preventing opamp from overdriving

billonious · Post by **billonious** » 19 Dec 2010, 18:22

this is a kind of "thinking loud" post, I just want to share thoughts and take ideas from other users.
The background is that regarding high gain ss distortion, many guitarists, including myself, find the built-in distortion better than of stompboxes (more alive, roar, less dirty though high gain ...). I conclude that the cause is the supply voltage. Opamps in amplifiers are fed with +-15volts. Clipping is done mostly at diodes (leds/si) and less at opamp itself. Contrary, stompboxes supply is +-4,5Volts, pushes the opamp to rails quickly and clips deadly before the signal reach at diodes. No doubt how dirty is the distortion of an heavily overdriven opamp.

So, the key to prevent (or reduce significatly) opamps' own clipping without sucrificing the high gain. Let's move in my next though.

Let's say, I am building a 9V stompbox with some (3-4) low gain opamp stages (x5 - x10 each one), with clipping diodes/leds in each feedback to prevent opamp from clipping (rather than produce soft knee distortion), and maybe one-two voltage dividers across the signal. Also, some bass & treble roll of and some assymetry of the signal. Total gain is about 2500. This circuit could (I am just wondering) produce a polished and noble high gain distortion. Something like the transistorized Big Muff but with opamps. I draw a draft schematic to make my thoughts clear.

Is there something similar in the stompbox land?

Hides-His-Eyes · Post by **Hides-His-Eyes** » 19 Dec 2010, 18:45

Nope, I can honestly say I prefer pedal distortion.

Anyway, in theory the voltage in the op-amp never gets above V_fwd of the diodes; so there should be no op-amp clipping going on... Unless the signal in exceeds about 3V p-p!

billonious · Post by **billonious** » 19 Dec 2010, 18:54

even for palm mute playing?

Hides-His-Eyes · Post by **Hides-His-Eyes** » 19 Dec 2010, 18:55

Ah, like more like heavy metal distortion? Out of my expertise.

Check my edit, anyway.

billonious · Post by **billonious** » 19 Dec 2010, 19:03

Hides-His-Eyes wrote: Anyway, in theory the voltage in the op-amp never gets above V_fwd of the diodes; so there should be no op-amp clipping going on... Unless the signal in exceeds about 3V p-p!

this exactly is what I mean

Website · Post by **Lonkero** » 19 Dec 2010, 19:45

Intresting topic!

A couple more theoretical questions to the table

Lets say you wanna achieve a signal amplified as much as possible, with as low as possible amount of distorting

So could the one key be in the voltage and raising it to get more headroom?

How much can you use voltage raising to prevent signal getting distorted?

Dirk_Hendrik · Post by **Dirk_Hendrik** » 19 Dec 2010, 20:18

It's worth to give this matter some thought from the perspecive of achieving maximum distortion at minimal gain.

But an opamp gain stage with diode limiting (clipping) will not amplify up to the rail voltage. The diodes prevent that.

billonious · Post by **billonious** » 19 Dec 2010, 20:50

Dirk_Hendrik wrote:It's worth to give this matter some thought from the perspecive of achieving maximum distortion at minimal gain.

But an opamp gain stage with diode limiting (clipping) will not amplify up to the rail voltage. The diodes prevent that.

I have thought it and voila the answer: Use two pairs of green led in series with a large resistor (: 10k) in feedback plus a rail2rail opamp. Two green leds conduct at 4V, plus the resistor ...If knee is to be produced, it will be a very-very soft (what is called compression), dynamic, with much headroom.

billonious · Post by **billonious** » 19 Dec 2010, 20:52

Lonkero wrote:Intresting topic!

A couple more theoretical questions to the table

Lets say you wanna achieve a signal amplified as much as possible, with as low as possible amount of distorting

So could the one key be in the voltage and raising it to get more headroom?

How much can you use voltage raising to prevent signal getting distorted?

Yes, using two 9V batteries (18V) with a rail to rail chip could make our lives better. To be specific I have to check tlc2262 datasheet.

Gerry de la Sel · Post by **Gerry de la Sel** » 19 Dec 2010, 20:56

billonious wrote:, many guitarists, including myself, find the built-in distortion better than of stompboxes

Let's say, I am building a 9V stompbox with some (3-4) low gain opamp stages (x5 - x10 each one), with clipping diodes/leds in each feedback to prevent opamp from clipping
Is there something similar in the stompbox land?

