Best practices using 555/4017 & many LEDs

[Freddythunder]

Hello,

I'm new here.. I have a question about running a lot of LEDs in a couple scenarios. I'm building a full-size lightbar for a fake police car to be used in movies. I found a schematic for the red & blue LEDs that I'm going to duplicate using some 555s or 556 here:
http://www.instructables.com/id/41-LED-Flasher-Circuit-using-555-IC/

But I would also like to (easily & cheaply on a no-production-value film budget) get yellow LEDs on the back that would show direction by illuminating at one end and moving across. I've made this before using a 555 and a 4017, but only one LED at a time. I want to make it realistic so I'm dropping about $80 in LEDs so each segment would be 6-10 (30mA) LEDs instead of just one little one.

What would be the best, cheapest and most reliable way to do this? Can I use a transistor for the added current or a relay of some sort? I've been searching on and off for months for this solution, but I don't think I know the proper nomenclature to find what I am looking for.

Thanks!

 

[I_am_learning]

Both transistor or the relay will do.
If you use relays, then you would need to freewheeling diodes to prevent from inductive over-voltage.
Or you could use transistors just as well. Make sure the transistor can provide the current required, and you bias the base with appropriate current.

I think using transistors would be cheaper and more reliable .
And, welcome to PF.

 

[mdjensen22]

You probably don't want to switch relays as fast as your application will want. In addition the power supply requirements are much higher for driving relays as opposed to driving transistors.

Go with the transistors.

 

[Freddythunder]

Awesome, transistors it will be. Thanks for the help. I think that opened a lot of doors. This whole time I had never googled "transistor as a relay" (duh) and got a bazillion results. I'm going to read a few thousand of those and then I'll come back if I have any questions. Thanks!

 

[Freddythunder]

Actually, hang on, I have another question. I ran across this before about a year ago:
http://wild-bohemian.com/electronics/flasher.html

I remember that transistor configuration to make LEDs blink. I actually tried to get my dad to look at it (actually went to school for this stuff... I didn't) and he never got back to me. Since the LEDs are on the positive side of the transistors, is it feasible for me to run an array of LEDs off of this or will I blow up the transistors because the current still needs to course through them. This is for a car so I'll have tons of current at 12-14 volts. ...Or should I stick to the 555/6 flashers and make arrays through transistors?
Thanks again! I'm very happy I found someone who can help me with this stuff!

 

[NascentOxygen]

But I would also like to (easily & cheaply on a no-production-value film budget) get yellow LEDs on the back that would show direction by illuminating at one end and moving across. I've made this before using a 555 and a 4017, but only one LED at a time. I want to make it realistic so I'm dropping about $80 in LEDs so each segment would be 6-10 (30mA) LEDs instead of just one little one.

How many segments will comprise the full arrow? Each segment being made up of 6-10 LEDS one above the other in a vertical line?

 

[NascentOxygen]

'm building a full-size lightbar for a fake police car to be used in movies. I found a schematic for the red & blue LEDs that I'm going to duplicate using some 555s or 556 here:
http://www.instructables.com/id/41-LED-Flasher-Circuit-using-555-IC/

How many LEDs do you plan to combine for the blue light? That circuit as it stands can probably power at least double the number shown when you operate it off 12 volts, so maybe 40-50 blue LEDs. Much the same for the red.

 

[Freddythunder]

I really want to do this right and have it look as non-cheesy as I can, so I'm actually upping the amount of LEDs from what I said before and doing it right. So, I'd like to have 80 red on one side, 80 blue on the other. If I could use a power transistor or some kind of configuration for this circuit:
http://www.555-timer-circuits.com/police-lights.html
(noted the wrong one before - sorry!)

The yellow ones in the back would be 10 segments of 20 LEDs each segment.

