Unlimited led sequential blinking circuit

Status
Not open for further replies.

Ian0

Joined Aug 7, 2020
5,114
What happens after phase N?
You need a shift register such as 74HC164. Hold D high and apply signal to clock.
 

BobTPH

Joined Jun 5, 2013
4,899
All you need s a 6 pin micro and 1m WS2812B led strip.

@Audioguru again: I have 30 m of WS2815B LED strips (1800 rgb leds) in my ballroom and I haven’t gone blind yet. With all of them on bright white, it lights the room at a moderate brightness.

Bob
 

crutschow

Joined Mar 14, 2008
29,785
Below is the LTspice simulation of a partial sequential circuit using 74HC164 shift-registers, which will do 8 LEDs per register.
Only two of the 7 registers to do 56 LEDs are shown.
A low MR signal resets all the outputs to zero until MR goes high to restart the sequence.

Each output can drive about 4mA of LED current.
More than that will require a driver transistor at the outputs for each LED.

1651113015338.png
 

ElectricSpidey

Joined Dec 2, 2017
2,146
My only criticism of using addressable LEDs would be the limited spacing between the LEDs if strips are used and the huge amount of work if not using strips to create other layout options.

Using the shift registers allows using a much more varied type of LED and layout configurations.

But if all you need to do is light up one LED at a time in a straight line with little concern with the spacing, then addressable LED strips are the way to go.

I'm always amused by the less chips is better option, when there are many other considerations.
 

BobTPH

Joined Jun 5, 2013
4,899
If these are to be spaced widely, the wiring is horrendous for 60 individual LEDs. The strips requite only 3 wires.

And, as far as spacing goes, WS2812Bs also come on tiny individual boards that you daisy chain with 3 wires. That is how the lights on my wife’s dress in my avatar picture are wired.

Bob
 

Ya’akov

Joined Jan 27, 2019
5,659
Sure 10 chips is always s better than one , even when the 1 chip can down so much more than the 10.

Bob
Well… to be more accurate (though not to disagree) number of LEDS + 1 chips using a serial bus, since each LED has it’s own on-board IC.

Also, as @BobTPH says it would be much better to either cut up a WS8212B strip (which is relatively trivial), or to use the readily available single LED breakout boards which a dirt cheap in 100 quantities, than to try to wire up 60 LEDs individually. I am not even sure how you’d practically manage that if the LEDs are spread over even a modest area, but if they are distant (on the order of meters) it is even more fraught.
 

AnalogKid

Joined Aug 1, 2013
10,054
I am not even sure how you’d practically manage that if the LEDs are spread over even a modest area,
It is only 1/8th as bad as it might seem at first. If each SR is on its own small board and drives the 8 LEDs nearest it, those are home runs back to the nearest SR board, not all the way back to the power source. If there is no need for the Reset signal when the system is up and running, then each SR board has a local R-C power-on reset. That leaves only four wires that run between SR boards. Three of them are in parallel (common to all boards) - Vcc, GND, Clk; the enable signal input is daisy-chained from the previous SR board, and the enable output to the next board is driven by Q7.

With multiple, daisy-chained SR boards, the common VCC and GND wires carry 100% of all LED current and will have to be heavier than the clock line needs to be. This system is relatively easy to expand in multiples of 8 LEDs.

Lotsa wires, agreed, but not nearly as cumbersome as "more than 50" home runs back to a single controller. Although ... if the LED locations do not have to be adjustable, ribbon cable would make quick work of managing 50 home runs back to a single power source/controller. Just sayin' . . .

TS: Please give us more details about the application and the physical layout of the LED locations.

ak
 
Last edited:

Ya’akov

Joined Jan 27, 2019
5,659
It is only 1/8th as bad as it might seem at first. If each SR is on its own small board and drives the 8 LEDs nearest it, those are home runs back to the nearest SR board, not all the way back to the power source. If there is no need for the Reset signal when the system is up and running, then each SR board has a local R-C power-on reset. That leaves only four wires that run between SR boards. Three of them are in parallel (common to all boards) - Vcc, GND, Clk; the enable signal input is daisy-chained from the previous SR board, and the enable output to the next board is driven by Q7.

With multiple, daisy-chained SR boards, the common VCC and GND wires carry 100% of all LED current and will have to be heavier than the clock line needs to be.

