I have an AC-powered circuit which flashes a series of LEDs in seemingly random sequences (not actually random). I'd like to replace all of it with a battery-powered circuit.

I don't want to use a flip flop circuit, and I also don't want to use the popular "simplest LED flasher" circuit which omits the base pin of a transistor—because many comments on these circuits say that using a transistor this way damages it. 555 timers and microcontrollers seem like overkill—I want to keep it as simple as possible.

EDIT: And self-blinking LEDs all have the same frequency; I want to be able to control the frequency myself.

Here is the AC circuit:

Obviously I need to ditch the rectifier and find new pot, resistor and capacitor values.

But to replace the DIAC I'm a little lost. My first thought was to use a transistor with a capacitor triggering the base, but I can't figure out a way to do that. Another thought was a MOSFET. Does anyone have any suggestions? Thanks in advance!

 

Since you have a long list of parts you don't want to use, have you considered self-flashing LEDs? Doesn't get any simpler than this.

 

Since you have a long list of parts you don't want to use, have you considered self-flashing LEDs? Doesn't get any simpler than this.

When I first saw them I thought they'd solve the problem, but these LEDs have a set frequency which can't be changed: "Frequency: 1.5 Hz (90-96 times/minute)"

While I could try staggering them, it's more important to me that the frequency is different for each LED.

 

My first thought was to use a transistor with a capacitor triggering the base, but I can't figure out a way to do that.

Neither a bipolar transistor nor a FET are "triggered". They both are at heart an analog, quasi-linear amplifier. In circuits with more components they can have bi-stable action, acting as a switch that can be triggered.

The DIAC diode in your schematic is a non-linear device that changes from a virtual open circuit at low voltages to a quasi-short-circuit at higher voltages. It is the same function as the two-terminal transistor circuit you already have ruled out. What is wrong with using it?

Another option is a unijunction transistor (UJT). This again snaps from a non-conducting state to a conducting state when the voltage across it exceeds a certain value. It is essentially the one-transistor oscillator circuit approach, but with a device grown specifically for it.

AND - what is wrong with "a flip flop circuit". In electronics that term has many implementations and variations. Which one are you referring to? Or are you talking about a two-transistor multivibrator circuit?

Another option is a Schmitt trigger oscillator. You get 4 of them in a CD4093 quad NAND gate package, or 6 of them in a CD40106 hex inverter package. Each oscillator is individually tunable.

https://2n3904blog.com/cd40106-schmitt-trigger-relaxation-oscillator/

ak

 

Years ago I tried to use a small SCR (like 1A) in this circuit instead of diac and it worked. But the problem was if the cap is charged very slowly the SCR wasn’t triggered.
Maybe you can try replace diac with two transistors connected as SCR and get a better success.

 

But to replace the DIAC I'm a little lost. My first thought was to use a transistor with a capacitor triggering the base, but I can't figure out a way to do that. Another thought was a MOSFET. Does anyone have any suggestions? Thanks in advance!

A unijunction transistor flasher is very similar to that diac circuit. Not sure about availability though.

How many LEDs? Don’t know why you say a microcontroller is overkill, it would be the simplest, lowest parts count, and cheapest circuit by far, and you could use a pseudo random number generator to make a very long sequence generator looking far more random than what you are likely to come up with any other way.

 

Maybe you can try replace diac with two transistors connected as SCR and get a better success.

Good point. If you don't like Dick Cappel's 1-transistor circuit, you can synthesize it with two transistors acting as a breakover switch. In this way, you get the same effect as the 1-transistor circuit, but with both transistors operating "normally".

ak

 

It is the same function as the two-terminal transistor circuit you already have ruled out. What is wrong with using it?

As I stated above there are comments on different videos and postings of this circuit (and I cannot remember the exact wording so I will probably get this wrong) which state that running power through the transformer this way—taking advantage of the breakover or break down point—will damage the transform over time. Again, I probably didn't explain it exactly right, and today I cannot find any of the comments I read last night about this.

If using the transistor this way does not actually cause damage, this seems a reasonable solution.

AND - what is wrong with "a flip flop circuit". In electronics that term has many implementations and variations. Which one are you referring to?

Most of the LED flash circuits online are either Dick Cappel's no-base transistor version or two LEDs alternating with a flip flop circuit. If a flip-flop can be made with a single LED, it could be an option. But I'm hoping there's a simpler design similar to the one in the original high-voltage circuit.

How many LEDs? Don’t know why you say a microcontroller is overkill, it would be the simplest, lowest parts count, and cheapest circuit by far, and you could use a pseudo random number generator to make a very long sequence generator looking far more random than what you are likely to come up with any other way.

