More op amp clipping questions

[blackieNYC]

I have a simple parametric eq - geofex single op amp gyrators, operating at 9 v. Built  a single band before. Don't think I've been clipping it. But now I'm doing something a little different. Despite my lengthy attempt to convey the details, this may be a simple question for many folks.  Please tell me if I've got this right:

I have a two band geofex parametric, 9volt supply, placed after a distortion circuit (a 4049 "forty-niner" and alternately, a 386 distortion). These dirt circuits are certainly capable of clipping the eq. I have breadboarded and now perfed this, doing final tweaks. It seems I did roll back the distortion's output level, dialed up a boost in the eq, and was able to avoid clipping the op amp, but I wish I had spent much more time exploring that. I have a volume control on the output of the distortion circuit because there is also a gain control and I will need to make up some volume with lower gain settings of course. I wish to keep the ac voltage below the clipping level of the TL072, by capping the output by adding a series resistor before the volume control as a voltage divider. (I generally prefer my pedals to be set for unity gain, I don't make use of pushing the all's front end. Distortion circuits seem to have tons of gain and my volume controls are almost always well under 12 o'clock.  But I don't wish to limit the output volume to unity, only to limit it to the clipping threshold. A little higher than unity I hope.)
Make sense so far?
Biased at 4.5v, a rail will clip at a peak to peak signal of,what, 3 or 4 volts?  I need to choose a value for the voltage divider resistor just before the volume pot. I can use a large variable resistor to find this value. I'm thinking I can simply set the dirt circuit for something just a little beyond unity gain, by turning the volume pot all the way up and adjusting the variable resistor.  I can bang on the open strings (with my loudest humbuckers,maybe a compressor with a wee bit of gain) and tweak the variable resistor up to the audible clipping threshold and then back it down a bit. Now my volume control, at maximum, will not allow the distortion pedal to clip the EQ. And hopefully this will be a bit below unity and I can have my volume control dialed back a little, so I'm close to unity.
Still with me? Thanks.
When I do this test I should also set my eq to a narrow-Q boost. The maximum boost hopefully, which is about 10dB or so, somewhat like a fixed-wah kind of thing.  Perhaps this is the first place that will clip - the input to the output op amp, after the gyrator stage. 
I think it's pretty easy to hear when your nice smooth distortion is clipping an op amp downstream. It's a bit of a fart, leaps out at you. Even though I'm only supplying 9v, it seems like i should have sufficient headroom. What would you expect here?
I also have a tone generator and a scope that I could do this with. When I've heard about a peak output of a humbucker as being 1v, is that 1 volt p-p?  What should I use-1.5 v p-p?  That boost of 10 dB from the eq - that will take 1 Vp-p up to 3.2Vp-p, right?  Looks like I'm close to having a problem. I really don't want to have to rebuild something. I think PRR may have once told me that rail to rail op amps aren't the cure-all we imagine them to be, compared to upping the supply rails. I'm up for experimenting. I could try a R to R op amp, or maybe use a TLC1044 to supply the EQ with 12 volts, if I really really have to. (Don't want to. I am sooo done with this pedal.)
Parametric Eq, 9volt, set to a narrow Q boost, after a distortion. Seems reasonable. Anybody been down this road? Maybe I could have reduced all the above to just that.

[PRR]

> ...my lengthy attempt... ... ... ... ... ... that.

TL;DR.

If it's clipping, turn it down. Basic audio hygiene.

If you want the front to clip (fuzz), but not the back, put a pot there and turn it down.

[anotherjim]

Both the 368 and 4049 will run with less than 9v. Use a 78L05 regulator to run them on 5V (or a 317 adjustable to find an optimum) and when they will clip it will be a smaller p-p for downstream headroom. Might be a simpler/smaller way than having to fit trimmers.

[Groovenut]

What if you set your max output from the overdrive circuit to well below the clipping of the eq, using an overly high input signal (1.5VPP) then adjust the overal gain of the eq stage to make up the difference so the output of the pedal is slightly above unity even at low overdrive levels?

