Weirdness with osc -> trigger circuit

[elmegil]

I'm working on what I want to be a variable clock circuit, generating triggers rather than gates (i.e. pulses instead of square waves).

Step one: simple LFO with a 40106, check.

Step two: condition the output with a trigger conditioner. Fail.

I've tried this both with the one that Fonik has in the baby-10 sequencer schematic (upper right: http://modular.fonik.de/pdf/baby10.pdf) and the simpler one from Thomas Henry's Electronic Drum Cookbook (although I didn't swap the LM358 for a 741 when I simplified it, but a single sided op amp is a single sided op amp... for this purpose anyway). The simpler circuit omits the first 100k, the diode, and the 47k from Fonik's schematic.

Both of them exhibit the behavior illustrated in the photo. Yellow is the output of the 40106 osc, and blue is the output of the trigger conditioner.

1) I have a nasty >5V triangle wave across the top of it, limiting my amplitude.

2) I am periodically getting triggers on both the rising and the falling edges of the clock signal..

If you look carefully, you can see some distortion in the 40106 output that maps exactly to the triangle wave (and it's 180 degree mirror).

I've also noticed that the apparent triangle is insensitive to the time scale. Switching from 50ms as pictures to 200, the triangle appears to have the same shape-- it doesn't compress like the width of the pulses, for example. I've seen it before and always written it off as an artifact of my scope. It also doesn't change if I use the 10x setting on the probe, or if I move the probe (I thought at one point I was injecting it from the scope, and the glitchiness in the square wave was causing the variability in the trigger pulses).

Now, I'm not in a pristine RF environment, but 5V?? I don't see that in the ground, or V+ lines. Bypasses don't make any difference. Grounding all the various unused inputs doesn't make any difference.

I *am* using full length leads on several of the resistors, rather than cutting them to length for the breadboard.

But I'm lost here.

Any ideas?

---

2013-01-29_22-59-45_262.jpg
Description:
Filesize:	2.59 MB
Viewed:	24 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[marvkaye]

Pretty interesting.. it looks like your conditioned triggers are being modulated by another LFO. Just curious, but do you get that same triangle wave shaped artifact when you scope a different signal.. like one of your operating VCOs or LFOs? IOW, is it something from within your scope? Seems to me if it's not affected by the channel timebase it's something internal, maybe somehow related to the screen power... maybe you could isolate it using the scope's block diagram or schematic. The weirdest part is that the triangle and its mirror are actually visible on both channels (I didn't notice it at first in the triggers, but see it now... strange...)

If it doesn't show up when you scope another module that would mostly eliminate the scope itself as the culprit (I would think). The curious thing for me is that it's a constant relative to the screen and not the signal being displayed (at least that's what I take from your description...).

I'll be anxious to hear what it finally turns out to be. By all means, keep us informed if you make any progress with it.

<marv>

[elektrouwe]

[frijitz]

[elmegil]

Do you think that's true of the double triggers as well?

I've tried slowing things down and see similar results, which in the past has solved more obvious aliasing problems.

Any suggestions for how to test in the absence of a handy analog scope?

[brock]

[elektrouwe]

[elektrouwe]

[elmegil]

I haven't looked yet, but do you know of a 40106 oscillator schematic that lets you have other than a 50% duty cycle?

If I could do that I wouldn't need the trigger conditioner. I'm not terribly hung up on having only a 1-2ms pulse for the clock.

I will draw up the schematic when I get a chance. Right now I've got Jury Duty every day, and no end in sight yet so I don't have the time I'd normally have to do this.

[PHOBoS]

you can change the dutycyle by placing a diode in parallel with the feedback resistor.
With or without a small series resistor depending on how long the pulse needs to be.

[diablojoy]

if your already using part of a 40106 for the oscilator why not just use another 2 sections of it as a one shot and an invertor ?
that will give you a determinable trigger pulse of what ever length you like.

[elmegil]

How does one set up a 40106 inverter as a one shot?

[diablojoy]

very simple
try breadboarding this, i havent checked it
but i think it should work with the values given

---

40106 one shot plus output driver.pdf
Description:		Download
Filename:	40106 one shot plus output driver.pdf
Filesize:	10.29 KB
Downloaded:	33 Time(s)

---

[elmegil]

Ok, that's basically doing the same thing I was trying to do with the comparator, which is where it appears to be getting double triggered. I'll give that a shot. This is all on breadboard at the moment, for exactly these reasons.

Thanks

[brock]

Looking at the scope shot for the third time, it is pretty obvious you have double triggering. I think you need some hysteresis (positive feedback) on the comparators. I'd try 1M to see if that fixes it and go down by 50% steps if it doesn't.