There are five LEDs. Perhaps I should have used a different word than random; the LEDs flash in a regular pattern EDIT: at a regular frequency, although precision isn't necessary. (EDIT: It's not important to replicate the frequency of the original circuit.) Explaining what this outdated obscure device does would fill an entire thread with even more exasperated replies!

While a microcontroller could easily get the job done, to me programming one and wasting its potential on such a small project seems overkill. "It's synonymous to using a table saw to cut a piece a paper when a standard pair of scissors will do the job just fine." (Dustin Christopherson) To me it feels like putting a microcontroller between a light switch and the bulb to turn it on. (In this case I suppose we're removing the switch entirely.) I guess it's really a matter of preferences and I'm stuck in the old days.

Thanks everyone for the suggestions in this thread, when I get a moment I guarantee I'll research all of them!

 

What is the pattern? And by that I mean the on and off times of the 5 LEDs, if two or more are on at the same time, etc. Physical location is totally irrelevant to the circuit design. 5 LEDs in a sequence pattern certainly is doable. Shift register, Johnson counter; and if you like analog stuff, a linear ramp driving five voltage comparators.

To be clear and unambiguous, consider labeling the LEDS 1 through 5, and making a table of which ones are on and off at each step through the pattern. Note that any time all of the LEDs are either on or off counts as a step.

You mentioned an adjustment. Is that one adjustment for the speed of the overall pattern; five adjustments, one for each LED; or other?

ak

 

What is the pattern? And by that I mean the on and off times of the 5 LEDs, if two or more are on at the same time, etc.

The frame rate loses some of the flashes, but each LED has a different but repeating pattern EDIT: frequency. Like I said accuracy isn't important, but I'd like to be able to adjust each one as needed.

You mentioned an adjustment. Is that one adjustment for the speed of the overall pattern; five adjustments, one for each LED; or other?

There's a potentiometer which adjusts overall flash frequency ("Master Rate" in the schematic), and each LED has its own resistor which is added to the Master Rate to control its individual flash frequency. Again, the frequency isn't important.

 

Does anyone know if it's true that this circuit by Dick Cappels will damage the transistor over a period of time?
http://cappels.org/dproj/simplest_LED_flasher/Simplest_LED_Flasher_Circuit.html

 

Neither a bipolar transistor nor a FET are "triggered". They both are at heart an analog, quasi-linear amplifier. In circuits with more components they can have bi-stable action, acting as a switch that can be triggered.

The DIAC diode in your schematic is a non-linear device that changes from a virtual open circuit at low voltages to a quasi-short-circuit at higher voltages. It is the same function as the two-terminal transistor circuit you already have ruled out. What is wrong with using it?

Another option is a unijunction transistor (UJT). This again snaps from a non-conducting state to a conducting state when the voltage across it exceeds a certain value. It is essentially the one-transistor oscillator circuit approach, but with a device grown specifically for it.

AND - what is wrong with "a flip flop circuit". In electronics that term has many implementations and variations. Which one are you referring to? Or are you talking about a two-transistor multivibrator circuit?

Another option is a Schmitt trigger oscillator. You get 4 of them in a CD4093 quad NAND gate package, or 6 of them in a CD40106 hex inverter package. Each oscillator is individually tunable.

https://2n3904blog.com/cd40106-schmitt-trigger-relaxation-oscillator/

ak

YES!! and with the CD40106 IC you get six per IC. and then you can have that group drive a CD4049 Hex inverter and it can drive six LEDs. Not sure you can do better than that.
But then we learn that it must not be truly random, but some specific pattern So now I suggest using an 8 bit prom, and load in the desired sequence for up to eight transistors that can each control one or more LEDs.

 

You claim it must be a repeating pattern, and you also claim the circuit you are replacing had separate unsynchronized oscillators. That would not create a repeating pattern. It might look like that short term, but the pattern will change over longer periods.

 

While a microcontroller could easily get the job done, to me programming one and wasting its potential on such a small project seems overkill.

I am having a little trouble understanding that reasoning. Why does it matter that the micro could be doing something else?

If micros were only used at their full potential, practically nothing would use them. Take an IR remote for example. It sits there in sleep mode 99.99% of the time, then comes alive for a fraction of a second when someone presses a button. What a waste! But I wouldn’t think of doing it any other way.

Parts list for your 5 LED sequencer:

1 8 pin PIC
1 capacitor
5 LEDs and 5 resistors (needed for any solution)

Perhaps you have no experience with micros. or programming. In that case, the learning curve would be a large negative and I would understand.