I have put together an online model to explain in full detail what I would like to create. To activate it, click the grey buttons labeled "code3" and "yellow"
http://www.freddythunder.com/film/lightbar.html

Then once all that is done, if I want the yellow lights to blink like the red/blue ones, I'll just duplicate that circuit. I'm thinking that each section of LEDs will be a module with wires coming off of it and I can build a "brain" into one circuit board to control everything with a few 555/4017 combos. Whaduyathink?

 

[NascentOxygen]

In that circuit, they are supplying one of the colours with ten times the current as the other. I haven´t played the video, does it tell you which of their LEDs needs the greater current?

On your animation, I think you have the blue flashing differently to the red. In any case, it is too fast for me to figure out: are you wanting to flash any blue simultaneously with any red?
Maybe you could slow it down to one quarter actual speed?

 

[Freddythunder]

No I'm not sure why any of them would need 10 times the current. I'm too much a noob.

On the animation, the way I'm thinking, there's 2 blue and 2 red sections that fire and blink and then shut off, then it alternates to the remaining 2 blue and 2 red sections. The 555 sends a box wave (right?) to the 4017 making it alternate which 'side' fires. I'm sure I can get the circuit to work like the video shows, but my only question is can I use a bigger transistor and run 80 LEDs on each side of that circuit, or do I have to make multiple circuits.

As for the speed & the way they blink, quite frankly, I don't care. Just as long as they blink. It's just special lighting for the movies ;)

Thanks for your help

 

[NascentOxygen]

A pair of transistors should be able to drive 40 LEDs, at least. But first we need to know the voltage and operating current of your LEDs, an this will differ for red and blue. Have you chosen the LEDs? There are basically two choices: standard LEDs and high brightness LEDs, the latter costing more and using more current.

Different coloured LEDs differ in their efficiency at producing light, and the eye has different sensitivities between red and blue, so to have similar apparent brightnesses it can be desirable to power the different coloured LEDs with different currents.

The 555 generates a square wave.

 

[NascentOxygen]

A selection of web pages illustrating details to consider, I'm not endorsing any sites.

Two pin or four pin LEDs:
http://www.superbrightleds.com/cat/...3mm,18,166:LED_Package,5mm_High_Flux,18,1221:

Flat-top LEDs for wider viewing angle:
http://www.plasmaled.com/bright-led-lights-3mm-5mm-8mm-10mm-flat.htm

Cheap 100 packs:
https://www.sparkfun.com/categories/172

 

[Freddythunder]

Thank you for those links! I would have loved to [see if my budget allowed to] get those instead with the larger viewing. I have already bought 110 of the LEDs already from here:
http://ledshoppe.com/led10mm.htm

The white and blue are:
Emitted Colour : WHITE
Size (mm) : 10mm
Lens Colour : Water Clear
Peak Wave Length (nm) : N/A
Forward Voltage (V) : 3.2 ~ 3.8
Reverse Current (uA) : <=30
Luminous Intensity Typ Iv (mcd) : Average in 750

00
Life Rating : 100,000 Hours
Viewing Angle : ±10°
Absolute Maximum Ratings (Ta=25°C)
Max Power Dissipation : 80mw
Max Continuous Forward Current : 30mA
Max Peak Forward Current : 75mA
Reverse Voltage : 5~6V
Lead Soldering Temperature : 240°C (<5Sec)
Operating Temperature Range : -25°C ~ +85°C
Preservative Temperature Range : -30°C ~ +100°C

and the red/yellow are:

Emitted Colour : Red
Size (mm) : 10mm
Lens Colour : Water Clear
Peak Wave Length (nm) : 640 ~ 645
Forward Voltage (V) : 1.8 ~ 2.2
Reverse Current (uA) : <=30
Luminous Intensity Typ Iv (mcd) : Average in 10000
Life Rating : 100,000 Hours
Viewing Angle : ±10°
Absolute Maximum Ratings (Ta=25°C)
Max Power Dissipation : 80mw
Max Continuous Forward Current : 30mA
Max Peak Forward Current : 75mA
Reverse Voltage : 5~6V
Lead Soldering Temperature : 240°C (<5Sec)
Operating Temperature Range : -25°C ~ +85°C
Preservative Temperature Range : -30°C ~ +100°C

These LEDs have a very narrow viewing angle, but they are as bright as the sun. I'm thinking with the right blinking speed combined with camera height, shutter speed and aperture settings, I'll get real close enough to one of the new $500 lightbars on real cop cars - at least on film.

