0
\$\begingroup\$

I would like a slow PWM output, driving a relay turning an AC heating element on and off. I desire a period time of around 10s.

What is a realistic upper period time limit for a PWM circuit like the one found here?

According to EveryCircuit, a 100 kOhm potentiometer and 100 uF C1 capacitor should leave me with my desired ~10s period time, but I suspect noise could become an issue with resistances so large and period times so slow.

\$\endgroup\$
9
  • \$\begingroup\$ Are you running this open-loop? (Are you just manually set the PWM duty cycle and have no electronic feedback used to adjust the duty cycle based upon readings of some process variable under control?) \$\endgroup\$
    – jonk
    Commented Jan 2, 2019 at 21:16
  • 2
    \$\begingroup\$ I'm not sure why you think noise would be an issue. One of my very first electronic projects ever was a 1-hour 555-based timer that used something like 1000 uF and 3.3 M -- well outside the recommended limits, but it still worked fine for many years. \$\endgroup\$
    – Dave Tweed
    Commented Jan 2, 2019 at 22:14
  • \$\begingroup\$ @jonk Yes, open cycle it is. I would just put some trial-and-error scale on the potentiometer - doesn't need to be precise / repeatable. \$\endgroup\$
    – towe
    Commented Jan 2, 2019 at 22:37
  • \$\begingroup\$ 10s has definitely been done, but I'll be honest and do whatever I do whenever someone asks about the 555 in a PWM application: Chances are, considering cost of passives, a cheap microcontroller would do just as well but be way more flexible. In this case, any cheap microcontroller hooked up to some kind of analog or digital thermometer would, in the next stage of development, allow for a 10 line addition of source code to convert this to a PID controller – which would, in a very rough analogy, be the equivalent to a trial and error scale, only that the controller does the trying and \$\endgroup\$
    – Marcus Müller
    Commented Jan 2, 2019 at 22:46
  • \$\begingroup\$ "homing in" on the desired temperature, instead of you sitting in front of whatever you're heating and turning the nob, waiting, checking, making a mark on the scale, turning, waiting… \$\endgroup\$
    – Marcus Müller
    Commented Jan 2, 2019 at 22:47

3 Answers 3

Reset to default
1
\$\begingroup\$

With tantalum electrolytic capacitors you can choose ones with ultra low leakage @ Vmax equiv to 1M to 10M +.

But it is far better to use a Triac and temp sensor and driver to triac with hysteresis that controls the drift temp. about any setpoint or use phase control for linear PWM control with some gain. to reduce the error or PID loop for ultra fine control. (<<1'C)

\$\endgroup\$
1
  • \$\begingroup\$ +1 for suggesting a tantalum capacitor. \$\endgroup\$
    – Caleb Reister
    Commented Oct 23, 2019 at 18:17
0
\$\begingroup\$

At 100uF that's either a very large film capacitor or it's an electrolytic capacitor, the latter mar vary in capacitance significantly with temperature,

100K should be fine, so long as the wires to the 555 are short, even 1 meg would work here.

The analogue inputs are connected to the timing capacitor so they won't see much noise.

The capacitor on pin 5 (control voltage) significantly reduces noise succeptability By filtering the other internal comparator inputs

Supply ripple is not an issue so long as it's less than about 25%

\$\endgroup\$
0
\$\begingroup\$

I second the microcontroller approach. Attiny10 is less than a 555, support components would be less too. Gives you flexibility if you want to change the time in the future without having to remove & replace components.

\$\endgroup\$

Not the answer you're looking for? Browse other questions tagged or ask your own question.