Mutron 3 type filter with H11F3 problem

[grapefruit]

I've built a SVF similar to the Mutron 3 filter. It works fine using a dual gang pot to control the frequency, but when I use a H11F3 fet output optocoupler for frequency control it is distorted, especially at high resonance settings. It's not clipping. It looks more like slew rate distortion...

If I put a 400Hz sqaure wave in, and turn the frequency control up...

With dual gang pot:
Starts of with pretty much  sine wave and smooth humps (harmonics) are added as I raise the frequency.

With the H11F3:
Starts off a bit more triangular, and as the humps come in they're not smooth, a bit more triangular - straight line-ish. And you can hear the distorion in the output just as the humps come in.

I have a 2k2 resistor in series with the H11F3, and a 100k resistor in parallel.
I hope that made sense! Any ideas would be greatly appreciated...

Stew.

[free electron]

H11F3 opto-fets works perfectly as switches, when the LED is fully on or off, but used as a variable resistance they can work only with very small amplitudes, even guitar signal can be to big. Try o put a sine wave into the h11f3, thru some resistance (2k2 for example) and observe the output when changing the led current. Once I bought about 100pcs of H11F3s in hope to use them as variable resistance, as datasheet says. I found them no use in such application, at least with reasonable signal amplitudes. But, they work very nice as opto-isolated switches. I use them a lot in anti-pop automatic muters controlled directly from digital logic.

Regards
Piotr

[grapefruit]

Thanks Piotr,

I was starting to think that may be the case. I'll try something else...

Regards,
Stewart.

[Paul Perry (Frostwave)]

This excellent article www.pisotones.com/MuTron-III/MuTronIII.htm (in Spanish) says the H11F3 saturates too easily. Personally, looking at the H11F3 data sheet, I think it should be possible to overcome the problem, maybe the peak current to the diode is just too high?
If it is a matter of severe non-linearity caused by the audio thru the fet of the H11F3 beign large, maybe there is some way to reduce the audio signal thru it & make up gain elsewhere?

[grapefruit]

I don't think it's a problem with the LED current. I'm controlling it manually and even with low LED currents it has the problem. I try reducing the input signal more and see how that goes...

The data sheet says it handles high level signals, but in this case it doesn't seem to.

Stew.

[markphaser]

Sounds like its overloading the H11F3 fet output optocoupler but why or how?

[markphaser]

here is the schematic

http://experimentalistsanonymous.com/diy/Schematics/Filters%20Wahs%20and%20VCFs/Mu-tron%20III.jpg

What is that 0805 component is it a FET or optocoupler?

where is the H11F3 fet output optocoupler at on the schematic?

[markphaser]

On the schematic op-amp (A5) has 2 diodes what do those diodes do?

What kind of stage is that? is it a peak circuit?

[grapefruit]

In the Mutron schematic, the 0805 is an LED/LDR type optocoupler. The circuit around A5 is a precision full wave rectifier, used in conjuntion with C9 to detect the envelope of the signal.

I'm not building a Mutron. The filter circuit I'm using is similar to the Mutron filter circuit though.

Stew.

[markphaser]

Thanks for the information

A5 is a precision full wave rectifier

So what does the full wave rectifier do ? it changes your guitar audio into a rectifier square wave?

Is A6 a driver for the 0805 is an LED/LDR type optocoupler?

[grapefruit]

After a quick search, all you want to know is here...
http://www.geofex.com/PCB_layouts/Layouts/neutronpub.pdf
RG has done it again!

[oli]

Hi,

the rectified signal is used to drive the LED of the optocoupler, but this rectiefied signal does not appear at the output of the effect.
And yes A6 is a switchable (inverted or not inverted) driver for the LED - but also just a simple OP amp.
RG´s layout is good, but it has one disadvantage - it needs two batteries or a charge pump, but this circuit works with just a little modifications with a single supply.
Regards,
Oli

[markphaser]

Thanks

Its a Precision Rectifier/filter.The rectified signal is filtered into a variable DC Voltage in C9 4.7 envelope cap to the LED driver
driving a proportional current throught the LED the LDR changes the filters frequencys

1.) H11F3 LED/photo FET
2.) CLM6000
3.) NSL32
4.) VTL5C3

Whats the difference between all these? they are replacements but are doing to have a different attack times and sweep/tapers i would think than the 0805

The orginal 0805 had 2 LDR's and one LED how does it make it different having 2 LDR's in one package with a LED than to have one LDR in a package with one LED? I'm guessing that the 2 LDR's would be matched and ganged together more in sync?