Thanks!

 

[NascentOxygen]

The white and blue are:
Emitted Colour : WHITE
Size (mm) : 10mm
Lens Colour : Water Clear
Peak Wave Length (nm) : N/A
Forward Voltage (V) : 3.2 ~ 3.8
Reverse Current (uA) : <=30
Luminous Intensity Typ Iv (mcd) : Average in 750

Those blues are so dim in comparison with your reds that you may decide to buy brighter blues. Could you power a handful side by side with an equal number of reds, and view them head-on from a distance to see how they'd go? The yellows and reds sound good.

At 10mm, they are big ones. You've bought 100, but on the figures you gave you'll be needing at least 400?

Yes, you could work within viewing angle restrictions by always filming head-on.

 

[NascentOxygen]

Just check your figures on your whites and blues, as that site currently doesn´t list 750 mcd LEDs, certainly not in 10mm.

At first glance, I think it will be feasible to use a driver module, and with care operate parallel strings for the blinking LEDs, e.g., the decade counter can control two 4015 modules each driving 84 LEDs to produce the blinking. To keep down cost, the yellow LEDs can be switched by transistors.

Have a look at the 4015 BoostPuck http://www.ledsupply.com/led-drivers.php Without boosting the +12v from the car battery, you´d be limited to powering 3 of the blue LEDs in a string; by boosting it to +28v you can power 7 blue LEDs per string.

 

[Freddythunder]

Okay. Sorry for the delay. Got laid off from my job and have been job hunting ever since.

I read somewhere on here that I can't count and it seems I cannot. Sorry my led count was off. But it looks like I still have some planning to do. I took a look at a real lightbar a couple days ago and I think the LEDs are ten on top, five on bottom so it may be less LEDs anyway.

So what you are suggesting instead of using the bc547 transistors (or alternate higher power transistors) I could use these boost puck thingies? Looks like a new world.

And can you show me how I would get the yellow ones to work like that using just transistors? Maybe a link to a tutorial or something?

I'm real close to just making the red and blue blink back and forth with just transistors to save time and money. As for the brightness, no matter what the numbers say, both the red and blue are bright enough for filming. They may actually be too bright and I may have to use larger resistors, that is if I ever get that far.

Thanks for the help!

 

[NascentOxygen]

Okay. Sorry for the delay. Got laid off from my job and have been job hunting ever since.

Nasty. Not something you saw coming?

So what you are suggesting instead of using the bc547 transistors (or alternate higher power transistors) I could use these boost puck thingies? Looks like a new world.

I was suggesting this, it would save transistors and resistors and a lot of connections. However, in light of your opinion that at full current the LEDs may be too bright, this eases the design constraints I was working within.

It's tempting to connect 3 of the blue LEDs in series with a resistor and power that string from the +12v. But there's a difficulty in designing for say, 22mA of LED current at a battery voltage of +12, while avoiding having that current soar to almost double that value as the battery warms up under charge and its terminal voltage rises to around +13.5v.

That regulator module I pointed to would have solved this. However, until you work out precisely how many LEDs need to be powered in unison, this can remain as "to be decided".

Regardless, some tests for you: connect one of your blue LEDs in series with a 390Ω resistor to a +12v supply, and record the voltage across the LED. Now substitute one at a time another fifteen blue LEDs, recording their voltage each time. That will indicate the voltage spread to allow for.

Can you do the same for your YELLOWS and, separately, the REDS.