Lotsa wires, agreed, but not nearly as cumbersome as "more than 50" home runs back to a single controller. Although ... if the LED locations do not have to be adjustable, ribbon cable would make quick work of managing 50 home runs back to a single power source. Just sayin' . . .

ak
I still find it hard to justify discrete logic when the alternative solution is so simple, cheap, and flexible. It gives you arbitrary addressability, RGB arbitrary color, and microprocessor control for free in the bargain.

No doubt your scheme is far better than a naive home run approach but I sincerely believe it has been superseded by the newer technologies.
 

AnalogKid

Joined Aug 1, 2013
10,054
I sincerely believe it has been superseded by the newer technologies.
I do not disagree at all. BUT, to me, this:

Does anyone have an idea about how I can realize it without using arduino?
eliminates all programmable devices and methods. Step #1 in forums: answer the question that was asked. A shift register solution will be way better than any variation of the "stacked transistor" approach he links to. For me, that is good enough until we get more details.

Also, (metaphor alert), getting to Cleveland by car is
simple, cheap, and flexible
***only*** if you already own a car, know how to drive, have a license, and can read a map. If you are doing all of that from scratch, it will take weeks. Still, that's less time than it would take to muddle through your very first ever programming project.

ak
 
Last edited:

MrSalts

Joined Apr 2, 2020
1,762
How can an Arduino/microcontroller-based solution be "better" if it doesn't meet the specification?

Does anyone have an idea about how I can realize it without using arduino?
For example...

All you need s a 6 pin micro and 1m WS2812B led strip.

@Audioguru again: I have 30 m of WS2815B LED strips (1800 rgb leds) in my ballroom and I haven’t gone blind yet. With all of them on bright white, it lights the room at a moderate brightness.

Bob
addressable LED strips are the way to go.
Splice the Main Brace
Well… to be more accurate (though not to disagree) number of LEDS + 1 chips using a serial bus, since each LED has it’s own on-board IC.

Also, as @BobTPH says it would be much better to either cut up a WS8212B strip (which is relatively trivial), or to use the readily available single LED breakout boards which a dirt cheap in 100 quantities, than to try to wire up 60 LEDs individually. I am not even sure how you’d practically manage that if the LEDs are spread over even a modest area, but if they are distant (on the order of meters) it is even more fraught.
I still find it hard to justify discrete logic when the alternative solution is so simple, cheap, and flexible. It gives you arbitrary addressability, RGB arbitrary color, and microprocessor control for free in the bargain.

No doubt your scheme is far better than a naive home run approach but I sincerely believe it has been superseded by the newer technologies.
Now, you've all proposed addressable LEDs. Finish your answers - show us how to control those addressable LEDs without an Arduino /microcontroller.

The OP probably wouldn't have even asked for help if an Arduino was (allowed, possible, interesting) to him. Ignoring the hard part ALWAYS makes an answer easy.
You are not bringing a revolutionary idea to the conversation by suggesting an Arduino - spoiler - the OP already considered arduino/microcontroller and dismissed it (it was specifically excluded from his design specification).
 
Last edited:

Ya’akov

Joined Jan 27, 2019
5,659
I do not disagree at all. BUT, to me, this:



eliminates all programmable devices and methods. Step #1 in forums: answer the question that was asked. A shift register solution will be way better than any variation of the "stacked transistor" approach he links to. For me, that is good enough until we get more details.

Also, (metaphor alert), getting to Cleveland by car is

***only*** if you already own a car, know how to drive, have a license, and can read a map. If you are doing all of that from scratch, it will take weeks. Still, that's less time than it would take to muddle through your very first ever programming project.

ak
This seems a perpetual debate. I will just add a point and leave it at that:

Many times the TS is no more skilled at executing the discrete solution than the MCU solution so the learning curve question is different. It’s not this case I don’t own a car, but I do own a motorcycle (for example).

It is the case that a motorcycle or car will have to be purchased and the person will have to learn how to drive one of them. Is one more easily learned than the other?

In our case I don’t believe that it is commonly the case people are more qualified to use the discrete solution than the MCU. I think they are mistakenly intimidated by the MCU and imagine discrete is easier; and often in their ignorance they think “Arduino” means a big expensive development board while means a few simple components.

As you know with chips like the ATTiny which is easily programmed in the Arduino IDE a single 8 pin chip replaces a lot of discrete logic, and no evaluation board is involved. It also allows quick revisions with no rewiring So it comes down to why the TS specifies “no Arduino or MCU”, it may be from ignorance not for a valid reason.

Of course if they “just want to”, that’s their business
 
Status
Not open for further replies.
Top