problems with a 40106 monostable

JingleJoe · Posted: Fri Apr 20, 2012 5:09 pm Post subject: problems with a 40106 monostable

I had this bloody circuit working fine, now that I'm acctually using it in my synth it doesn't bloody work properly Evil or Very Mad

I'm making a delay circuit for a pulse, it is only about 1ms long but thats enough; it activates the first monostable which is approximately 2ms long, which then activates another monostable with a pulse length of 1ms again. This delays the first 1ms pulse by about 2ms which is plenty for what I want.

Simple, right? Wrong. First one of these I built oscillated and wobbled so badly I just started again.
This one works considerably better but I still get erratic oscillation, adding a pull down resistor to the input of the inverter after the first monostable helped, but not enough, I still get bursts of oscillation after some input pulses. The oscillations seem to come from directly after the first stage of the delay circuit, the 2ms monostable.

Could this be down to the crappy ceramic caps ringing? Would perhaps polyester work better? Or maybe my scope is introducing some interferance which causes it? Or is there anywhere in my circuit which could use another pull down/pull up resistor?
I've been working on this one circuit for more than 6 hours tonight. Please help me I am very discombobulated. Sad

pulse thingyo.png

Description:

Phase relationship of pulses shown throughout circuit. Chip: 40106 inverting schmitt trigger buffer.

Filesize:

7.27 KB

Viewed:

466 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

elmegil · Posted: Fri Apr 20, 2012 8:58 pm Post subject:

If you think it's your scope...do you have/have you set the probe to 10x? I found that some of the circuitry I was working on with a DSO Quad baby scope would get significant capacitance effects (square waves becoming ocean waves Smile

) at 1x.

Not sure whether that would make you oscillate, but its at least something to rule out....

diablojoy · Posted: Fri Apr 20, 2012 9:42 pm Post subject:

try a pull down resistor right at the start on the input
if you have this built on a breadboard try a different position on the board also be aware breadboards have loads of stray capacitance
did i mention that i hate breadboards ?

Inventor · Posted: Fri Apr 20, 2012 10:24 pm Post subject:

Hey Joe, I know you must be frustrated because your post lacks it's usual wonderful sense of humor! I would not generate pulses the way you are doing it, I'd use either an LMC555 timer chip or chips or one of the CMOS beasties of the same variety.

Remember that a CMOS input is super high impedance, so when you put a cap in series between the output of one inverter and the input of the other inverter, you are relying on the leakage of the protection diodes to discharge the pulse, and to what value we don't know. All very quirky. Adding el super-high value piece-de-resistance to ground or Vcc will help but is still possibly too weak at 220k.

Realize also that you are creating a high pass filter between the inverters, which will create a pulse, but I think you are pushing the duration of that trick too far with large time delays and small capacitors and huge resistors.

Just my two cents. Back to your normally scheduled lab work!

Les

JingleJoe · Posted: Sat Apr 21, 2012 1:58 am Post subject:

Inventor, are you suggesting that I use lower value resistors and higher value caps? And yes, the CR networks make filters, which give sloping edges to the pulse, resulting in a delay. But surely the diode network inside is fine? It should not affect the cap if it has a pull down/pull up resistor?

Inventor · Posted: Sat Apr 21, 2012 2:20 am Post subject:

Ah, that helps, Joe, to hear that it worked fine on a freadboard. The 10nF caps are only 10,000 pF in value, which is really quite small. remember that a breadboard has a LOT of capacitance from every adjacent node to any given node. Your actual breadboard circuit had a 10nF cap in it and a bunch of little caps to all the other nodes including power and ground. That will affect it.

Also your scope probes being bits of wire will have a lot of capacitance to them. every time you probe a node, you change the capacitance to ground from that node, and in a big way compared to 10,000pF. This is a common thing that intermediate circuit designers run into, the maddening scope capacitance! As you prove different places you get different behavior and that's confusing as all get out.

Bigger caps and smaller resistors are the way to go here IMHO. Either that or the CMOS timer chip option.

Good luck with it, and I'm awake now so post if you get any other problems. Inventor Hotline: 1-800-electro-music.

Les

JingleJoe · Posted: Sat Apr 21, 2012 2:56 am Post subject:

Thanks inventor Smile

I will try all those suggestions! I will be going out today to see Sea Odyssey so I will pop over to the elecronics shop too and get some cmos monostables incase I can't get this one to work.

Inventor · Posted: Sat Apr 21, 2012 3:40 am Post subject:

OK keep us posted!

Les

elmegil · Posted: Sat Apr 21, 2012 5:14 am Post subject:

Yah, what inventor said about the scope... You'd want high impedance on your probes. My first thought was "well put a big resistor in series with your scraps of wire" but then there'd be no calibration, so.... probably not the greatest idea, but might help in a pinch.