At this stage I'm needing to know how many RED and BLUE together will flash in unison, to see whether they can all be economically powered via a single transistor.

As for the job hunting -

http://img502.imageshack.us/img502/7752/goodluckemoticon.gif​

 

[Freddythunder]

As for the job, I had an idea it was coming soon. But two Fridays ago my boss said not to worry there was plenty of work and then Monday the company couldn't pay me anymore. I got feet on the pavement though.

Anyway, I'm not sure I understood what you meant about testing voltages. I started with a 330 ohm resistor and one blue, the voltage was 3.29. 2 went to 3.18. When I got to three, the first led was 3.05 and the third was 3.18.
4, 1st was 2.92 and the last was 3.06
5, 2.76 and 2.86
6, 2.54 and 2.56
By the time I got 7 LEDs hooked up in series, it was just under 2 volts and they were barely on. At that point I figured I'd check to see what you ment cause I think I'm testing things wrong.

I'm going to do some camera tests tonight to find out exactly how many LEDs will be in each section. I have 10 red and ten blue plugged into my breadboard just running solid at I think 30ma each. Its painful to look at so I might drop it down to 10 LEDs per section as its for movies and not real traffic visibility. I'll keep you posted!

 

[NascentOxygen]

By "substitute" I mean remove the first LED and replace it with a different one, so that at anyone time there is just the resistor and one LED. Different LEDs may have slightly different voltages, those differences are what I'm interested to know.

Since you're starting over, can you use a 470Ω resistor this time, the results will be a little more representative; I'm getting the feeling we can aim for less than the full 30mA.

To answer an earlier question in this thread, I'm sure you'll be able to operate all the sequences off the one 555/4017 counter, you won't need to duplicate the clocking part. This means your sequences will all be fully synchronized.

Just confirm: the yellow strip comprising 10 blocks, when it's activated, will have the left-most block constantly lit, and the others will sequentially fire and each then remain lit until the right-most block has been lit, then 9 blocks will go dark together but the left-most block will never go dark at any time?

(Suggestion: if that is the case, then it may be worth making the yellow strip 11 blocks, since the counter can operate 10 and your left-most block does not need to be clocked if it is to stay lit constantly.)

I haven't been able to figure out the whites—what they do, why they are located where they are, and how many there'll be.

 

[Freddythunder]

I really appreciate your help with this. I really have no one else to ask this stuff to.

Okay, now I know what voltages you need. Here are the results with the 470 ohm resistor (yellow purple brown)
Yellow, 2 volts
Red, 2.1 volts
Blue, 3.27 volts

That is awesome that the red and blue is only one 555/4017 they cab be synced, or not, won't matter.

As for the yellow, that is completely right about the sequence. I think it would be more realistic if the first one on the left shut off when it resets, but whatever gets me close enough is close enough.

I'm going to go test the 20 LEDs I have on the breadboard to see how it looks on the camera. After that I'll post back exactly the final count of how many LEDs

 

[Freddythunder]

Alright, film testing done. I will definitely be going with only 10 LEDs per section, not 20. These LEDs are so bright I may have to use larger resistors on the blue so they look more blue in the camera. So, finally, I'll have 40 blue and 40 red. 8 total modules of LEDs where four and four will alternate with the flashing like the animation I put together.

What do ou think?

 

[NascentOxygen]

I really appreciate your help with this. I really have no one else to ask this stuff to.

There are plenty of others here who'll help you.

Okay, now I know what voltages you need. Here are the results with the 470 ohm resistor (yellow purple brown)
Yellow, 2 volts
Red, 2.1 volts
Blue, 3.27 volts

I was hoping you'd measure a dozen of each colour! Especially the blues now. They won't all be 3.27v and I'd like to see how much they can differ.

