555 ringing

Thread Starter

Sam1951

Joined Nov 4, 2019
4
Hi All:
I'm new here so hope I'm in the right place..... As part of a larger project ( Computer P.S. for bench top power ) I thought it would be fun to include some square and sine wave signal sources ( 1Khz, 10khz, 100Khz, 1Mhz) using 555 timers which I have more than a few of. These 555 timers ring like the bells of St. Marys. My question is about the 100 Khz design.
After some experimentation, I've got an acceptable square wave output from the 2N3904 emitter follower. Adding C2 (top of page) connected directly to pin 8 (555) helped a lot. Also the 9.1v zener in series with the 1N4148 helped. I still see some ringing on the sine wave output but various attempts at filtering the sin wave are not helping.
I was wondering if anyone has a suggestion for how to make the sine wave cleaner. The ringing seems to be in the 1-10 Mhz range ( near as I can tell on my trusty old B&K 30Mhz scope ). Maybe a series tuned circuit across the output of the 2N2222 ? I don't have a lot of inductors to experiment with. My technician level education is bit old (1970's) maybe something new is out there. Thanks
 

Attachments

MrChips

Joined Oct 2, 2009
30,720
Many times, ringing is not at the signal source but at the measuring device, i.e. at the oscilloscope.
Hence, before we start pointing fingers it is wise to characterise your test equipment.

What is the make and model of your oscilloscope?
What are you using for an oscilloscope probe? You should be using a proper probe with x10 attenuation.
Is the probe properly compensated for frequency?
What are you using as a ground on the oscilloscope probe? You ought to be using the grounding clip that is attached at the tip of the probe.
 

AnalogKid

Joined Aug 1, 2013
10,989
The output edge of a 555 is more than fast enough to cause a scope probe to ring like a resonant circuit. This is because the ground lead with the clip on it, usually something like 4"-6", is an inductor.

Also, what is the purpose of having a hard diode clipper on the 555 output? All this does is make the chip run hotter.

Make sure C2 has the shortest possible leads and is connected close to U1 pins 1 and 8. Add a 10 uF electrolytic cap in parallel for better performance.

Please add reference designators to all components on the schematic.

The 330 ohm resistor in the upper right corner (see how easy it would be to say "R4") seems a bit low. Consider scaling all resistors up by a factor of 10.

The standard astable circuit you are using is very asymmetrical. This increases the energy in the higher order harmonics, making it more difficult to filter off a clean square wave. On the CMOS LMC555 datasheet is a more simple variation that comes close to a 50/50 duty cycle and holds it regardless of frequency.

ak
 

Audioguru again

Joined Oct 21, 2019
6,674
The very old original 555 produces a 400mA shoot-through current each time its output switches which overloads the power supply wiring causing ringing. The modern Cmos 555's (LMC555, TLC555 and ICM7555 do not produce the shoot-through current and are rail-to-rail producing a proper completely square wave.

Your filter is not an active filter. It has a few simple filters in series so the response is very droopy. I made a perfect low distortion sinewave from a squarewave using a "switched-capacitor Butterworth lowpass filter IC" (Maxim make some). I used a compact layout soldered on a stripboard, may you are using a Mickey Mouse solderless breadboard that increases your ringing.
 

Thread Starter

Sam1951

Joined Nov 4, 2019
4
Thanks all for so many quick responses. I hope to address all points made in this post.
Scope model ..... B&K 1479b (30 Mhz)
Probe .................... PR-40 (set on x10)
I re-checked the calibration with the 1 Khz source provided on the scope (they were very close to optimum)
I un-floated the scope ground. It was common practice to float the scope ground all the time at work. Being ever
mindful of possible hurtful voltages present on the scope chassis depending on where you hook the ground lead.
However, it could also ruin your day to blow up some expensive stuff, perhaps take a few hours to fix it and even need a
circuit board overnight shipped thus requiring explanations to bosses.

I took out the 9.1v zener & 1N4148, I had hoped to clip 555 pin_3 @ ~ 9.8v but it affected the rise time on Pot_3 some.
I left R_3 & C_4 across 555 pin_3 as they helped the output on Pot_3 look better. I hooked the probes directly to or with a
very short straight wire to the point measured. It all made a little difference.

I suppose I should have known this but I was surprised to see that such a big improvement in the C_1 signal was obtained by
grounding that probe also. Only grounding 1 probe is not enough, even at only 100 Khz. Ooops, lesson learned, re-learned.

I've include some pictures. The square wave output (pot P_3) looks rather good now, (as long as I don't buy a new scope) :)
Not badly asymmetrical and limited ringing (see picture) . I still see a little on the sine wave output though. I wonder to what
extent it's actually there or more a product of still imperfect grounding? Anyway if it is really there (10-20 Mhz ?) I found a .76
to 1.25 uH coil that I've long since forgotten the original purpose of. It would be cool if I could use that. Across C_11 (series)
or after C12 (parallel). I'll have to carefully grind the mounting leads so it will plug into my proto-board, it seems worth a try.

