Mini Sample Hold - Transfer Layout

[Marcos - Munky]

I don't have any 3904s here, so I may get some when I go to the parts store. I did tried a few 2089s, but nothing really interesting. Also, setting the trimpots was pretty hard too.

On the other hand, I found out the body is a nice source of noise I had wires going to the breadboard, and holding the emitter wire made the oscilations more random, more pronounced and with a bigger range of tones.

[Kevin Mitchell]

I'm sure most NPNs are candidates here. Sadly I haven't researched what factor make one a good noise transistor.

-KM

[Mark Hammer]

I don't have any 3904s here, so I may get some when I go to the parts store. I did tried a few 2089s, but nothing really interesting. Also, setting the trimpots was pretty hard too.

On the other hand, I found out the body is a nice source of noise I had wires going to the breadboard, and holding the emitter wire made the oscilations more random, more pronounced and with a bigger range of tones.

It would be nice if a bowl of Rice Crispies had an emitter and collector contact. 

[Marcos - Munky]

It would be nice if a bowl of Rice Crispies had an emitter and collector contact. 

That! But seriously, I found pretty interesting tones from touching some parts of the circuit. For example, the jfet near the 3906 on Kevin's layout. This one looks like a good candidate for circuit bending.

I'm sure most NPNs are candidates here. Sadly I haven't researched what factor make one a good noise transistor.

I tried what I had on stock, but I don't have so many options. Today I played a bit more, this time most focused on the guitar itself than the transistor, and with the amp set at a higher volume. And I must say I'm very happy with the BC107B I tested today.

... but I'm not still happy with the resonance. Now I got the transistor selected and the trimpots tuned by ear to the settings I think it sounds best, seems like I'm having too much resonance. I'll put the 1M pot back to the circuit, and maybe try another resistor like a 470K in series with it.

Btw, I found out the bias trimpot have some interesting spots. Like Mark said, only a small part of the range is useful, so it may or may not be worth to use it as a external pot. Fixing it to a smaller range and adding a fine tuning pot may be interesting. On the other hand, I think the depth trimpot was such a pita to be set that it really should be fixed. I also suggest to use a multi turn trimpot, because it's very sensitive.

Also, I'll probably add a led indicator for the speed, as Kevin's 2nd layout.

[PRR]

> It would be nice if a bowl of Rice Crispies had an emitter and collector contact. 

https://laughingsquid.com/cereal-bowl-amplifies-rice-krispies/
http://dominicwilcox.com/portfolio/snap-crack-and-pop-amplifier-cereal-bowl/

[Mark Hammer]

I tried what I had on stock, but I don't have so many options. Today I played a bit more, this time most focused on the guitar itself than the transistor, and with the amp set at a higher volume. And I must say I'm very happy with the BC107B I tested today.

... but I'm not still happy with the resonance. Now I got the transistor selected and the trimpots tuned by ear to the settings I think it sounds best, seems like I'm having too much resonance. I'll put the 1M pot back to the circuit, and maybe try another resistor like a 470K in series with it.

Btw, I found out the bias trimpot have some interesting spots. Like Mark said, only a small part of the range is useful, so it may or may not be worth to use it as a external pot. Fixing it to a smaller range and adding a fine tuning pot may be interesting. On the other hand, I think the depth trimpot was such a pita to be set that it really should be fixed. I also suggest to use a multi turn trimpot, because it's very sensitive.

Also, I'll probably add a led indicator for the speed, as Kevin's 2nd layout.

The usefulness of the Depth trimmer as a control will depend on the noise transistor selected.  Remember that it simply attenuates the amplified random voltage from the noise source.  If the gain of the op-amp isn't enough to make that random voltage big enough to justify attenuation, then the problem lies not with the Depth trimmer but with the choice of transistor and gain of the op-amp.  When I finally found the "right" 3904 as a noise source, I found the Depth control quite useful over at least 60% of its range.  That's not 100%, but it's quite enough to be useful as a panel control.

As for the usefulness of a clock-rate indicator LED, it may be useful for slower clock rates, but I imagine once you get above 4hz or so, the flashing won't be particularly informative.

[Marcos - Munky]

On my build, I think less than 60% of the range was useful. I mean, I could get from "just a little bit of effect" to "more effect than useful". The thing is all of this is in just a little bit of the turning range. So a panel control may be used, but probably the taper have to be a different one. Or maybe adding resistors to set the limit values and use a fine tune depth pot. Just a suggestion if anybody wants to use it as a panel control and have the same issues as me.

On the led, yeah you're right. I was thinking on "tremolo/phaser" clock rates, but thinking again on the clock rates of this one, the result probably'll be just a lit led.

[Mark Hammer]

Tinkering with mine this evening, I swapped the 1meg feedback resistor in the noise stage for 1M5.  Rummaging through the miscellaneous transistor, I found a transistor whose noise qualities I liked but which needed a bit of a boost (an A817).  The 1M5 did the trick.

I traded the 10k depth trimmer for a 5k panel mount pot with a 5k1 fixed resistor in series with the ground leg.  So you can't kill the noise feed completely, but you can trim back the variability of the noise.  I'll figure out a replacement for the Bias trimmer to connect it to a panel-mount control tomorrow.