As for the yellow, that is completely right about the sequence. I think it would be more realistic if the first one on the left shut off when it resets,

So you want it to differ from your web simulation? You want the whole yellow panel to be fully blacked out for 1/11 of the cycle?

but whatever gets me close enough is close enough.

Not the way I look at it. You can get precisely what you want, but the time to decide is now, before final calculations, since design constraints are in some respects tight if you want lowest cost.

 

[NascentOxygen]

On examining the way the 4017 operates, I see you have to choose between:
(a) 10 blocks of yellow, with the leftmost never darkened; or
(b) 9 blocks of yellow, with the leftmost being darkened for 1/10 of the cycle.

 

[NascentOxygen]

This should be right for driving up to twelve of the yellow or red LEDs. Construct one to start with, and test it.

With it being able to drive 12 LEDs, consider making your yellow and red blocks multiples of 12.
The circuit to drive blue and white LEDs will be a bit different.

 

[Freddythunder]

A couple questions about that.. I see the 547's set up as a darlington pair (read about that it my travels) is that to allow for higher current? Is this circuit going to condition the voltage or current to save the life of the LEDs? You had mentioned that using car power voltage will fluctuate. I'm guessing that the LEDs are on the right, are those other 5/6 on the left diodes for one way power travels? And what value is C1?

Last question because I'm newer to electronics, I'm cheap, and I'm excited so I'm in a hurry (I'm working on that one) is it not advisable to just do something like this:
led.linear1.org/led.wiz/?VS=13&VF=2.8&ID=30&N=10&output=s

Sorry for all the questions, I do not doubt your help. I want to learn and understand as well as accomplish my goal.

Thanks!

 

[NascentOxygen]

A couple questions about that.. I see the 547's set up as a darlington pair (read about that it my travels) is that to allow for higher current?

It means the circuit will take very little current from the 4017. The 4017 data sheet shows it can output currents of only about 10mA. This is not sufficient for your needs, where one of the 4017 outputs will be driving maybe 14 separate blocks of LEDs, so a Darlington pair is needed to minimise the load on the 4017.

Is this circuit going to condition the voltage or current to save the life of the LEDs?

This limits the current to the LEDs to about 22mA. should battery voltage rise to 13.8v.

I'm guessing that the LEDs are on the right, are those other 5/6 on the left diodes for one way power travels? And what value is C1?

The small signal silicon diodes on the left form an OR gate and they come from outputs 1-9 of the 4017. Different blocks will have different numbers of diodes connected to different 4017 outputs to set it in the blinking sequence. C1 is an electrolytic, say 220μF, 25v but not needed at this stage.

Last question because I'm newer to electronics, I'm cheap, and I'm excited so I'm in a hurry (I'm working on that one) is it not advisable to just do something like this:
led.linear1.org/led.wiz/?VS=13&VF=2.8&ID=30&N=10&output=s

That calculator should give the same results I have.

 

[Freddythunder]

I can't seem to find 547 transistors very easily. Can I use 2222 or 3904 instead?

 

[NascentOxygen]

I can't seem to find 547 transistors very easily. Can I use 2222 or 3904 instead?

The 2N2222 should do just as well.

 

[Freddythunder]

Oh and btw, the white ones are just on off. The are like flashlights. 2 on the front and one on each side. For realism. I was planning on just running power to them straight

 

[Freddythunder]

On examining the way the 4017 operates, I see you have to choose between:
(a) 10 blocks of yellow, with the leftmost never darkened; or
(b) 9 blocks of yellow, with the leftmost being darkened for 1/10 of the cycle.

Then B. I would like the leftmost one to darken. That is how the real ones do it so I'll shoot for that.

I totally misunderstood what you meant about the LED voltages, but got it. Grabbed the DVM and started testing. Here is a link to a table of the data that I recorded:
http://www.freddythunder.com/film/leddata.php

I'm using some of the blue ones in a different configuration so if you need more than those 10, let me know I'll test the others.

