Mini Sample Hold - Transfer Layout

[Kevin Mitchell]

This jfet pin thing has got my head spinning 
Just to verify...


Looks good to me 


Also - I don't think you can fit the circuit into a 1590A without some serious work. I've tried and decided to stick with a 125B. Maybe an SMD version?

-KM

[Marcos - Munky]

I always got confused on jfet pinouts. We're supposed to spin the jfets, not our heads . The image got everything clear now.

As for the 1590A layout, yeah, it'll probably be a not so easy work, even because it's been some time since I did a not so simple layout. And no smds allowed

Edited: so I checked the last schematic again, and the part that got me confused is still there. The arrow should be the gate, but on the schematic it's labeled drain. That means drain and gate labels are reversed if you check jfet symbol against the labels. The layout is correct if you check the labels, so it's incorrect if you check the symbol. So, just to be clear, should I forget the jfet symbol and just follow the pin labels? I'm just missing the jfets part to finish the schematic on eagle and start the layout attempt.

[Kevin Mitchell]

Edited: so I checked the last schematic again, and the part that got me confused is still there.

I know what you're saying. And the more I dig into it the more confused I get. I pulled up the GGG layout and those jfets are connected differently than the tonepad one as well as mine(since it was a rip from the tonepad one). The GGG is drawn correctly as the symbol shows but I can't tell you if it even works.


If you're not confused enough - here's the tonepad one that I've followed and labeled my jfet connections to accommodate (remember, they said to use BF245 but labeled the "G" pin on the wrong side);


I can vouch that I've followed my schematic on breadboard a few times without any trouble. I have a couple of videos proving so. Though the GGG is drawn correctly, I can't tell you if it works the same or at all.

Since the tonepad one is more dated - I wonder if GGG copied the schematic and had not noticed the connections aren't as expected. Has anyone successfully built one? I have no idea.

I'll have to breadboard this thing again and try both configurations. Again - I know my drawing works.

-KM

[duck_arse]

I think in circuits like this it is safe to say you can swap your drain pin with your source pin at will. Kevin's circuit three posts back, however, takes the pin name swaps a bit too far. the "tonepad" seems to have taken the most convenient of the S/D pin/positions to easy-up the layout.

[EBK]

I'll put in another plug for my preferred way of laying out JFET stuff to avoid pinout problems:
https://www.diystompboxes.com/smfforum/index.php?topic=120323.msg1126755#msg1126755

[Kevin Mitchell]

Kevin's circuit three posts back, however, takes the pin name swaps a bit too far.

I've labeled the schematics to show how they are connected in relation to the jfet pins. I didn't take it too far - I recorded tonepad's layout to prove that I followed a working circuit.
-KM

[Marcos - Munky]

The more I look at the pinouts, more confused I get . It seems to me tonepad layout have the gate on the correct side, but got the jfet orientation reversed (for a BF245). Anyway, I'll just start my layout based on what I got from your layout. If it works, it works. If it doesn't, it's just a matter of rotating a jfet or twisting a few pins

[Mark Hammer]

Checking the datasheets for the 5457s I have, I discovered that the pinouts were actually as Kevin originally suggested, with the flat side facing outwards, instead of inwards.  I reinstalled the FETs and the circuit sprang to life.  The filter and noise sampling seems to  work fine, but the input buffer/gain stage seems a bit problematic.  I have checked and rechecked the PCB and solder joints.  I also swapped an LM358 in place of the TL072 to lower audible ticking.  However, the signal level is still well below bypass level and the filtering limper than the resonant noise would suggest.

[Mark Hammer]

Got the 2nd one working properly, after finding a nice noisy 3904.  I'm beginning to think that, while Kevin shows it as a 2-knob effect, the Depth and Bias trimmers really need to be panel-mount controls to provide the most control over the range  of filter steps produced.  I won't say that the entire range of settings of those two controls yields something desirable, but a considerable portion of their rotation is useful and usable.

I suppose the only thing I don't like about it is that a teensy bit of glide would be nice. so that the steps aren't robbed of their character and contrast, but don't produce as hard a step as to sound like a pop.  Is that 220nf cap on the way to the CV feed game for tinkering with?

[Mark Hammer]

But since I don't think I have stated it explicitly enough, nice job Kevin!