 

Unijunction transistors are still available but at outrageous prices.
However, PUT devices 2N6027 and 2N6028 are still available at reasonable prices. And with a single transistor per LED, it doesn’t get any simpler.

 

OK, based on the video ...

First, all of your descriptions of what you want are pretty poor.

It looks like whenever an LED blinks, alone or with others, all LEDs blink for the same number of milliseconds. Also, it looks like the blink pattern is being clocked by a single frequency. That is, no matter if 1, 2, or 3 LEDs are on at the same time, they all go on and off together. And the times when all LEDs are off also fit in the clock frequency.

I've seen the video only twice, once an normal speed and once at 0.25 speed, but I can't see any repeating pattern. If what you want is a clocked 5-bit random number generator, a uC should be part of the discussion.

Please try again to describe what you want.

ak

 

You claim it must be a repeating pattern, and you also claim the circuit you are replacing had separate unsynchronized oscillators. That would not create a repeating pattern. It might look like that short term, but the pattern will change over longer periods.

I don't really "claim" anything; I've never even heard of unsynchronized oscillators As I stated above the overall flash rate isn't important, I simply said that each LED flashes repeatedly at the same interval (I used the word "pattern" but should have said frequency). I didn't bother to explain that the interval is determined by the resistor labeled "RATE" and that each LED has a different value of "RATE" resistor because it wasn't important to my original question of replacing the DIAC.

I'm just looking for help in replacing the DIAC with a low-voltage alternative, using discrete components (if that's the correct term). My circuit won't flash exactly the same, but as I said that's not important.

I am having a little trouble understanding that reasoning. Why does it matter that the micro could be doing something else?

It doesn't. I fully understand that a microcontroller is a fine device to use for flashing five LEDs in a random or a rigid pattern. But I just don't want to use and program a microcontroller, and I'm sorry if that disappoints you. My parameters were set in post #1: I would like to use simple non-programmable components (discrete components?) in a similar fashion to the original design, and microcontrollers are just not on the menu today.

Perhaps you have no experience with micros. or programming. In that case, the learning curve would be a large negative and I would understand.

I don't have any experience with them yet. I'm sure I'll use them in the future—I have an unused Arduino Leonardo and a Raspberry Pi in my drawer.

So far I have lots of great suggestions to research—thank you everyone!—and although a microcontroller is also a great suggestion, for now I'd like to learn how to use components which don't require programming for this project.

 

First, all of your descriptions of what you want are pretty poor.

I'm sorry about that; in post #1 I said:

But to replace the DIAC I'm a little lost. My first thought was to use a transistor with a capacitor triggering the base, but I can't figure out a way to do that. Another thought was a MOSFET. Does anyone have any suggestions? Thanks in advance!

I came for help with replacing the DIAC which flashes the LEDs, and I've gotten several suggestions to research, including a unijunction transistor and Schmitt trigger oscillator from you, which sound very promising; thank you!

You had asked about the flashing pattern because of my answer to Bob—I should have used the word frequency instead of pattern, as there is no relationship between the five LEDs. However each time I did say that precision wasn't important. I didn't elaborate because I'm not trying to replicate the flashing frequency, I'm trying to replace the DIAC trigger mechanism with a low-voltage method. Once I have that in place, hopefully I can tweak the flash frequencies with different resistor values. I've edited my original reply to your question so that others will understand that the flashing pattern (or lack thereof) is not important to the circuit.

I'm sorry you spent time analyzing the video. There's no clock frequency, just 10 resistors, 5 capacitors and 5 DIACs for the five LEDs as shown in the blue box in the schematic. It's very simple. However your observations are probably correct: Each LED has the same components except that the resistor marked "RATE" is a different value for each LED. This resistor probably determines how quickly the capacitor charges, and thus determines the flash frequency. Since the rest of the circuit is the same, my guess is that the amount of voltage/current fed to each LED is the same.

 

Parts list for your 5 LED sequencer:

1 8 pin PIC
1 capacitor
5 LEDs and 5 resistors (needed for any solution)

Perhaps you have no experience with micros. or programming. In that case, the learning curve would be a large negative and I would understand.

You misspelled “ATTiny13A“ as “PIC”. I know the letters are all close together on the keyboard, it’s an easy mistake to make.

 

Does anyone know if it's true that this circuit by Dick Cappels will damage the transistor over a period of time?
http://cappels.org/dproj/simplest_LED_flasher/Simplest_LED_Flasher_Circuit.html

Transistors will be damaged by excessive current causing excessive heat, or by excess voltage causing a junction breakdown, leading to excess current. So there is no reason why the relaxation circuit shown would cause damage unless some limitation is exceeded.