I'm off to radio shack to get a couple 2222's. I will mail order some 547s, they're like $6 per 50 on amazon, but want to get started with this right away. I'm going to get a small board and some solid wire and put together what you suggest. I'll let you know what it is complete.

 

[NascentOxygen]

Important: the 2222 has a lower gain, so change R1 to 3.9kΩ (Always check initial operation: no components should get hot, not even warm to the touch.)

You might find it useful to construct just one circuit for testing purposes, and power all three colours from it simultaneously so you can photograph them together and, if necessary, set each to a more-suitable brightness.

So, using that basic circuit, instead of the 10 yellow LEDs, you could have:
one string of 4 yellows in series with a 220Ω, plus
a string of 4 reds in series with 220Ω, plus
a string of 2 blues in series with 390Ω, plus
another string of 2 blues in series with 390Ω

This would give 4 of each colour, and give you the opportunity to modify the resistor values to determine best luminosity for camera work. Don't forget that when the car's voltage varies these LEDs will change brightness a bit, too. https://www.physicsforums.com/images/icons/icon6.gif

 

[Freddythunder]

Well went to an electronics parts place in Mesa today in hopes of finding the 547s. The 2222 are 2 bucks each at the radio shack so no thanks. I'm ordered 50 547Cs for $6, should be here in a few days.

I'm planning on doing the first circuit on a breadboard with the multiple colors to test like you suggested (good thinking!) and the solder together my sections of 10 LEDs each.

I'm going to use a 555 to turn the blocks on and off to test from pin 3. I think that's right. I'll alk to you in a few days.

Btw, found new job, start Monday.. All very exciting.

 

[NascentOxygen]

Btw, found new job, start Monday.. All very exciting.

Great. That's fast work!

 

[Freddythunder]

I actually got really lucky. Got 2 job offers after one week of looking. I'm a PHP developer and there were a ton of postings, but then this week - nothing out there. I'm thanking God above for the opportunities & the timing.

Anyway, while I wait impatiently for my transistors to come in, I came across another question. I was going to post a new thread, but figured I'd ask you first. I know that at Walgreens, they sell these balls that have a little flasher toy inside of them, when the ball gets knocked around, a blue & red led blink just like the police flasher thing. Years ago when I was a bartender, there were a few promotions where the booze company would give out these novelty ice cubes that had a blinking led inside of them, but it would blink like 1-2-3 pause 1-2-3 pause - much like I want my project to do. Being a kid inside, I've torn a toy or two apart like this to find a very small circuit board with some black plastic shell over any electronics.

My question is, is there some more simple way to do what I'm trying to do - or do you think that these toys have some sort of 555/4017 inside of them that make them do the same thing I'm trying to do? Is there some sort of other chip out there that is manufactured just to blink.

Or, maybe it's two 555 timers, one set to blink fast and another slow and the one 555 triggers the next 555 so you get that blinking, then pause, then blinking, then pause. Hmm. Maybe I'll play with that while I wait for my transistors.

Just a random thought I had.

 

[NascentOxygen]

Using 4017 is by far the simplest method, in principle, for what you are doing, because the sequences are locked together and controlled by a single clock. To speed up or slow down the rate of blinking you'll just adjust one 555 and there are no timing issues. However, it also follows that you can't independently control things, so in that sense your options are limited. If you wanted 3 successive blinks, you'd need to use two 4017's with this timing method.

I thought you said you had already built the basic 555 + 4017 circuit? How much did the 4017 cost?

While cooling your heels, you could turn to the task of figuring out how many input diodes you'll need for each of the 9 yellow blocks. This means you have to work out what combination of outputs from the 4017 is needed for each yellow block to turn it on and hold it on. Maybe start with the timing diagram here: http://electronics.stackexchange.co...-multiple-connections-in-this-traffic-light-c

You'll also need to search for the full 4017 pin-out.