Experimenting with different resistor values, I found that 2M2-2M7 worked well for the "droop" function.  The 47nf storage cap goes between the JFET gate and V-.  The droop resistor goes between the gate of the second FET and ground,, to bleed off the random voltage stored in the cap.  It provides a very pleasant downward sweep with each step; a sort of "pyoo-pyoo-pyoo" sound.  The price you pay for this is that it limits the highest filter steps, necessitating turning the Depth control up a bit to compensate.

Finally, it can stand to have a little more resonance than it currently does, and does not suffer from having less.  So consider increasing the 330k in series with the 1meg resonance pot to 470k or even 680k.

It's a nice little circuit.  Limited applicability - much like a ring modulator - but lots of fun nonetheless.  Thanks again, Kevin.

[Marcos - Munky]

I didn't tried the "droop" resistor, but one of the sounds I got from touching around the circuit board seems to be similar to what you're describing. So I'll really try it! And also increase the 330K resistor to a bigger value, since I too agree it does not suffer from having less resonance.

I kinda boxed mine today. Kinda because I found out my bias trim is faulty and have to replace it, but I don't have any 100K trim on stock. Had to remove one electrolytic cap and connect it to the board by pieces of wire because I drilled the hole for one of the jacks on the wrong place.

And I also agree it's a fun circuit. Regular chords sounds awesome with it.

[Mark Hammer]

Here's my build.  The audio is a little overdriven.  I should have lowered the guitar amplifier volume or the mic sensitivity, or both.  But it demonstrates the working circuit, verifies the layout (thanks again Kevin!), and demonstrates the "Droop" function.  Enjoy.  More details are provided at the Youtube posting.

[Kevin Mitchell]

Excellent, Mark! That sounds exactly as it should 

I'm on the fence about the depth and bias being mounted controls. But then - I'd likely only play with the speed myself. Very cool nevertheless!

-KM

[Kipper4]

Wow I've never heard a sample hold before Mark. The randomness is cool. Great effect.
Imagine.....

Thanks guys

[Kevin Mitchell]

Here's a no-fill version. I didn't do solder mask for my first one and the bridges are a nightmare. The spacing isn't wide enough to avoid bridges. I'd need a thinner solder tip for sure.

PDF link here


-KM

[rankot]

Try adding more space to resistors and use rounded rectangle shaped pads for ICs.

[Marcos - Munky]

In my opinion, diylc's "factory setting" for solder pads is a little too big, at least on the version I have on my flash drive. You can make them smaller, so there'll be more space between them and the traces.

Btw, here's mine. The initial idea wat to have speed and resonance pots, then changed my mind to a fixed 1M resonance and the droop switch. After Mark's demo, I changed my mind again and put the bias as a panel pot. I was going to try a water slide-like finishing based on the original artwork, but I liked the bare aluminium look so much it ended up with no artwork at all, just the knob colors to resemble the original artwork a little bit. First photo with flash, second photo with some pink lens flare .

[Mark Hammer]

Just an update.  Finished another one for a buddy today.  I used J113s for the JFETs, a BC307B for the noise source, and a TL062 for the dual op-amp.  Sounds great, and maybe even better than the one I Youtubed.

This is not to say you MUST use those same components.  Rather, there will be many different JFETs and noise transistors that can work just fine with the circuit, so feel free to experiment.

[Kevin Mitchell]

Sweet deal! Thanks for updating. I'm considering getting some boards made for this. Would be an easy job and the folks who don't etch can get in on the fun.

[Mark Hammer]

Might be worth considering a "Range" switch for the filter, so that bass players can get in on the fun.  I find I can move the range around with the Bias control, but just not low enough to do justice to a bass, 7-string, or baritone guitar.

[rankot]

Just finished basic version of this. I will try all the board mounting pots, as I found BIAS trim very interesting. DEPTH has really limited range, so I will try what Mark suggested - 5k+5k pot.

[rankot]

OK, now the final build report - although I used my own PCB design - but some findings can may be useful to everyone:

• I can't find any difference with Resonance pot. Tried to put it in series with 620k (as Mark proposed 470-680k) and use B1M pot, but I can't hear any difference, so I will put fixed 1M resistor here. I don't know why is this so, and what does it have to do anyway?
• I also found out that the best range for Depth is from 8k2 (from ground to CCW pin), with B2k2 pot, so I left it like that in my circuit.
• Speed pot is fine, but I have reduced 150k resistor in series with it to 51k. It can be even set for shorter times, which can be interesting, since I don't find too long time much interesting.
• I use B50k for Bias pot, together with 62k in series with CCW pin to -9V. It gives a good range (and could maybe enlarged to 68 or even 75k), but it is also interactive with the position of Depth, so I will go with this combination.
• I tried 2N3904 as a noise generator, but it's far from expected, so I tried few of my 2N2222A and find them almost all the same and sounding very good. It's socketed, anyway, so I will try with different transistor types later, if I don't get bored with this.

Hope I didn't forget anything - also, there's a question for Mark - he noted that Bias range can be altered for bass, but didn't mention how. Help would be appreciated.

Also, what was the purpose of the filter section of the original FSH-1? I thought of possibility to make an additional filter PCB with a 3PDT switch. It will require 1590BB, but may be interesting.

And I have some bubbling sound even with no sound input (both magnet volume pots at 0), is it normal or not?