Integrator/readout for output of a 555

[brett]

Hi.
My multimeter doesn't have a capacitance meter on it, so I'm building one. (It's a Schmitt trigger clock driving a 555 in monostable mode.)

The capacitor under test will modify the time high/duty cycle of the output.

Now I'm wondering how to measure the time high/duty cycle.  I was thinking of some filtering (simple RC network??) and a voltmeter, but that doesn't seem to work very well.

Any help would be greatly appreciated.

[brett]

Bump.
Help anyone?

[Rob Strand]

>I was thinking of some filtering (simple RC network??) and a voltmeter, but that doesn't seem to work very well.

It does work.  It depends on the actual frequency you are switching at but I think ~10k and 22uF would be OK.

The problem with the common  bipolar NE555 is the output doesn't switch between 0 and Vcc.  There is a saturation voltage drop at both high and low outputs and they are tempreature dependent.  The CMOS 555's like the TLC555 and 7555 switch to the rails and work better for pwm voltage outputs.  The other issue all the 555's have a an effective input capacitance of ~30pF which creates a capacitance offset in the measurements (there's a number way of dealing with that).

[brett]

Thanks Rob.
I've been switching at 50Hz, and was using 10k and 22uF (exactly what you suggested!!).
I *think* my problem was more to do with having a potentially ong duty cycle cut short by re-setting from the clock.  I'll have to have switchable R values, so that RC doesn't lead to "on" times of more than about 10mS.  I was confused because I was changing C and the duty cycle was staying at 50% and the output voltage was sitiing at 1/2 the supply.

And yes, I'm using a CMOS 555.  For a few extra cents, they have so many advantages over ordinary 555s.

I'll post a schematic when I get this thing going properly.
Thanks again

[Rob Strand]

> I *think* my problem was more to do with having a potentially ong duty cycle cut short by re-setting from the clock.

A 555 monostable needs a negative going pulse that is shorter than the shortest pulse.  There's a couple of ways to do this.  The easist is to take a cap from the output (pin 3) of the astable to the trigger (pin 2) of the monostable.  You then have a resistor from pin 2 to vcc and in parallel with that a diode with the anode on pin 2 and the cathode to vcc.  Values which come to mind for a cap meter are 4.7k and 330pF.  What this circuit does is create short negative pulses.

[brett]

Hi Rob.
Thanks for that idea.  I've used a similar idea before, but to keep the duty cycle of an astable short.  I've also used a diode to bypass R2 in the usual R1, R2 and C schematic.  That way, the duty cycle is set almost independent of R2 (there's a small effect of the 0.7V voltage drop), and can be as low as you want.

So far, I was using a Scmitt trigger for a simple oscillator, but the output is a fixed 50% duty cycle.  Your ideas make me think that I can use an RC filter tied between the astable output and Vsupply so that there is only a transient negative pulse as the astable goes low.  Thanks heaps for your help.

[Rob Strand]

The diode method to get a narrow pulse does work but if the capacitance is large you will find the pulse width is actually longer than expected, which can cause problems for small capacitance readings.  The problem is he 555 as an effective impedance on the discharge pin which cause the disharge time to get longer.  When you get into that region the size of the timing capacitor is the only way to control the pulse width (smaller = shorter pulse).  I had a minimum parts version where I didn't used a diode, I shorted "R2", chose C to get the pulse width (as I recall C was around 47nF) then choose R1 to get the astable period - it works quite reliably and consistently across devices.

The 555 has many quirks! (especially when the you get near the minimum pulse widths and periods)

>I was using a Scmitt trigger for a simple oscillator

Yes using that ckt will require the pulse shaper.