MXR 10 band clipping circuit

[Max9999]

Hi stompboxers,

I would like to know the clipping leds light up point of schematic below. D1, D2 and D3 are BAS40 general purpose Schottky. Forward voltage 380-500mV.
So as I see it ( and I am usually wrong) clipping voltage is that + whatever IC4B is doing. I am expecting some kind of comparator action of IC4B but I cant find how its connected here anywhere else. Thanks in advance!

[Rob Strand]

I would like to know the clipping leds light up point of schematic below. D1, D2 and D3 are BAS40 general purpose Schottky. Forward voltage 380-500mV.

The circuit is a little difficult to understand because the diodes are always biased on via the 1M to +V.   The idea of use of using two diodes (+2k2) is to detect if either IC4C or IC2C clip.   The 1M + 100n seems to delay the detection unless it clips long enough.

Once you get to D3 it looks sort of normal.

So as I see it ( and I am usually wrong) clipping voltage is that + whatever IC4B is doing. I am expecting some kind of comparator action of IC4B but I cant find how its connected here anywhere else. Thanks in advance!

IC4B is acting as the comparator.   IC4B is wired as a Schmitt trigger.  So instead of the threshold being defined by a comparator it is defined by the Schmitt trigger thresholds.   The Schmitt trigger hysteresis prevents the comparator from glitching.

This page analyses the circuit,
https://www.electronics-tutorial.net/analog-integrated-circuits/schmitt-trigger/inverting-schmitt-trigger/

The addition of the 1M feedback resistor is another complication.   I suspect it extends the time the LEDs are on once clipping has been detected (but I didn't sit down and analyse the whole process.)

[Max9999]

Thanks a lot Rob Strand! The fact that the lights start to blink when clipping is detected does look like the Schmitt trigger output in that link.

I tried to LT Spice it to try to bypass the problem that I dont understand it. The roadblock I hit is that I get the blocky Schmitt trigger output even when I feed it with 0.1V. That surely must be too low to clip this thing .. could you give a quick glance to see if you notice where I messed up? Or should I not measure V but I?

[Rob Strand]

I tried to LT Spice it to try to bypass the problem that I dont understand it. The roadblock I hit is that I get the blocky Schmitt trigger output even when I feed it with 0.1V. That surely must be too low to clip this thing .. could you give a quick glance to see if you notice where I messed up? Or should I not measure V but I?

What you have looks OK.      Looking at V is fine.  You can try setting the input to different levels but I'm pretty sure it won't change much because there's a small bug somewhere.

If you look at the circuit with D3 removed it forms a one Op amp Schmitt trigger oscillator.    Without analysing the details I just assumed the presence of D3 and a zero input signal would kill the oscillation.   It looks like that's not the case.

I need to work out the precise reason for that.   You might need to do a mod to get it to work reliably.  I can see if you pull D1 and D2 then the 1M to Vcc should stop oscillation.  That's only an observation, not a fix.

As a side note, the intent of the D1 and D2 is to compensate for the voltage drop on D3.    This is a common trick.  For some reason the way the circuit is drawn didn't let me see that yesterday.


OK I worked it out just before I pressed post.   The input needs to be wired back to Vref (4.5V), not ground.   The output of opamps float at Vref.

[Rob Strand]

Here's a sim using a square-wave input.   The peak level is stepped.   The clip detector kick in at 3.25V, so the threshold is between 3V and 3.25V peak.   (The LED actually flashes at a more or less constant rate.)

[note: R41 = 470k instead of the R41=620k on the schematic]

[Max9999]

That did the trick! Thank you so much! I hope you have the same victory feeling that I am having by proxy, you deserve it you genius!

With a 9V supply the treshold is about 6V p-p, 1.5V from the rails, which is a little tight for this opamp. But with 18v ( the intended voltage for this pedal) it lets us do 12V p-p and has 3V headroom to the rails, which is perfect.

I can adjust the trigger voltage by varying the R41 R42 ratio right? VT=R2/(R1+R2 ) Vout?

[Rob Strand]

I can adjust the trigger voltage by varying the R41 R42 ratio right? VT=R2/(R1+R2 ) Vout?

Yes, threshold increases when R41 decreases.

If you want to do it precisely in spice you need to make sure the opamp swing matches the actual opamp.   If you don't then the sim will use a higher (ie. rail to rail) Vout and you will see a higher threshold than the real circuit.

The sim will show you the difference between the actual behaviour and the theoretical number.  The actual thresholds are larger because of the lag from the detection time constants

FYI, my sim had R41 = 470k instead of the R41=620k on the schematic.

[Rob Strand]

I was about to update my sim to show how the circuit trips at different peak voltages with square-wave and sine-wave input but then I realized something wrong.  With zero (AC) input the clip-detector output voltage is sitting at 0V, that will turn on the LEDs.   That seems to be a problem.  That got me thinking does the circuit re-arm correctly when the clipping stops?

You often start these threads assuming things are mostly OK but then you realized there's a few cracks and you need to spend time on it.

(FYI, another detail is the opamps are biased slightly above Vcc/2, since R4 is less than R2.   That's going to change the clip voltages and the trip threshold.)

Here's a sim showing the Schmitt trigger always resets to output zero regardless what state the Schmitt-trigger is in when the signal is remove.

So the gross problem is the clip LEDs are on when there is no signal.



I had a look at the pedal and I think the "reversed behaviour" of the clip LEDs is intentional.   

The 10 x band-leds, which are always on, are part of the slider pots and are used as position markers.   The two clip LEDs are the position marks for the Volume and Gain slider pots.   So in order for the all the slider LEDs to be on in the normal state they decided to blank the Volume and Gain indicators for clipping (instead of the usual ON = clip type clip indicators).     The manual says very little.