I'm trying to use the ~0.1Hz output of one timer on a 556 to induce a slow-changing shift in frequency of the second timer at ~0.5Hz, by routing the output through R5 to the control pin of the 2nd timer and C3 in the attached. The timers are both working nicely on my breadboard.

I'd like to maximize the shift in frequency of the 2nd timer, up to a factor of two in either direction from the mean, ie. from 0.25Hz to 1Hz. With the current component values, I think I can see some variation but it's hard to tell and probably not a halving-doubling.

The math on this is a bit much and I'm hoping someone has experience with this approach or could simulate this for me. How much shift can I get, and how can I get this range to be stretched out to, say, 5 seconds of increasing frequencies followed by 5 seconds decreasing?

 

Your diagram shows the reset pins open, this is an illegal condition. Tie those suckers to Vcc!

 

The resets on bjt 555's have the equivalent of 100k pull-ups on the reset pins, but the CMOS versions don't. Either way, you risk getting "bitten" if you don't tie them to Vcc.

You're not going to get quite as much shift in frequency as you'd like; maybe from 0.33Hz to around 0.85Hz. The big thing that will change is the duty cycle; at the higher frequencies the duty cycle will decrease very significantly. That's because when the threshold is so low, the timing cap charges really quickly, but the discharge takes a relatively long time.`

 

... you risk getting "bitten" if you don't tie them to Vcc.

Easy enough to do that. But does the very low frequency off my application make this less of an issue? It's obviously working fine w/o this. I suppose an occasional reset wouldn't be an issue as long as it doesn't lock up altogether.

...maybe from 0.33Hz to around 0.85Hz.

I can live with that. I just want to maximize whatever I can get with this dual-555 strategy, given the general timing goals (~1 second pulses getting faster for ~5 seconds and then slower for 5).

 

Well, I posted a week or two ago about using a 4017 to change timing caps for a 4093 NAND free-running astable multivibrator if that would be of interest.

 

Well, I posted a week or two ago about using a 4017 to change timing caps for a 4093 NAND free-running astable multivibrator if that would be of interest.

Yes, and that would give greater control, but I opted for the lower parts count and smaller footprint of this "good enough" approach.

 

The resets on bjt 555's have the equivalent of 100k pull-ups on the reset pins, but the CMOS versions don't. Either way, you risk getting "bitten" if you don't tie them to Vcc.

<snip>

Uhhh, no.

It may depend on the 555 make you are getting, but I don't recall ever seeing any schematic of a 555 with what you describe.

This is from the NE555 datasheet.

The reset pin is a digital input, you must tie it to something. Always a good rule for logic, except maybe TTL.

I admit I'm curious. I'll have to try it out sometime. I have run experiments to see how much voltage you need to turn it on, it was a minimum of 0.7VDC (call it 1.0VDC). I've never tried it totally open before.

 

I just did that experiment, without realizing it. But I'll nail them to Vcc for my build.

 

You're not going to get quite as much shift in frequency as you'd like; maybe from 0.33Hz to around 0.85Hz. The big thing that will change is the duty cycle; at the higher frequencies the duty cycle will decrease very significantly. That's because when the threshold is so low, the timing cap charges really quickly, but the discharge takes a relatively long time.`

I finally took the time to run thru the math and (no surprise ) you're right on, more or less. With the values given and ignoring the effect of C3, driving the control pin low can get the frequency close to 1 but never quite there. Driving it high pushes the frequency down to about 0.07 as a limit. So the maximum frequency range shift is about 10X using this "police siren" strategy of alternating Vcc and ground to the control pin.

In my case I want a slowly moving frequency, not just two levels, so I use C3 to "integrate" the voltage at the control pin. I have to use a larger value of R5 so that C3 can have an effect without being impractically huge. At R5=630Ω, the upper frequency is 0.888 and the lower is 0.222Hz, giving a 4X frequency shift as I first said I wanted. However that would still require a very big C3 to give reasonable smoothing.

Compromising a bit more with R5 at 1000Ω as first shown gives 0.85 and 0.25Hz, a range of 3.4X. But the RC time constant for C3 is still only one second, and my period is 5 seconds for the first timer. Doubling R5 to 2000Ω narrows the frequency range to 2.6X but makes the effect of C3 more pronounced. If I want a wider frequency range I'll have to find a bigger C3, so I can reduce R5.

And yes the duty cycle changes because the discharge stays the same while the charge time gets much smaller as the frequency is driven up by grounding the control pin. I'm only using the 556 as a clock for a 4017 counter, so the pulse width doesn't matter to me.