Possible mod to Maestro FSH - idea

[jrc4558]

I was playing with the sample-hold function today at the rehearsal. It was fun, but too unpredictable at times. So I decided to think of a way to make it more consistent, and the first idea was that the filter section can be driven by an oscillator, instead of peak detector, or a S&H. But, instead of using a triangular function, the filter can be driven by an LFO that would produce a step-like function. Can anyone please tell me what would be the proper name for such function? English is not my first language, and I don't know the particular term for it. And i need that term to search such oscillator by...
thank you.

[Seljer]

I've been thinking about the same thing.

My best idea (as it'd still be implemented into the basic structure of the pedal) is instead of the random noise being used as the input for the S&H you'd place an LFO of any kind there, and depending on its speed and if it syncs with the S&H's clock or not (particulary at higher LFO speeds) you could get some pretty neat stepped waveforms.

Or you could add an actual LFO that generates stepped waveforms -> http://geofex.com/Article_Folders/LFOs/psuedorandom.htm
or a sequencer (like in the Zvex Seek Wah), also described on there but heres another one: http://www.geocities.com/thetonegod/vanpoint21/van_point21.html
and another ones http://www.dedalofx.com/bioroids/ideas/idea_02.php (this one here is specifically desinged for OTA based effects like the FSH...)

[Paul Perry (Frostwave)]

Seljer is on the right track.
If I was doing it, I'd be thinking about having a control that panned between the noise signal that usually goes tot eh sample and hold, and a LFO.
Then, as you vary the mix between noise and LFO, the output of the sample & hold varies from totally random to a stepped function. The LFO has to be slower than the sample and hold frequency, of course.

[jrc4558]

I liked the 16-Steps Triangle LFO on dedalofx site. I will try to breadboard his design of the stepped LFO and tie it to the 22k resistors thet send CV to pins 1 and 16 on the LM13600. I mean this one: http://www.dedalofx.com/bioroids/ideas/i02_lfo.jpg
Any know layouts to this perhaps? If no - I could make one, I guess!

[Mark Hammer]

Ironically, when it comes to synths, one of the "tests" of how good a noise source is, is whether feeding that noise source to a sample & hold for control voltage purposes sounds too much like a stairstepped LFO!

[moosapotamus]

I recall being able to get some repeated (non-random-like) patterns from the stock FSH-1A circuit by various different settings of the trim pots. In fact, when I built mine I had considered using multi-turn trim pots because I actually had a hard time setting them to get a good random-like sound.

~ Charlie

[jrc4558]

Ironically, when it comes to synths, one of the "tests" of how good a noise source is, is whether feeding that noise source to a sample & hold for control voltage purposes sounds too much like a stairstepped LFO!

Then, I could build a triangle LFO, feed it to the 0.05uF capacitor on the noise input of S&H and get a stairstepped LFO on the output? Then the LFO speed will be the speed of modulation, while the S&H speed will vary the number of steps per cycle, correct?

[Paul Perry (Frostwave)]

I recall being able to get some repeated (non-random-like) patterns from the stock FSH-1A circuit by various different settings of the trim pots. In fact, when I built mine I had considered using multi-turn trim pots because I actually had a hard time setting them to get a good random-like sound.~ Charlie

That shouldn't really happen. Maybe there was hum gettig into the "random noise" before it was sampled?

[moosapotamus]

I recall being able to get some repeated (non-random-like) patterns from the stock FSH-1A circuit by various different settings of the trim pots. In fact, when I built mine I had considered using multi-turn trim pots because I actually had a hard time setting them to get a good random-like sound.~ Charlie

That shouldn't really happen. Maybe there was hum getting into the "random noise" before it was sampled?

Hummm... dunno, maybe. I built the GGG version with all of Tonda's mods and used a MAX1044 for the +/- 9V supply. What's the function of those trim pots (R26 & R27), anyway? R26 looks like offset? R27 looks like intensity? It's been a while since I played with it. I might have to pull it out and have some more fun.

~ Charlie

[Eb7+9]

In one of my re-designs I used the LFO from the MS-20 and designed a different noise gen - the Maestro/Oberheim noise circuit doesn't spit out white noise, it's more of a PWM Clock really (same idea used in EMS vco's) ... I then coupled both outputs through a mixing pot to get step plus noise - giving straight staircase, staircase with a random component, or straight random out of the sampler ... I have an original Oberheim VCF, it exhibits some of the problems I've heard in the clone sound files - the originals did vary quite a bit I hear and many people have problems with their clones ... I don't think it's an accident - this can be explained by studying how the voltage to current converter driving the OTA's works ... basically, there's an offset issue that arises from the second jFET in the sampler since that device also operates as a level shifter - everything is DC coupled in that section ...

Its VgsDC bias voltage affects where the whole voltage range driving the converter lies - if VgsDC is too high the filter will never enter the realm of lowF settings and can even lie too high where the filter doesn't really do too much ... there's two ways to fix this: swap that jFET or, better yet, take the Base connection on the current follower PNP device that's connected to ground and hook it up to the sweeper of a 10k or 50k pot instead of ground - with both ends of the pot tied to the rails ... this will allow you to fine tune the vertical OFFSET of the filter ... another thing you can do is replace the fixed resistor going between the Source on the jFET driver and the Emitter on the PNP follower by a larger valued pot (think I used 100k) - the role of that resistor is to provide a SCALING function  - it determines how wide/narrow the frequency range in the filter will be for the amount of voltage range feeding the converter ...

Providing for offset and scaling adjustments and using a better noise source can bring much improved performance from this circuit ... also, if you go for a ramp generator idea you'll find the offset and scaling controls handy for setting exactly where the filter starts and stops in response to swept voltages ...

~jc

[bioroids]

I liked the 16-Steps Triangle LFO on dedalofx site. I will try to breadboard his design of the stepped LFO and tie it to the 22k resistors thet send CV to pins 1 and 16 on the LM13600. I mean this one: http://www.dedalofx.com/bioroids/ideas/i02_lfo.jpg
Any know layouts to this perhaps? If no - I could make one, I guess!

Hey, you have some sound samples of that effect on the site, but applied to a synth. I don't have a pcb layout, sadly. On some cases it has turned out noisy, depending on the cap used for filtering. I guess there's lot of room to improve on that circuit.

Luck!

Miguel

[chemosis]

is it possible to get a bigger fatter sound out of filter. it sounds thin and small at times. also wish to add cvs but lack knowledge

[Kevin Mitchell]

In response to the original post, what I've done on breadboard (if I recall correctly) was replace the NPN noise transistor with a triangle wave. Put an output of a simple LFO onto the 47n cap instead of the noise transistor being the frequency source. The only thing is the notches of the triangle wave may play funny with the s/h speed so a symmetrical, repetitive cycle would be tricky to maintain. So you'd have to find the sweet spot between the two controls when dialing it in.

Here's a post I kept up with earlier this year regarding a newer FSH. I'm throwing one together tonight for a holiday gift.
Mini FSH Thread