Op amp is clipping too early

[Tube2stomp]

Hi,

I'm breadboarding some circuits that have low input impedance, so instead of using a pedal going in, I added on the board an op amp input stage to get high impedance input and low impedance clean boost output.
I'm using a simple tl072, V+ is 33.4V, half voltage bias the usual two 10k resistors with 100uF bypass, feedback resistance is 100k pot with tail resistor 2k thru 47uF cap to ground, for the tests I have a 500k pot on the output.

Really simple, right?
With such high V+ and just 50x gain I didn't expect any clipping from just a guitar input, but it does clip.
I measured my guitar at max 0.3V and i measured the output of the opamp just before it sound like clipping and it gives 6.6V.
Why is that? shouldn't it give clean boost up to at least 25V if not more?

I don't have much experience with solid state stuff so I'm really not sure what am i missing.

[PRR]

> just 50x gain

That is a LOT of gain in almost any audio path.

Why so much gain??

If the original concern was just impedance, why not unity gain?

And why buffer down to a low-low impedance, then put a 500K pot *after* the gain? That pot can be 125K impedance, losing all the impedance advantage of the opamp. Also (as you know) allowing 50X gain to clip before you come to a place to turn-down.

Change your 100K to 2K or 5K.

> I measured my guitar at max 0.3V and i measured the output of the opamp just before it sound like clipping and it gives 6.6V.
Why is that? shouldn't it give clean boost up to at least 25V if not more?

Measuring true Peak voltage from plucked instruments is tough. Also you may be mixing RMS, Peak, and Peak-to-Peak. Given 33V supply, the most output is 33B P-P, 16.5V Peak, 11V RMS. TL072 will lose about 2V each side, giving 28Vpp 14Vp 10Vrms.

10Vrms out and gain of 51 says 0.2Vrms in will clip. You already said 0.3V in, so clipping is expected. I would pencil a higher voltage for many guitar rigs. I would not take more than gain of 10 here. There are few situations (except gross overdrive) where you need even a gain of 10.

[amptramp]

I have given up on using meters to determine the signal levels.  If you use a scope, you can see peaks that exceed the average or RMS levels you would see on a meter.  Guitars can surprise you with the output levels sometimes.

[Tube2stomp]

I'll explain.
I'm testing to see how tubes behave at low plate voltages. Started from 9v to 12v to 18v and went up to 33.4v so far (power supply of a printer). At those voltages there is grid current so input impedance is low, too low for direct guitar input... Hence why I need high input impedance and low output impedance from this stage.
I do want and need input gain going into the tube, if any just to see who it behaves driven hard. My highest voltage clean boost ever was a Barber Launch Pad (9v in doubled to 18v),  so I was curious how high of clean signal I can get from the same op amp when powered from this 30v+ "HT".

I'm using a Fluke 17B+, it has frequency test capability at least on paper from 10hz to 100khz so guitar frequencies are easy. It also has a min/max feature so plucking my guitar like an idiot way above what I really play showed max result at 0.312v.
I assume it measures AC voltage at peak to peak,  don't think it has RMS capability and measuring mains i get 230v so its not just peak either.

Oh and the 500k pot was just what I had at the output of the whole circuit and used it for measuring just the input op amp. In the circuit it will have to actually drive anything from 10k to I assume 50k.

With that said am I wrong or am I getting clipping (at least to my ears) at 6.6vpp?
Don't have a scope here (looking for one locally) but it could only show that the signal is actually clipped earlier then my ears say...

[Rixen]

can't quite visualize your circuit, do you have a dc path for all the input bias currents ? Hope I'm not telling you how to suck eggs..

[Tube2stomp]

The op amp input stage?
Really run of the mill setup, like the shaka boost circuit.
Using non-inverting input, biased to half HT via two 10k resistors with 100uF bypass.
Have 100k pot in feedback and 2k in tail, thru 47uF to ground.

I'm at work right now but could it be the 500k load at the output? as in the op amp needs a lower load to get higher voltage swing with this setup?
Just guessing even though it doesn't make sense to me...

* if it wasn't clear the tests on the op amp are done detached from anything "tubes". just the op amp by it self.

[PRR]

Bench meters without RMS are often Average Reading, RMS calibrated (on the assumption the wave is sine).

