Thanks, Ill try out the first one as I already have all the parts, though the lf398 seems to be quite reasonably priced too.
I just switched around the components.
AndâŚused a 10nf instead of a 1nf. Both changes where necessary for my circuit to work at the time. I havenât tried to build another one.
I definitely want to build a couple more sample and hold circuits. I will likely put the Schmittz circuit on breadboard this evening with some noise circuits. I intend to build YASH but I need to get a few supplies for that one.
This should work but I dont know why the original arrangement didnât work as well, what sort of waveform were you using to trigger it?
1 Like
Ok I just ran a couple of simulations of both versions of the trigger ciruit the trigger was a 5 volt peak to peak squarewave. It seems that the original circuit didnt produce a significantly large pulse to flip the cmparator, Im guessing Rene must have been using a 10 volt pk to pk waveform.
Rene? Thought we were talking about Moritz Kleinâs. Their comparator is 100k over 33k, so 3 V should trigger it (with a 12 V supply). I was doing a simulation too and it said with a 0 to 5 V (or -5 V to +5 V) square wave input the pulse height at the comparator input goes up to about 4.4 V, so that should be fine. If the diode and cap are switched around the pulse height goes a little lower but still above 4 V, and itâs about 2x wider; there also are negative going pulses on the falling edges of the square wave which is what the diode after the cap is intended to prevent. However that doesnât disturb the comparator.
With a 10 nF cap the trigger pulse is 10x wider and lower in amplitude â in my simulation, too low to trigger the comparator with a 0 to 5 V square wave input. So itâs odd that it worked better for Aaron. The width shouldnât matter as far as triggering as long as itâs large compared to the integration cap charging time, and with a J113 JFET thatâs quite short.
With a 10 V square wave it does trigger. The integration time is 10x longer, about 1 ms, which is a lot longer than it needs to be but probably still short enough not to be objectionable.
I can get it too trigger with square or triangle. The LFO I originally tested it on has square and triangle. The sqr is like right at 5v+/-, and the triangle is right at 7+/- on that particular circuit. I have been using a 4017 clock divider to trigger it lately.
You are right, I was talking about Moritz Kleins circuit. Doh 
Regarding my simulation results, the simulator is s not very consistent with simulating these sort of circuits, Ive tried out several and the end result is very dependant on the simulation speed. I assumed that because my result backed up Dogstars finding that my simulation was accurate in this case.
What surprised me was that Dogstars circuit was able to be triggered by a triangle wave!
1 Like
Can you remember how fast the LFO was running when you used it to trigger the S&H circuit using a triangle wave? Im a bit perplexed by this, hence the need for more info.
Blue noise, thats new to me. I need to try this out, Im guessing 3dB/Oct High pass, like a mirror image of pink noise filtering?
1 Like
Yeah itâs something like that. 3 approximated 3db/oct slopes put in parallelâŚor something⌠
Its been a while since I tested my sandh with a triangle trigger, and my memory says one thing, but reality says different. I just tried the same triangle wave that I ârememberâ getting it to trigger. Seems my memory was part fantasy. I canât get a triangle to trigger it.
Youâre supposed to get blue noise by filtering white noise with a 12dB highpass filter at 6kHz. There is also orange noise, yellow noise, etc, in the same analogy of colour spectrum.
The circuit shown above is not a 12dB per octave HPF. It appears to be a similair slope as the pink noise filter but sloping upwards rather than downwards. Perhaps ir was misnamed by the original inventor. Perhaps it should be called a pastel blue filter. 
My guess is you probably used a sawtooth (fast rise, slow fall)
My powers of recollection are a bit shaky these days as well.
1 Like
Per Wikipedia:
Blue noiseâs power density increases
3.01 dB per octave with increasing frequency (density proportional to f ) over a finite frequency range.[5] In computer graphics, the term âblue noiseâ is sometimes used more loosely as any noise with minimal low frequency components and no concentrated spikes in energy.
[Note this is 3 dB/oct power density, not amplitude.]
and also
Some of those names have standard definitions in certain disciplines, while others are very informal and poorly defined.
I think if the discipline involved is synth DIY, itâs likely to be informal.
3 Likes
Good points. I was about to say that the analogy to the light spectrum is not set. A high pass filter should aim to get one to the âblueâ territory of noise that we perceive as âhissingâ.
hiiii⌠can u give me solution i couldnât fine a jfet is there some way to us a common mosfet
I also have some logic mosfet lying around id there some way to use those ??
Btw nice project !!!
1 Like
You can find different ways in which you can replace the more hard to find JFETs with more available MOSFETs in a S&H circuit. I canât remember whether it was Ken Stone or someone else who had a schematic using four MOSFETs, and there are others as well. I havenât built one (yet), but the design proposed by Trevor van Ginkel here using two MOSFETs seems simple enough.
Thanks! I wouldnât know where to start without the jfet. I had to order the jfet. Iâm still quite new at this and I got lucky with this circuit. I wouldnât know how else to do it with MOSFETs.