[garcho]

"basic audio hygiene" ha! i'm gonna steal that

[blackieNYC]

PRR's eyes glazed glazed over halfway thru - I don't blame you! I do have an output volume pot on the distortion. Pre eq. I want to find a solid way to limit the 5 o'clock position of that volume pot to just below clipping of the following eq. A large resistor as voltage divider before the pot.
Fairly good hygiene for a guitar player.
Groovenut and Anotherjim - thanks for playing. 5v supply sounds good, and I hadn't though of make up gain at the eq output. I think I can simply up the NFB 2.7k resistor (in the geofex Simple Easy Parametric.)  but first, I need to reassemble the thing. It's perfed, but disconnected from the breadboard because now I'm wire wrangling it to a cardboard mock up of a front panel. All the above mods are doable. Not too invasive. But first, I need to see if I can currently get unity with all possible eq settings, and maybe try a r-r op amp cuz that's easy. Think 1.5 Vp-p is plenty? But not too much? I'll do that, and do it by ear as well. Thanks, all three of you.

[Bill Mountain]

I built a 4049 based overdrive.  I was placing an active EQ after the drive section.  I was also worried about clipping so I went the "klon" route and simply increased the voltage for the EQ stage.

I also had a stand alone graphic EQ pedal that I run at 24 Volts to avoid clipping as well.

I like the suggestion to limit max voltage before the EQ.

You could put some limiting diodes before the EQ after the distortion's level control.  If the EQ has a max boost of say 15dB (voltage gain of 5.6V) then I would use series antiparallel pairs of Si diodes to get about 1.2V max.  Times that by 5.6V and you get 6.72V before clipping.  Use a rail to rail opamp for the EQ and you're golden.  These are some huge voltages for the max signal that you'll likely never need in a guitar application.

Or be lazy and use a bipolar +/-9V supply (not on the 4049 though).

[garcho]

I like the suggestion to limit max voltage before the EQ.

i agree, no point in having tons of headroom clipped off for a dirty guitar tone, go the 5V route

Or be lazy and use a bipolar +/-9V supply

lazy?!    too bad guitar lives on a different planet from the rest of audio, i would prefer bipolar

[ashcat_lt]

So...  You know the maximum output level of the dirt box.  You know the maximum gain of your EQ and its clipping level...  Divide? 

Or you don't know the maximum levels and asking how to find out?  If you connect the input of the opamp stage to one of the rails, its output will tell you how close the thing can get to that rail.

[PRR]

4049 and '386 can swing 4V peak.

TL072 can swing 3.2V peak.

If the '072 is unity-gain, then you just need a 4/3.2 divider.

Output level is 3.2V peak. But what does it go into? Many guitar-amps overload at 0.5V peak input. And if you are already distorting, you perhaps do not want to clobber the amp's input. Making heroic chips and rails pointless.

If the '072 can have a 20dB bump, then you just need a 4/0.32 divider.

Output level is 0.32V peak. Which is a happy zone for many guitar amps.

Lose gain?? Horrors!! Except: 4049 has gain of 50-300, '386 gain of 20 to 200. You have GOBS of gain already. Not for "gain", but to get your weak guitar signal up to the 4V clipping points so the chips will distort. (As said, you could use a lower rail on the clipper. But you can't go a lot lower, maybe 3V or 4V B+ so 1V-2V peak.)

Unity gain?? What is "unity gain" on a clipper?

Unity gain implies a linear amplifier. Say gain of 10, no supply limit.

Input - Output
10mV  - 100mV
20mV  - 200mV
100mV - 1V
200mV - 2V
1V ---- 10V
2V ---- 20V

If you applied a loss of 10:1 after this, the output would be equal to the input: "unity gain".

But a *clipper* is linear only up to some (low) input. Say gain of 200 and 4V clip level.

Input - Output
10mV  - 2V
20mV  - 4V
100mV - 4V
200mV - 4V
1V ---- 4V
2V ---- 4V

For any input over 20mV, the output IS 4V.

Or if you apply a 10:1 loss after, the output is 0.4V:

Input - Output
10mV  - 200mV
20mV  - 400mV
100mV - 400mV
200mV - 400mV
1V ---- 400mV
2V ---- 400mV

Yes, at *400mV* input, the output is 400mV, "unity gain". But at ANY other input (I assume you play with some dynamic), the output is not equal to input.