I don't get it. You say you prefer amp distortion, but you've designed a distortion pedal? All distortion pedals already use diode clipping, just like amps; they don't rely on opamp clipping. SS amp distortion is nearly always just a tubescreamer variant circuit, inside the amp. Apart from the rail voltages they're basically the same thing.

billonious · Post by **billonious** » 19 Dec 2010, 21:08

Gerry de la Sel wrote:
billonious wrote:, many guitarists, including myself, find the built-in distortion better than of stompboxes

Let's say, I am building a 9V stompbox with some (3-4) low gain opamp stages (x5 - x10 each one), with clipping diodes/leds in each feedback to prevent opamp from clipping
Is there something similar in the stompbox land?
I don't get it. You say you prefer amp distortion, but you've designed a distortion pedal? All distortion pedals already use diode clipping, just like amps; they don't rely on opamp clipping. SS amp distortion is nearly always just a tubescreamer variant circuit, inside the amp. Apart from the rail voltages they're basically the same thing.

Read the whole post. I explain why pedals with diodes are not just like amps. Because the opamp's overdrive mess the sound, diodes only clipping is much more gentle. What do you mean that they don't rely on opamp clipping? The Rat (an example of dirty rotten distortion) doesn't rely on opamp clipping, is it possible?

Gerry de la Sel · Post by **Gerry de la Sel** » 19 Dec 2010, 22:00

Read the whole post. I explain why pedals with diodes are not just like amps. Because the opamp's overdrive mess the sound, diodes only clipping is much more gentle.

The opamps in pedals don't clip (unless you're accidentally overdriving a clean pedal?).
Both distortion / overdrive pedals and amps use diodes to clip the signal. They nearly all use the same two or three basic circuit topologies.

What do you mean that they don't rely on opamp clipping? The Rat (an example of dirty rotten distortion) doesn't rely on opamp clipping, is it possible?

The Proco Rat uses diodes for clipping.

billonious · Post by **billonious** » 19 Dec 2010, 23:26

Gerry de la Sel wrote:
The opamps in pedals don't clip (unless you're accidentally overdriving a clean pedal?).
Both distortion / overdrive pedals and amps use diodes to clip the signal. They nearly all use the same two or three basic circuit topologies.
The Proco Rat uses diodes for clipping.

If you are correct that opamps don't clip (that seems nosense for me ), please explain me with simple words what will happen if:
1. you set gain of 3.000 with input signal 100mV and 9V supply. You still insist that it won't clip ??????

What about when an op-amp can't put out more voltage than you feed it?
2. you put out the diodes from Rat, you think you 'll get a clean signal ?????????????? You insist that diodes do the job in Rat?

A third person could explain it to you better than me, because I am incapable of. Read in this page http://www.geofex.com/effxfaq/distn101.htm, Jack Orman writting:

Code: Select all

 "In poorly designed circuits, the opamp gain is high, and the diode characteristics are lost, giving just a lower-voltage clipping op amp, which suffers from all of the problems of the bipolar common emitter stage, and perhaps a few of its own."

Or, just google for "opamp clipping".

Sorry for being ironic & rude but either my electronics knowledge is crap, or the sun is spinning around earth

Hides-His-Eyes · Post by **Hides-His-Eyes** » 20 Dec 2010, 00:20

gain of 3k, input signal 0.1V, = 300V, diode clipping gives a square wave in this instance. You don't seem to understand the role of the diodes. They prevent the gain from ever reaching > (V_fwd/V_in) =~ 6 in your example; any higher and the negative feedback "undoes" your gain.

billonious · Post by **billonious** » 20 Dec 2010, 00:35

Hides-His-Eyes wrote:gain of 3k, input signal 0.1V, = 300V, diode clipping gives a square wave in this instance. You don't seem to understand the role of the diodes. They prevent the gain from ever reaching > (V_fwd/V_in) =~ 6 in your example; any higher and the negative feedback "undoes" your gain.

why you people don't read carefully what is written? This example is for Gerry de la Sel to understand that his claim of "opamps don't clip" is nosense. Explain to him, not to me. You really seem to read two-three words only per sentence before replying. What you write is obvious for me. In the example of gain of 3k, it is supposed that no diodes are placed to let the opamp reach the rails. What you people don't understand? Are my english obscure?
Please read carefully before tutoring because you are out of topic!

