SOT: 555 voltages "overshoot" and "ringing"

[brett]

Hi
I've got a bunch of 555s that trigger each other and run as free oscillators, etc.

When I ran them off +5 V (from a 9V battery and 7805) and ground , they exhibited a little overshoot.  Not too bad.

Since then I've changed things to run then off +5V and -5V from a 7805 and a 7905, the overshoot has got worse, and there's an osciallation that lasts several uS after each switch.  The supply is two 9V batteries, 4 x 470uF caps, the 7805 and 7905, 2 x 22uF caps and 2 x 0.1uF film caps.  I had hoped that the film caps would prevent this sort of thing.

Would going to CMOS 555s help?

I need to have them switch from +5V to -5V because I want to drive both MOSFETs and JFETs from them.
thanks

[Cliff Schecht]

I assume that you aren't running on a breadboard? That could easily cause problems there.

Also, you mention that you are seeing a switching time in uS? 555 timers don't really like to go faster than a mS or so, past that and I've ran into problems myself (well, I needed 10 uS and the 555 can't really do that comfortably). For uS pulse widths, I went with a 74hc4538 and haven't had any problems. For oscillators, maybe look into some other 74xx series cmos parts?

Remember that overshoot is caused by the pole(s) of your system getting very close to the imaginary axis on your pole-zero plot. The closer they get, the worse your overshoot will be. Look for stray capacitances or inductances and make sure you have a strong ground, these can all be culprits of excessive ringing. 

[R.G.]

Unless your scope probe is a 10x and calibrated, it can inject overshoot all by itself. Never seen one ring though.

[alanlan]

Where are you connecting your oscilloscope ground connection?  To the -5V or 0V?

[brett]

Thanks for the help.
My switching times are in the 10 to 600uS range.

RE: Where are you connecting your oscilloscope ground connection?  To the -5V or 0V?

"Ground" is -5V for the scope and the 555s.
cheers

[Cliff Schecht]

Thanks for the help.
My switching times are in the 10 to 600uS range.

RE: Where are you connecting your oscilloscope ground connection?  To the -5V or 0V?

"Ground" is -5V for the scope and the 555s.
cheers

I don't think 555's are very reliable at 10 uS, you start running into parasitic capacitances that throw off the accuracy of the circuit.

Also, putting the scope at anything other than a true 0 V ground isn't a good idea. The inputs are designed to look at signals with reference to 0 V and nothing else, otherwise you start getting the -5 V return currents showing up in the scope outputs. The inputs are direct/capacitively coupled for a reason, to allow you to look at any signal whether it has a DC offset or not. Does it do the same thing when you put the scope at 0 V ground?

[Rob Strand]

The bipolar 555's have overshoot.  You can't get rid of it.  You can tame the overshoot by placing load resistor on the output, start around 2k2 to 4k7.  IIRC, connect the load to ground (as opposed to +V).     Run a 100nF in parallel with a 10uF tant. on the rails. 

The CMOS 555's are *much* cleaner.

[gez]

Brett, when you say you have 'overshoot' then presumably you're using pin 4 of each 555 to enable/disable oscillation?  Or am I missing something?

Anyway, if that's the case have you considered using a 4093 chip?  They can be wired as Schmitt trigger oscillators and the 'free' input used for gating.

[brett]

Hi
thanks for the ideas.  I'll pick up some TLC555s tomorrow and have a look. (I ALWAYS smile at "TLC" ! Who thought of that?)

Gez - I tie pin 4 to pin 8 to Vsupply (+5V).  Seems that 99% of astable and monostable circuits that I've seen do that. (or maybe I am just copying a bad habit?)

Fortunately, this is mainly a theoretical nuisance, as the ciruit seems to hold the MOSFETs and JFETs "on" and "off" at the right times.

Thanks all!