Issues with 556 on 12V

I started with an original circuit containing a 555 in astable mode driving a pair of 4015s and a 4017. These in turn were used to switch on 15 and 4 LEDs, respectively. This circuit ran fine between 9 and 12 volts, from both batteries and a power supply.

The values used on the 555 was:
C1=1uF electrolytic
R1=47k 1/4 Watt
R2=100k trimpot

The next iteration was to use a 556 so the 4015s and 4017 could be driven at independent rates. I did some 555 math to tweak the timing of the chips to allow a different "center" and the results for the 556 are as follows:

Driving the 4015s:
C1=1uF electrolytic
R1=22k 1/4 Watt
R2=50k trimpot

Driving the 4017:
C1=2.2uF electrolytic
R1=100k 1/4 Watt
R2=50k trimpot

This 556 is also being run in astable mode. The issue that I have is that while this circuit runs beautifully at 9V, when powered with 12V, it proceeds to light the LEDs in an erratic pattern. I've found some mention of higher voltages and edge trigger and such, but nothing I could figure out.

I'm not sure if the caps need to be a different value, or if the resistors are wonky or what. I've tried slightly different values for each, and it's not changing anything. It's just really strange to me that (V would work perfectly, and 12V messes up, and all that really changed was going from a 555 to 556.

Any help would be appreciated, since I'm getting frustrated.

555 chips create nasty spikes on the supply lines. Make sure you have 0.1μF and 10μF decoupling capacitors across Vcc and GND.

I'm guessing you mean the VCC and GND pins on the 556? It seems like the same idea as decoupling a long length of wire, so does the value matter?

I soldered up a board yesterday based on the attached schematic and proceeded to test it. Once again, it runs fine at 9V, but when 12V is applied, the 556 (astable) seems to get noisy, and the LEDs driven off the 4017 are lit erratically.

Some explanation of the circuit:

The 4015/4017 are driving LEDs through 2N3904s, but I left them out for clarity.

JP1 in the lower right is a 4 pin screw terminal relatively close to the 556. Pin 1 is GND, Pin 2 is + from the battery. Pins 3 and 4 go to a SPST with an optional decoupling cap (tantalum) if the switch is on long wires. All this works as intended from testing.

The capacitors on the 556 are both electrolytic.

So in testing to reduce the noise issues at 12V, here's what I found:

1) A 0.1uF polyester film capacitor from Radio Shack put across Pins 1 and 2 on JP1 (so across the battery + and -) seems to eliminate any noise issue. This can run for hours without any problems at all.

2) A 0.1uF tantalum (taking into account polarity) seem to also work in the same location. Point of note, we only tested this for a few seconds.

3) Since the above seemed to work, we then placed the 0.1uF tantalum across the VCC and GND pins of the 556 itself, respecting polarity. This also seemed successful, but then the cap proceeded to blow the side off itself. This may have occurred in scenario 2 if we left the circuit run longer, but it wasn't something we wanted to try again.

For the time being, repeating option 1 for 12V operation isn't a huge deal, but ultimately, I'd prefer a more permanent solution. Any higher value of cap seems to induce its own sort of weirdness, so 0.1uF seems to be the value that works.

My questions are:

Should I stay away from tantalum caps for eliminating the 556 noise and use a different type, or did I do something wrong with the tantalum? It's hard to not want to use them because of the small form factor.

Is there a known cause for why scenario 3 occurred? Is there some backwards current or strange power issue on the 556 VCC/GND path?

I have an aerospace degree, but my practical EE training was more or less non-existent, so while I know 555 circuits are pretty simple/basic, I lack a lot of the tribal knowledge, so please, bear with me.

Here are my thoughts.

I don't think the voltage difference between 9 and 12 is the problem at all; I think the problem is the source of the voltage. The 12 V supply is noisier than the 9, which means...

1. I would go back to the 12 V supply and add some electrolytic caps in the appropriate places depending upon what kind of power supply it is. If, by chance, it uses a LM7812 regulator or similar, I would have a 1000 μF electrolytic and a .33 μF ceramic on the input, and a .1 μF ceramic and a 10 μF electrolytic on the output.

2. I would have a .1 μF ceramic across the power pins of every IC in the circuit. They should be as physically close to the power pins as is practical.

3. I would add a .01 μF ceramic cap from pin 3 of the 556 to ground.

4. I don't know that the tantalum caps are part of the problem, but I see no reason to use anything but ceramics. Those that are 35-50 V rated are plenty small physically; the leads fit nicely on .1" to .2" centers and the bodies are proportionately small.

FWIW


ETA: I forgot that the 556 has two control voltage inputs. As the OP has discovered, they both should be bypassed to ground. Although he used a .1 uF cap, the datasheet calls for a .01 uF.

Ceramics have a lower parasitic inductance and ESR then tantalums so ceramics are better at filtering high frequency noise. That's why they are always used for decoupling directly at the chip and why a small ceramic cap is often used in parallel with a larger electrolytic cap for filtering.

Check the datasheet of your 556 timer. The off state output voltage climbs with supply voltage and output amperage.

Driving the system at 12 volts is likely preventing the off state voltage from falling below the 0.6v threshold of your 2N3904 transistors - they will always be on.

Use a higher value base resistor (see datasheet of 556 - specific to your model and brand). TI claims an OFF state voltage at 15 volt input at 100 mA to be 2.5 volts on NE556 chip!

http://www.google.com/url?sa=t&rct=j&q=ne556&source=web&cd=1&cad=rja&ved=0CDIQFjAA&url=http%3A%2F%2Fwww.ti.com%2Flit%2Fds%2Fsymlink%2Fsa556.pdf&ei=c8hMUZ7PFMrS0gHNuoCQCw&usg=AFQjCNElkl0A6edyEjMZ_PZSVA1FLYZsFQ&bvm=bv.44158598,d.dmQ

After some putzing around, it seems like a pair of .1uF ceramics from pins 3 and 11 to GND fixes the issue nicely. On the next batch of boards I make, I'll probably include a pair of pads to add the polyester film cap on the power to hedge my bets.