If you play with very small dynamic, you could adjust your final clip-level to be "similar" to your "average not-distorted" playing level. Playing alone, I'm not sure that's really a thing. (Playing with most drummers or horn-sections, you play to equal them, and the level may be more consistent.)

[Groovenut]

Regarding my use of unity gain in my comment ( I realize I am using the term incorrectly to the techinical definition),

Unity gain in this context would be what your ears percieve as the same level out as in using the bypass as an example.

[blackieNYC]

Right. I too am speaking of fuzz pedal "unity" in the guitar pedal sense - jumping from cleanguitar part to dist. guitar part.
The idea being that a pair of doubled Si diodes, clipping the signal abruptly, sounds better than clipping the op amp? Gotcha.  Forgot about that option. Seen that protection method used in non-distortion circuits, havent tried. Certainly, my distortion circuit followed with 1.2 volt diode/ground clippers, isn't going sound very different. Thanks for the ideas guys.
ashcat - cool trick. The datasheets are always for bipolar - so this would give you a clear indication of clipping threshold for any supply you provide?
Paul - thanks for coming back! It's those gobs of gain that cause me to have my volume pot barely cracked open. Set like that, I haven't worried about clipping the following delays or phasers - it's just that eq with the peak I like so much. Anyway, I now have a variety of means by which to attempt put a ceiling on the output volume. I probably will end up with your 4/3.2 divider.
I plan to sell the pedal,  and I suppose most will expect Boss like performance, where any knob setting combinations must yield a workable sound. I don't want to have to tell someone not to turn the volume past 10:30. 

[ashcat_lt]

ashcat - cool trick. The datasheets are always for bipolar - so this would give you a clear indication of clipping threshold for any supply you provide?

Umm... Yes?  If you ask it to go all the way to the rail, it will try as hard as it can, and the output will tell you how close it actually comes.  Don't think that really has anything to do with what that rail voltage is.  You might have the idea to try putting a DC voltage in to see how much gain you're getting, but that won't work (in single-supply circuits at least) because we make damn sure that there is no gain (gain = 1) at DC, and especially in a filter stage, the total gain depends on the frequency content of the input.

The datasheet for the TL07x family says min supply is 5V.  I've never actually tried it with less.  The lowest supply spec I could find in a common package was the LM324 which will run at 3V.  That one has an asymmetrical limit, though, going pretty much all the way to the bottom rail, but only about a diode drop from the top.

[PRR]

> datasheet for the TL07x family says min supply is 5V.

There is an internal 7V reference to bias things up. Performance is down just a hair at 5V, and 5V is super-popular, so they spec that point. 4.5V will work also, but gain is falling and headroom gets silly small.

> LM324 which will run at 3V

'324 was specifically designed for small (or large) single-supply applications of un-critical performance. They had hoped to get the car-engine/instrument market, but also many other minor chores in all industries. They do say at 25C it can "work" at 2.3VDC.

Ah: found the gut-shot:  http://www.onsemi.com/pub_link/Collateral/LM324-D.PDF

Page 5. The bias scheme takes the difference of two Vbe junctions. It will be-alive with Q24 Base at 1.2V. You need another volt or so to wake-up Q22. This and the >1.2V drop from V+ to OUT set the useful lower limit.

But we are audio geeks. We should be able to take a signal at 1V or 50V or 0.010V, pad it down or amplify it up to what-EVER level we actually want. While starving and over-feeding amplifiers is often useful, it isn't always necessary or best.

[anotherjim]

Just to add re the 324/358 specs.
The datasheet value for +rail swing limit is that it's always 1.5v below V+. That's what I've seen when I've measured them too.
If you want maximum output  before clipping you need a Vref lower than 1/2Vcc to bias to the middle of its swing. Some designs I've seen do that with 2 diode drops from +supply to the usual resistor divider +end. 2 diodes isn't exactly 1.5v, but it's near enough and is constant whatever the supply volts is.
Forgot to mention these amp's input range is also Vcc-1.5v.