It has also gotten me wondering if the clock and noise-source could be harnessed to a simple tremolo circuit for semi-random amplitude modulation.  I mean, why couldn't the 2nd FET holding the "sample" be the one setting the moment-to-moment gain in this circuit (Q2), and the LFO be replaced with the noise/clock? I know that ZVex has a sequencer-based tremolo, but why not a random stepped one?

[Kevin Mitchell]

I'm glad it's working for you! Though I wonder what's going on with your first board.

I'm going to try to breadboard this and see if I can add a glide control. I think that's a brilliant idea and would be very useful for this circuit. I'll also see about limiting the depth and bias so they are useful as external controls. When I breadboarded these a couple years back I didn't like messing with them much. We'll see how it goes.

-KM

[Mark Hammer]

Thanks for the interest, Kevin. I think the Bias and Depth controls are actually fine as is.  Only a very small portion of their range yields nothing of interest or use.

As for the first one that isn't working properly, it's quite possible the 13700 is fried.  I'll replace it later today and see if that yields more success.

But I've gotten quite curious about the suggested semi-random tremolo, using the noise+sample subcircuit.  It's not so much a random modulation, given that it is clock-stepped.  Rather, it would be the amount of attenuation with each step that would be less predictable.  One of the effects on my Line 6 M5 is based on the Lightfoot Labs Goatkeeper tremolo, which essentially mixes 4 different modulation patterns to yield different results that can seem to speed up or slow down.  https://www.youtube.com/watch?v=3uj5sm0Uz8A  The ZVex Seek-Trem uses a simple 8-step sequencer to produce more-or-less nonlinear patterns, but it repeats the same pattern so that it is effectively a tremolo with variable waveform (although it permits "gaps" to be inserted).  https://www.youtube.com/watch?v=7KfKRhuCwFI  Using the S&H subcircuit to drive a tremolo would yield a consistent pacing of the modulation but vary the degree of modulation in largely unpredictable fashion.

[Kevin Mitchell]

Interesting! I may consider adding CV inputs on all of my modulation pedals 
I need another tremolo - probably the Lune. I'll see how the S/H lfo pairs with it.

I plan to build my MSH circuit next week. I guess I need to restock on some parts. I knew the day would come
I'll add a switching jack to the order so I can hijack the S/H CV.
-KM

[Mark Hammer]

Boss pedals also employ FET switching with a 1M resistor feeding the gate of a FET.  To make the "turn-on" less abrupt, they also include a 47nf cap to ground from the junction of the 1M resistor and FET gate,  I'll see if that provides some relief from audible step-popping.

[Mark Hammer]

So I got the two units working fine.  T
As suspected, the dysfunctional one had a bum 13700.  One side is working fine, so you'd hear filtering of noise but you wouldn't hear much input signal.

I tried the experiment running a 47nf cap from the gate of the switched JFET to ground, as seen in Boss pedals.  Didn't work.

[Marcos - Munky]

So, a 1590A build is doable:


I take no credit on the layout, since I basically used Kevin's 1st layout and modded it to fit a 1590A.

It's tested and working. But, even I like how it sounds, it's still very different from the demo. Resonance pot it's innefective for me, it just sounds cool when it's maxed out. Also, tonepad's calls out for a 2M pot, while Kevin's schematic calls out for a 1M pot. And the oscillations doesn't sound as "rangey" as the demo. So I'll try other transistors and add more resistance to the resonance pot and check how it goes.

[Mark Hammer]

Resonance is not generally one of those things one desperately needs a pot for.  A 3-position toggle can often give you all the variation you need....including values not available from standard value pots. 

[rousejeremy]

great project.

[Marcos - Munky]

Today I played with the circuit a little bit more. For the resonance, I ended up with removing the pot. I tried to add a 1M resistor in series with the pot and found out it sounds pretty good with no pot at all. For the transistor, I ended up with a BC107B with the collector floating. It was pretty hard to find a good noise transistor, I tested a lot of them. I'll play with it a bit more tomorrow.

[Mark Hammer]

Choice of noise transistor makes a huge difference.  I was working through my bag of 3904s, and first thought the circuit was malfunctioning in some way that would necessitate further troubleshooting of the board.  Then I found "the right one", stuck it in the socket and instantly he circuit sprung to life.