Hides-His-Eyes · Post by **Hides-His-Eyes** » 20 Dec 2010, 01:49

A rat is a good example; the diodes are to ground and so there is a chance of the op-amp clipping itself; sorry if you were talking about diodes to ground and we diodes in the feedback loop. You're welcome here even if english isn't your first language, but please don't get angry if we misunderstand you.

Sources of drive in a RAT:

-Slew distortion from the LM308
-op-amp rails (generally avoided deliberately)
-diodes to ground
-clipping the output buffer? (seems unlikely)

however, I just did some maths and it does seem possible to clip the op-amp in a traditional rat circuit depending on the frequency: ie infinite frequency = caps are a short = gain ~300 at least; largest possible input voltage without op-amp clipping = 10mv, approx? If it was 100 I might be more cynical, but that seems somewhat conclusive to me.

RnFR · Post by **RnFR** » 20 Dec 2010, 03:04

i think Dirk was talking about switching out the FET in a Rat for a Ge BJT in order to have different clipping characteristics. not sure if he was talking about a Ge emitter follower or a inverter, though. if indeed it was a buffer he was talking about, it would indicate that overloading the FET input would create distortion. i'm sure it's dependent on the FET used as well.

DrNomis · Post by **DrNomis** » 20 Dec 2010, 07:35

If you have an Op Amp set up as a basic amplifier circuit,and you run it from a +/- 15V supply,the output is going to be fairly linear untill the output level gets to within about a volt of the supply rails,where it starts hard clipping,this is a characteristic of all amplifiers,the output can never exceede the supply rails regardless of how much gain the amplifier has.....

billonious · Post by **billonious** » 20 Dec 2010, 08:21

DrNomis wrote:If you have an Op Amp set up as a basic amplifier circuit,and you run it from a +/- 15V supply,the output is going to be fairly linear untill the output level gets to within about a volt of the supply rails,where it starts hard clipping,this is a characteristic of all amplifiers,the output can never exceede the supply rails regardless of how much gain the amplifier has.....

DrNomis, your post was my first initial thought for which I wrote this topic. So, the ss preamplifier of a amplifier has much more headroom and much less (undesirable) opamp's distortion than a pedal, even if the pedal is the same exactly circuit!.

The point is that diodes are not to be blame for the harshness of solid state designs but the opamps themeselves. So,
an overdrive pedal with diodes (either in feedback or to the ground) could sound 1000% better as long as opamp's distortion (due to opamp's overdrive) is eliminated. It could be eliminated with, fx, low gain of each one stage, feedback leds, voltage dividers to lower signal.

DrNomis · Post by **DrNomis** » 20 Dec 2010, 08:37

billonious wrote:
DrNomis wrote:If you have an Op Amp set up as a basic amplifier circuit,and you run it from a +/- 15V supply,the output is going to be fairly linear untill the output level gets to within about a volt of the supply rails,where it starts hard clipping,this is a characteristic of all amplifiers,the output can never exceede the supply rails regardless of how much gain the amplifier has.....
DrNomis, your post was my first initial thought for which I wrote this topic. So, the ss preamplifier of a amplifier has much more headroom and much less (undesirable) opamp's distortion than a pedal, even if the pedal is the same exactly circuit!.

The point is that diodes are not to be blame for the harshness of solid state designs but the opamps themeselves. So,
an overdrive pedal with diodes (either in feedback or to the ground) could sound 1000% better as long as opamp's distortion (due to opamp's overdrive) is eliminated. It could be eliminated with, fx, low gain of each one stage, feedback leds, voltage dividers to lower signal.

Well,if the Op Amp is being used as a kind of "clean booster",it has to have more headroom than the diodes,otherwise at some point the Op Amp will start to clip due to the supply rails limiting the output,I found this to be a problem with the Tube Driver type circuits,since the Op Amp has a tendency to clip and contribute it's harmonics to the Tube's harmonics when the Drive control is maxxed....

All amplifiers,including ss preamplifiers will distort if a large enough signal is fed into them,and this is because the output of an amplifier cannot excede the supply rails....

Once the output tries to get larger than the supply rails,it will be clipped hard.....