Go Down

Topic: Problem when expanding on the ShifOut tutorial for LEDs (Read 8 times) previous topic - next topic

Nick Gammon

If you are not supplying +5V to the Vcc of the shift registers there is something seriously, seriously wrong. Solve that first.

David82

#56
May 07, 2012, 03:07 am Last Edit: May 07, 2012, 03:43 am by David82 Reason: 1
It's getting the 5v it needs from pins 10/11/13. It does work. Just has a tiny bit of noise issues and only works under a bizarre configuration as described earlier. In the diagram on your site there is a 10k resistor going from SS <D10> to ground.
1. What's that for and is it really needed? Could it help with the noise?

2. I noticed all of you caps are set diagonally. Is there reason behind that?

3. My caps looks different too. They are the brown (ceramic disc). Does that matter?

iyahdub

He might be right, Nick, if it was just a few leds. If it aint many, ive seen the mux'ers pass enough to light them... But what it is strange is the fact that he is talking about 80 plus leds ?!?!?  lol That TV analogy exactly what id used.
10 LET Loop=Infinite
20 GO TO 10

Grumpy_Mike

Quote
It's getting the 5v it needs from pins 10/11/13. It does work.

No it is not.

This is called paricitic powering and is a good way to destroy chips. Unless you connect all the chips to the 5V line you don't stand a hope in hell of getting anything working. Chips are made to take power from their 5V input and ground.
As you try and scale up you get two problems, first you get noise on the supply lines that distributing decoupling capacitors can help with. Second the signals that go to all the chips, like the clock signal, are not strong enough to power all the chips. Using two inverting buffers back to back will give you one non inverting buffer. You should use the buffer output to drive no more than 10 chips.
Finally make sure your power supply can handle the current for all those LEDs.

Remember you are asking for help, however you seem to be rejecting most of it. Think about it, is that a smart move?

Nick Gammon


It's getting the 5v it needs from pins 10/11/13. It does work. Just has a tiny bit of noise issues and only works under a bizarre configuration as described earlier.


You have a fault in that case. You need to rectify it. The fact that power is somehow working its way into the logic lines and powering the chips would account for any number of problems.


Quote
In the diagram on your site there is a 10k resistor going from SS <D10> to ground.
1. What's that for and is it really needed? Could it help with the noise?


It's to keep SS low during the Arduino power-up sequence (it should probably be a pull-up but whatever). Its sole purpose is to stop some random pattern appearing before the first sequence is sent out by the processor.


Quote
2. I noticed all of you caps are set diagonally. Is there reason behind that?


Purely because of how far apart the leads were.

Quote
3. My caps looks different too. They are the brown (ceramic disc). Does that matter?


No, the type of cap shouldn't matter.

Go Up