Headroom, clipping diodes and the 1981 DRV?

[soggybag]

I'm looking at the 1981 DRV. It's a rat like op-amp into two Si diodes to ground style clipper. It's very similar to the Rat.

It has one big difference that it uses a charge pump. In my mental model of the situation it seems the signal passed the diodes can't ever get above +0.7 to -0.7v. Seems the charge pump is wasted here. Or it just gets us to clipping faster, is there audible differences that can't be covered by a 9v supply?

The circuit does have a buffered bypass which would possible benefit from a higher voltage supply.

http://3.bp.blogspot.com/-yJP7mB6f8rs/XQ5WS_HXBKI/AAAAAAAAGrA/QwSUspD6idQZ73_d-4o3Rp4l5WK_dPH8ACK4BGAYYCw/s1600/DRV%2Bschem.png

[ElectricDruid]

There's a couple of significant differences, I think:

The first gain stage IC2A will be less likely to produce op-amp distortion because of the extra headroom. So more clipping is coming from the diodes.

However, that's rather undermined by the second difference which is that there's *massive* gain in IC2B, the post-tone-control gain stage. There'll be heavy op-amp distortion from that stage. And that's after any tone control, so you've no further way to trim treble, except by putting less of it in the front. Ok, the 3n3 rolls off some edges, but that's all there is.

So, yeah, at first glance it looks very rat-like, but the gain staging is quite different, I'd say.

[GibsonGM]

If I were to build it, I'd run it on 9V and then compare with 18V.    Just because.  See which is smoother.  Pretty much what Tom said - wonder what 'tone' they were looking for out of IC2B, LOL.

[antonis]

there's *massive* gain in IC2B, the post-tone-control gain stage. There'll be heavy op-amp distortion from that stage.

IC2_B gain is unity plus 0.005 (1 + 10k/2M2)..

Probably a typo for IC2_A (46.5dB gain)

[soggybag]

Am I missing something by thinking the signal after D1 andD2 can't swing past +0.7 and -0.7v. So all the potential gain added by the charge pump is wasted? Though, it does help the buffer.

My mental model makes me think you could hit the clipping threshold on a 9v power supply with little effort and plenty more to hit the clipping threshold.

[Vivek]

there's *massive* gain in IC2B, the post-tone-control gain stage.

Please recheck. If the component values on the schematic are correct, I see roughly unity gain from IC2_B

The first gain stage IC2A will be less likely to produce op-amp distortion because of the extra headroom.

Suppose IC2_A gets signals of 4Vp and its max gain is about 200, we are asking it to output 800V at max.

A) IC2_A will produce massive amounts of distortion at most of the pot settings, irrespective of 9V or 18V supply

B) In the situation of higher gain settings on gain pot, the difference between 9V rail saturation and 18v rail saturation on IC2_A will not be audible.

(I had made graph of harmonic content versus clipping point.)

C) We could be driving ~8ma peak into the diodes

So the only difference I see with 18V supply is that first IC1_B does not clip. But is that a benefit ? slight clipping of first stage could add to the fun !!!



If I were to mod this circuit, I would think about

A) I would give last stage a gain of maybe 5 to 8x, to get massive output

B) I would use better tone controls

C) I would soften the last clipping  by changing CLR before diodes, or make that into a compliance pot

D) Everyone loves switches to play around with different diodes !!!!

E) I would make first stage clip gently too.

[antonis]

Am I missing something by thinking the signal after D1 andD2 can't swing past +0.7 and -0.7v. So all the potential gain added by the charge pump is wasted?

Α +/- 0.7V is further distorted at a gain of 6.4 for 9V supply where it needs the double gain (12. for 18V supply..
(what matters - aurally-, other than output amplitude, is the shape/slope of the clipped waveform)

[ElectricDruid]

there's *massive* gain in IC2B, the post-tone-control gain stage.

Please recheck. If the component values on the schematic are correct, I see roughly unity gain from IC2_B

Yeah, sorry - my bad. You're right. I somehow got R15 and R16 back to front. Ooops! Sorry for misleading everyone.