As you may have guessed, design decisions are made with what parts I can scrounge as a major consideration. Thanks again to all for taking the time to respond.

P.S. If I was going to buy a new scope, possibly with storage, 70 -100 Mhz. And willing to spend $1000 ( about what a descent
set of irons might cost). Any thoughts? So much has changed since I bought that B&K.Circuit.jpgPS_Project.jpgSine-wave_and_C1.jpgSquare-wave_and_C1.jpgEPSON002.JPGCircuit.jpgPS_Project.jpgSine-wave_and_C1.jpgSquare-wave_and_C1.jpgEPSON002.JPG
 

MrChips

Joined Oct 2, 2009
30,720
I don't see any "ringing" on the sinewave. What I see is "cross-talk".

Ringing occurs at the rising or falling transitions of the signal. Here are some scope examples of ringing.
The time scale on the first picture is 50ns/DIV.
The time scale on the second picture is 100μs/DIV.

1576371609607.png

μ
1576371645349.png


What you are seeing on your sinewave is cross-talk coming from the transition of the 555-timer signal.

There are many things in your set up that is making this worse than it needs to be.
Here is what you need to do.

1) Replace the LM555 with a CMOS version such as LMC555 or TLC555.

2) Add 0.1μF ceramic cap across your power supply rails at the 555-timer chip. Add 10μF electrolytic capacitor across the power supply rails.

3) Keep all your wiring short.

4) Get rid of the prototyping breadboard and do a proper PCB construction.
 

Thread Starter

Sam1951

Joined Nov 4, 2019
4
I already had a .1 uf on pin 8 of the timer and large electrolytic where power comes on the board. Those 2 things made a big difference in signal quality.
Oh yeah, of course. All that protoboard wiring is like little antennas. No way 20 Mhz is making it's way through the sine wave filter riding along with the .1 Mhz signal. It's getting broadcast to everywhere else in the circuit by all those little antennas.
Somehow I thought LM555CN was CMOS, it happens I have an LMC555. I just now tried it and that did the trick. The LMC555 is not generating all that "junk" to begin with. The signals look clean so now I can build....... Thanks
 

MrChips

Joined Oct 2, 2009
30,720
btw I notice on your breadboard layout a capacitor connected to pin-8 of the 555-timer IC via a nicely formed purple jumper.
Make sure that on your final construction that you do not do that.

Any 0.1μF decoupling capacitor must go from pin to pin of the IC, keeping the leads of the capacitor as short as possible. This is where a SMT circuit board layout is superior to thru-hole components.
 

ci139

Joined Jul 11, 2016
1,898
you may isolate 555 with (usually 1 to 100) ohm and no (or lesser) supressor cap to attenuate pull up - but makes the entire thing slower . . . or use output filter that makes only the output slower . . . _____________ . . . how long is your output line anyway

.... my bad, it's C1 ringing . . . that complicates the setup . . . ← the only thing that works → is above 300nH to 1.3µH parasitic series inductance to C1 (the worse is neg. peak)

https://www.edn.com/electronics-blo...ing-wire---loop-inductance--Rule-of-Thumb--15
https://www.edn.com/electronics-blo...ductance-in-the-return-path--Rule-of-Thumb--7
 
Last edited:

crutschow

Joined Mar 14, 2008
34,285
Only grounding 1 probe is not enough, even at only 100 Khz.
For a digital signal, it's not the frequency of the waveform that generates ringing, it's the rise and fall time durations.
The equivalent frequency of that is approximately 0.35/(rise/fall time) for the 10%-90% points on the rise/fall time.

Thus for the nominal 100ns rise/fall time of the 555 output, the equivalent edge frequency is 0.35/100ns = 3.5MHz.
That's also the reason the 555 needs to be decoupled directly across the power and ground pins with a 100nF ceramic capacitor.
 

Thread Starter

Sam1951

Joined Nov 4, 2019
4
Thanks to all for the help. The outputs look quite good now. I even tried removing C2 and they still look OK but will leave it in as decoupling caps are usually a good practice. I also removed R3,C4 and clipping diodes as unnecessary. I think I'll leave C11 in, removing it had minimal effect on the signal but might attenuate stuff I can't see on my old B&K. Will also measure [P7-R7] & [P2-R4] and replace with close fixed R's and try a couple of different Q1's and Q2's and make P1 a 20 turn. Will also leave P4,5,6 in as I have so many of them.
I don't know why I was thinking LM555CN was CMOS, changing to LMC555 did more to make the signals look good than any other one thing. Many times in the past, seemingly tough problems were caused by overlooking something simple. Same with Scope grounding, The nightmare scenario was forgetting that when looking at 2 signals, whatever 2 places you connect those ground leads are now connected together and to ground through the scope. You don't want to be the technician that causes some high profile customer to not get his elevator on time because you broke it while trying to fix something. Thus, float the scope ground and use only 1 ground lead. Confusing scope pictures could result, but at least you won't be the one who made it worse by breaking something else.
The soldering iron should be hot now, I think I have something worth building.
 

Attachments

Top