[MR COFFEE]

You can use the optoJFET as a sub for the photocell modules if you want to engineer a voltage-controlled PWM scheme and filter the switching frequency back out.

Parallel the fets with a high resistance and modulate the pulse width to get a varying resistance, i.e., if the JFET is in the circuit 10% of the time, the eqivalent resistance is the channel resistance (or resistor in series with the fet channel) times 1/0.10. If it's on 90% of the time, it's the channel resistance times 1/0.90.

This scheme is used in very high end compressors like the Crane, but takes good engineering chops to make it work well. Minimum switching frequency is greater than 150 khz and ground plane construction is highly recommended, shall we say? 

[Helmut]

I have the exact same problem with my circuit which uses the General Guitar Gadgets Neutron PCB.
http://www.generalguitargadgets.com/content/view/112/26/
I also chose the H11F3, and it distorts like crazy in certain settings (especially higher peak).

My assumption was that the R-on resistance is too low (100ohms) compared the original (don't know value).
I am contemplating inserting inline resistors, but I'd prefer not to cut the traces on the board.

Reading this blog though, it seems like the H11F3 cannot handle larger signals at all.
Per everything else that I have read, the H11F3 should work fine, so I'm a little lost.

Any ideas - what fixed the problem of the original post ?
Thanks!

[R.G.]

Time for general principles again.

The H11F3 is based on an LED illuminating a photo-FET.

Photo-FETs are J-FETs, they are just driven on or off by the presence or absence of light.

JFETs have a "triode region" near 0V/0A where they act like a variable resistor.

But there are limits to the range of voltage and currents over which acting like a resistor is true. Exceed those limits even a little, and the FET no longer acts like a resistor, it acts like - well, something different than a resistor. We hear this nonlinearity as distortion. See http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/info/comp/active/jfet/jfetchar/jfetchar.htm for an illustration.

Guitar signals from a single coil pickup are a tight fit into the resistor region, and humbuckers don't fit at all.

Using the H11F3 as a switch sidesteps the resistor region and uses only the cut off and full-on regions.

[Helmut]

Okay, I guess what I didn't understand is why this causes distortion, because the photo-FET is only used to vary the filters' cutoff frequencies.
My understanding was that the resistance goes to saturation when the controlling signal is out of range.
But what you're saying is that the FET side of the IC is non-linear even if the controlling signal (amount of light) is constant, correct ?

Thanks for the information!

[R.G.]

Okay, I guess what I didn't understand is why this causes distortion, because the photo-FET is only used to vary the filters' cutoff frequencies.

"Only used to vary the filter's cutoff frequencies" is true. However, this does NOT mean that the photo-FET, acting like a resistor, does not have signal across it. In fact it MUST have signal across it to vary the filter frequency in this and similar circuits.

Hence, nonlinearity in the photo-FET's resistance distorts the signal that does go across it.

My understanding was that the resistance goes to saturation when the controlling signal is out of range.

This it true too. They are not contradictory.

But what you're saying is that the FET side of the IC is non-linear even if the controlling signal (amount of light) is constant, correct ?

All things are nonlinear if you put a big enough signal across them. All things are close enough to linear for our purposes if the signal is small enough. What I'm saying is that the FET, biased right into the middle of its resistor range, is linear enough for small signals, but the nonlinearity shows up if the signal across it is big enough. This is independent of the amount of light. If the amount of light is zero, the FET is very linear - but an almost infinite resistance, so that doesn't help. Even in the resistive region, if the signal is big enough, the resistance is not constant.

As an example of what I said earlier about everything being nonlinear if you put a big enough signal across it, let's consider that paragon of musical virtue, the carbon composition resistor. This resistor has a voltage coefficient of resistance which is larger than carbon film, metal film, wirewound, practically all other kinds. The resistor changes with the voltage across it. Not a lot, but a few percent for a couple of hundred volts change in the voltage across it. So big signal swings across CC resistors are distorted slightly. This happens to be a sweet-sounding distortion, and it has become enshrined in legend and advertising.

At the other end of things, we commonly refer to diodes as clipping. However, to a small enough signal, they are not off/on devices, but exhibit a smooth change in resistance from nearly infinity to a few ohms. This all happens in the 0.1-0.2V "knee" of the diode. As an example, the tremolo in JMI Vox Supreme amplifiers and Thomas Organ Vox amps is based on varying the current through diodes with an LFO to make them change their resistance. The signal must be below 25mV, preferably below 10mV for this to work without apparent distortion, though.