So there's close to unity gain on that final stage, and a roll-off around 4KHz.

The first gain stage IC2A will be less likely to produce op-amp distortion because of the extra headroom.

Suppose IC2_A gets signals of 4Vp and its max gain is about 200, we are asking it to output 800V at max.

A) IC2_A will produce massive amounts of distortion at most of the pot settings, irrespective of 9V or 18V supply

B) In the situation of higher gain settings on gain pot, the difference between 9V rail saturation and 18v rail saturation on IC2_A will not be audible.

Yes, it'll clip. At x200, pretty much anything will! But the 18V should give you a *bit* more space. It's all a bit moot though, since everything gets sliced down to one-diodes-drop of volume straight afterwards. I suppose there's a slight effect there too - the 18V signal will get clipped slightly less by the op-amp and slightly more by the diodes, whereas at 9V, it's the other way around. But you probably won't notice at max gain - it'll be more subtle settings where that might come into play.

If I were to mod this circuit, I would think about

A) I would give last stage a gain of maybe 5 to 8x, to get massive output

B) I would use better tone controls

C) I would soften the last clipping  by changing CLR before diodes, or make that into a compliance pot

D) Everyone loves switches to play around with different diodes !!!!

E) I would make first stage clip gently too.

All good ideas!

[Mark Hammer]

My contention for some time now is that some pedals are effectively "double clippers".  That is, we think there is only one locus where they clip, but the op-amps preceding the diodes are also clipping because of voltage-swing limitations.  Bear in mind, though, that this reasoning only applies when the clipping diodes are on the op-amp output and going to ground.  If they are in the feedback loop of the op-amp - Tube Screamer style - the op-amp is prevented from reaching it's maximum voltage swing, because the diodes magically transform into ultra-low resistance conductors when the amplification of the signal hits around+/-650mv or so.***This, to my mind, is really the basis of the distinction many make between "hard" (diodes to ground) and "soft" (diodes in the feedback loop) clipping.

In this instance (DRV), IC2a normally pushes the signal well past what the op-amp can produce with a 9V supply, and THEN IC2a's output is clipped again by the diode pair.  Use of a charge pump, and higher supply voltage, will raise the maximum voltage swing, to feed the diodes a generally "cleaner" signal.  Note, however, that C6/R10 provide for a bass rolloff beginning around 1.5khz.  The Rat also provides proportionately more gain for content above that point also, but does so in the same stage as lower mids and bass, such that one increases gain for everything via the Gain control; just moreso for the highs.  I'm not suggesting this circuit is non-musical in any way, but by separating gain the way it does, it does NOT work the same way as a Rat.


***Never dawned on me before, but I wonder what you'd get if you used, say, a 2+2 diode complement to ground, but used a low-voltage op-amp AND a lower supply voltage.  So, if the forward voltage of the diodes was roughly +/-1.2V, BUT one powered the op-amp providing the gain with 5V (or even less!), would that get you "hard" or the equivalent of "soft" clipping?

[GibsonGM]

there's *massive* gain in IC2B, the post-tone-control gain stage. There'll be heavy op-amp distortion from that stage.

IC2_B gain is unity plus 0.005 (1 + 10k/2M2)..

Probably a typo for IC2_A (46.5dB gain)

Ya, I did the same thing, I looked at it backwards. I actually GOT the .005 result, and convinced myself that 'no, that is incorrect - they'd NEVER do that, it's the other way' ha ha!  You weren't alone, Tom...

[soggybag]

Great information, thanks for your insights.

I'm just wondering if you might get more range out of the clipping with some LEDs or by using series LEDs. My reasoning was that the higher clipping threshold would provide more space to use the higher output before clipping?

The other thought was that you could remove the charge pump and and save on some parts.

[Mark Hammer]

Both of those are sensible ideas.  The path to choose would depend on what you want/need the circuit to do for you.