage by using 1N914's or something similar, the LED may not glow as bright, so the amount of attenuation of the LDR may not be as drastic, and ultimately the amount of compression difficult to hear.

Adapting this to the use of Si diodes instead of Ge, I'd make the following recommendations:

1) If you can score an ultrabright LED (2000mcd or better), use one of those so you can squeeze more brightness out of a lower driving voltage.

2) Increase the gain of the first op-amp.  Ignoring the LDR, the stock unit has a range of gain from 3 (when the 100k pot is at max resistance) to 23 (when the pot is at min/zero resistance).  The LED brightness will be a partial function of how hot a signal that stage feeds the second op-amp (which is why the compression amount is essentially the stage-1 gain control).  

If you can increase the maximum gain of that stage, you get more envelope signal which begets greater LED brightness which begets greater illumination of the LDR which begets a lower parallel resistance in conjunction with the 220k feedback resistor which begets a drop in the gain of stage-1 which is your compression effect (sounds almost Biblical, doesn't it? ).  You can alter that gain in a few ways, since the gain of the stage goes up as the resistance to ground gets smaller or as the feedback resistance gets bigger.  My gut instinct suggests you opt for a 4k7 or 5k6 resistor instead of  10k, and a 50k pot instead of 100k.  A 4k7 resistor and 50k pot will give a range of gain (assuming everything is exactly on spec) from 5 to 48, which is a broad enough range.  A 5k6 resistor will get you a 5 to 40 range.  This ought to be able to goose the signal enough to light that sucker up in spite of using Si diodes.

3) You may be able to push the LED a little harder by reducing the value of the 330R resistor and upping the gain on the second stage.  Since this is a sort of push/pull circuit, though, I have some misgivings about this sort of adjustment since you'd need to work with both aspects to make it work optimally.  As a result, consider this your last resort because it would be the most labour intensive.

Ultimately, all of this is a question of getting the LDR to drop enough in "on" resistance to make a difference in stage-1 gain.  You may not need more than a single one of these assorted adjustments to have it fly right.

Good luck.