Dave Kendall · Posted: Sat Apr 21, 2012 7:05 am Post subject:

Hi Joe.

FWIW, the 4538 is a peachy little chip. It's accurate enough to do fast timing stuff for MIDI. Smile

cheers,
Dave

JingleJoe · Posted: Sat Apr 21, 2012 9:18 am Post subject:

I tried changing the caps and resistors, to 47nF and 47k and 22k, but to no avail Sad

Inventor · Posted: Sat Apr 21, 2012 10:22 am Post subject:

Joe, I just got back from an outing and saw your message so I took a closer look at your circuit. You've got that first resistor going to ground, the second to ground, and the third to Vcc. Tie the first one to Vcc, not ground. Here's why:

Your input pulse is positive so the first inverter produces a negative pulse as shown in your diagram. The resistor pulls the input of the second inverter down, so it's at ground. Now the negative pulse comes along and zap! it causes the input of the second inverter to go below ground, and the second inverter never triggers because it's input never crosses the threshold voltage of Vcc/2 (or thereabouts). The other two resistors seem OK to me.

When you think about it, since inverters invert, the resistors should be in the pattern of pull up - pull down - pull up, etc, either one way or the other (pull down - pull up - pull down). So try that and see what happens. I'll be around all day so post your results and I'll try not to lead you astray, we'll get this thing working for you.

Les

Inventor · Posted: Sat Apr 21, 2012 10:35 am Post subject:

Oh yeah, and congrats on your new scope! New gear is always such a thrill, even more so if you're poor lol. It's damn near ecstatic for me at my income level! Well, things seem to be getting better in that regard. Do remember to celebrate by posting a photo of it with some cool waves on it in the New Gear thread!

And do the New Scope Dance! "Hawonga, baronga, tonka wonka new toy!"

Les

JingleJoe · Posted: Sat Apr 21, 2012 11:19 am Post subject:

I am confused by that now inventor Sad

I will re-read what you said later and think about the circuit more. I don't think anything goes below ground though, I checked the cap waves, they only went one way; saw tooth like shapes. I must admit I am no expert on the above circuit however I did figure it out and have it working once Smile

I trust my past self not to fuck things up for me too much Laughing

The current issue is the formula for the 4528 time period Confused

not only is it bloody well hidden in the datasheet but it is confusing me, I think I am right in interpreting that it is: Time period = 0.2 * C * R * (ln V+)
I am only 50% sure that ln is the natural log so ln of 12 according to my calculator is 2.485, can anyone confirm this for me?

For now I'm going with; C*R*0.2*ln = Fuck it, I'll just use a 556 timer

JingleJoe · Posted: Sat Apr 21, 2012 2:25 pm Post subject:

Damnit, I'm still getting oscillation with a 556 timer. I'm going to try a different kind of capacitor, polyester or polystyrene or something. I have a feeling it's down to the ceramic caps.

Inventor · Posted: Sat Apr 21, 2012 2:42 pm Post subject:

wow, this could be an entirely different problem. what's the power supply setup and what else is in the circuit, is it decoupled properly, etc. let's look at the larger picture.

Les

JingleJoe · Posted: Sun Apr 22, 2012 3:25 am Post subject:

IT. IS. WORKING. Shocked

Inventor · Posted: Sun Apr 22, 2012 6:21 am Post subject:

Aha! well that's a good reason not to use those cheesy circuits, they have no hysteresis. low pass filtering the signal is a quick fix but the real deal is to add hysteresis, or positive feedback (very small amount).

What's happening, as you must realize now, is that when the circuit approaches the threshold voltage, the noise on the signal dances across that line causing a temporary oscillation. Then the inverter with all it's gain slams it's output back and forth and since it inverts, the feedback is negative and you get a sustained oscillation.

Also on the breadboard you had that much capacitance to gnd and Vdd, both causing a low pass circuit to exist, due to the parasitic capacitance of the breadboard. On your protoboard circuit such parasitics are much smaller so you had less of a filter to smooth things out for you.

It's all very easy to see in hindsight, and fortunately we have reached that "aha!" moment so all is well. The solution of course is to use a pulse generator IC to generate pulses, lol, not a chain of inverters. Lesson learned and good find JingleJoe!

Les

JingleJoe · Posted: Sun Apr 22, 2012 2:55 pm Post subject:

Those inverters do have schmitt trigger inputs, with a hysteresis of about 1 volt. I reckon they would have worked had I filtered the signal, however the 556 version is working fine now so to hell with them Laughing

Inventor · Posted: Sun Apr 22, 2012 3:53 pm Post subject: