I just happened to read your post of earlier today on another topic:
http://forum.arduino.cc/index.php?topic=169062.msg1257882#msg1257882
where you state (shout): ...no cap on ANY control line...
I have just made a little device in which I drive a 2-digit 7-segment led display by means of an ATtiny85 via 2 shift registers. The 7-segment displays are common anode so that when the SR pins are low the corresponding display elements are on. Initially, on switching on, all the elements of the leds were on which I found annoying. So, in my ignorance, I connected a 0.1µF cap between Vcc and pin 12 of the SR and found that that did the trick in clearing the display upon start-up. Is that wrong and, if it is, how should it be done? A pull-up resistor instead of the cap?
Are you bypassing each shift register and if not, Why? what you are doing is bypassing the power supply and creating speed issues with the R clock. A wire of any length is two things, first a resistor and second an inductor. The resistance will decouple (negate) any decoupling on the power feeding the IC and second the inductance of the wire is a bigger issue because the final term (reactance) is frequency dependent, Not independent. Inductive reactance rises with frequency. There is an illustration that shows a bypass on one of the control lines. The drawing is WRONG and for whatever reason hasn't been changed by the Arduino Gods... Neither the lack of bypasses or the "extra" cap although I am sure they are well aware of it.
Proper engineering REQUIRES a bypass on each active device... Or so have I been taught and my training is reinforced by the times I thought it wouldn't be necessary.
I was ALWAYS wrong and I've fixed my mistakes... Sometimes 100 boards at a time. I did this for a living for 20 years, You?
Is that wrong
Yes. You are getting an effect known as parasitic decoupling. You are driving an arduino pin directly into a capacitor and a largeish one at that. When a capacitor is discharged it looks like a short circuit causing a damaging amount of current to flow from the arduino pin for a short time.
The capacitor should be between the power and ground of the chip. This is called decoupling:-
http://www.thebox.myzen.co.uk/Tutorial/De-coupling.html
The cap needs to go from Vcc to Gnd. Not on any control line.
The arduino IO pins default to Inputs after a reset, and stay that way until the sketch starts to control them.
I would add a 10k pulldown resistor to hold them low during the reset period.
Yes, I do understand the purpose of a decoupling cap between Vcc and ground.
But here is shown a cap connected between pin 12 (latch pin) and ground:
@ CrossRoads: Does that violate your rule, or do I not understand what you mean by ‘conrol line’?
Does that violate your rule, or do I not understand what you mean by ‘conrol line’?
YES it does, that tutorial is WRONG as has been pointed out to you more than once before.
A control line is an output line that controls the shift register.
We've requested the tutorial be changed numerous times to no effect.
As was also pointed out to him