The "10hz to 100khz" is the response into the detector. The detector is followed by filtering so the needle does not jiggle so bad (equivalent for digital display: changing too fast to see).

So with 33V supply, 28V peak to peak, a sine clips at 10Vrms. With steady sine, the meter will agree. With plucked sounds, it will be similar but not exact. 6.6V'rms' may be "correct" for lightly clipped guitar.

If you are running the tubes on 33V, there is NO need to run the grid buffers on more than that. If the tube has gain of even 10, then same-supply driver can hit 10X OVER-drive, pretty gnarly.

[Tube2stomp]

When I measured my guitar I measured plucking (and getting MAX 0.312V) but when I measured the opamp stage I fed it with a 1kHz sine...
Knowing 0.3V will be the verge of clipping, I first fed in 0.2V (measured at opamp input) and got out 4.4V.
Then tried 0.3V and got 6.6V so amplification was on par and still no clipping.

I stated the "10hz to 100khz" just to be clear my DMM is measuring AC voltage on all frequencies and not just up to 200Hz or 400hz like cheap DMMs.
Meaning the guitar plucking signal of 0.312V was a true measurement.

So now, knowing I measured 6.6V with "steady sine", how's that is sitting with 10Vrms/28Vpp? 

p.s. Oh and yea I know that at 33V even as is that opamp stage has way more drive then tubes can handle. I'm just trying to understand what I'm actually measuring here with the 6.6V

[PRR]

You may need an oscilloscope to know what's really coming out of the preamp.

[Rob Strand]

Have you measured the DC voltage on the output of the opamp?

You should put a 22pF cap across the 100k feedback pot (10pF to 47pF OK).

You do have input and output caps, right?

[DrAlx]

Have you measured the DC voltage on the output of the opamp?

+1.
With no scope that would be something worth checking. One way you could get a lower RMS level is if the square wave deviates from 50:50 duty cycle. E.g. if if bias point is far from the "middle".  Easy experiment is to vary the bias point with a trimmer and see if that has an effect.

[Tube2stomp]

Just want to update:

I measured exactly half V+ at output and the same at inverting input... but non inverting input was 0.8V less
I have no idea how could this be cause the output and inverting input should get their operating point from the non inverting input, right?

So out of frustration I pulled out everything from the breadboard, wiped out the whole project tubes and all and started from scratch, this time from the op amp input stage.
It's all OK now
The clean headroom with 33V power supply is great, At the moment I tested up to x75 amplification and dinging into the strings produce no clipping.
It's 2AM here and in 4.5 hours I need to wake up  (well I have to fall asleep first!) and go to work so measurements will wait for tomorrow.

Thanks for the help guys! zzz

[highwater]

I measured exactly half V+ at output and the same at inverting input... but non inverting input was 0.8V less
I have no idea how could this be cause the output and inverting input should get their operating point from the non inverting input, right?

You are correct. When the meter's input impedance loads-down the non-inverting input, the inverting input and output also drop by the same amount... you just can't see it without a second meter, because it goes back up as soon as you remove the probe from +in.

So out of frustration I pulled out everything from the breadboard, wiped out the whole project tubes and all and started from scratch, this time from the op amp input stage.
It's all OK now

Glad to hear you got it working.
Unfortunately, for simple circuits with not-so-simple mistakes, rebuilding is often easier (or at-least less frustrating) than debugging.

[amptramp]

Yet another instance where the breadboard was the source of the problem.  I have used them in the past and I don't recommend them.

[Tube2stomp]

You are correct. When the meter's input impedance loads-down the non-inverting input, the inverting input and output also drop by the same amount... you just can't see it without a second meter, because it goes back up as soon as you remove the probe from +in.

Yea I'm aware of meter impedance loading and it does seem like that's what happening... but i'm not sure...
See, the bias divider ladder is of low impedance (10k + 10k). To load and shift the bias point by almost 1V I need to have roughly 100k parallel to the 10k going to ground.
Since I have a 470k input/bias resistor and my meter has 10M DCV input impedance means when measuring the non inverting input bias at the op amp pin I put 10.470M in parallel to 10k resistor of the divider going to ground. unless I'm missing something.

I can say that after starting from scratch and the op amp stage performed as expected, I measured same DC bias at all three pins (inverting and no inverting inputs and the output).

Another story for the (breadboard) books, ey?