You might re-consider the 2222's. If you use 547's throughout you will have to construct at least 5 blocks to do what a single 2N2222 can. The 2N2222's are rated for more current, so one 2N2222 can drive half of the blues together with half of the reds (providing my hunch is right that you'll settle on less than 20mA drive for each LED). You'd need just two 2222's, all the rest can be 547s.

 

[Freddythunder]

I thought you said you had already built the basic 555 + 4017 circuit? How much did the 4017 cost?

On my iPad and I can't use the nifty quote feature... So I made a circuit about 2 years ago using a 555 & 4017 that sequentially turned on and off 10 single LEDs. I took a video of it, but I don't know where it could be. I bought the 4017 at fry's electronics, actually bought a backup still in the package. Brand new NTE4017B for $1.69. Also have a 4013 CMOS flip flop. No idea what that does butive kept it save and static free for years.

As for the 2222s they were too expensive so I never bought them. I was thinking of making circuit boards for each segment of lights like your schematic so each segment would be 2 transistors, the LEDs and resistors. Then I'll just have three wires on each for power and the signal from the respective 4017 output pin. Right?

I made a blinking circuit with a 555 and a potentiometer to adjust. I ran pin 3 to your circuit of led array. I'm thinking it's working as it should because it did not fry the 555.
http://m.youtube.com/watch?v=w8EhJIqXOYI

The only thing I see that looks off is the array of 10 LEDs seems to ramp up instead of just turn on. I slow it down in the video so you can see it. This may make a difference with the red and blue because of how fast they're going to blink; they wouldn't be on long enough to get bright.

This may be the issue too. I had to use 66k resistors instead of 68k and 4.7k instead of the 5.6k. Other than that I don't have the understanding to know where to start to look for the issue. The led off the 555 does not do it. Thanks!

[EDIT!]
I turned up the pot so they blink really quickly - there's no difference, they are as bright as the sun still when they blink fast. Proof positive is I cannot see the screen right now due to black dots left in my vision from 11 rapid fire yellow LEDs shooting the back of my retinas. So I don't think that issue is an issue anymore.
[/EDIT!]

 

[NascentOxygen]

I was thinking of making circuit boards for each segment of lights like your schematic so each segment would be 2 transistors, the LEDs and resistors. Then I'll just have three wires on each for power and the signal from the respective 4017 output pin. Right?

That should work okay.

I made a blinking circuit with a 555 and a potentiometer to adjust. I ran pin 3 to your circuit of led array. I'm thinking it's working as it should because it did not fry the 555.

That's progress. Did you put your finger on the transistors, to feel their temperature? The component most at risk is the transistor driving the LEDs.

The only thing I see that looks off is the array of 10 LEDs seems to ramp up instead of just turn on.

I see what you're saying. It looks like you are driving 5 strings of LEDs? In the schematic, I showed only 3, as 3 is the maximum I'm comfortable with for the BC547. So remove all but one string of LEDs and see how it behaves with just one string. What value of resistor are you using for each pair of LEDs? Did you measure the supply voltage?

Have you confirmed that the brightness will be sufficient if you use a string of 4 LEDs and a 220Ω resistor?

Also remove that single LED and the resistor feeding it. Does that improve things? Between what two points do you have that LED connected?

After all this, if the LEDs continue to exhibit steps in brightness, connect a large electrolytic between +12v and ground right at the circuit board. At least 2000μF and note any improvement.

 

[Freddythunder]

Did you mention something about your circuit running the LEDs at 20ma? Aren't they rated for 30? I tried your resistor like you said and they are bright enough. I got a circuit board and cut it to pieces and have started making my arrays. I cut the pieces to narrow and don't want to crowd everything so I'm going to put the transistors on a separate smaller board between each two segments. it's good though. I want it to be real low profile.

I had to sand down the edges of these LEDs so they would fit next to each other but that's why God gave me a dremel.

http://twitter.com/freddythunder/status/238136953231536130/photo/1

So my next steps are to finish making the arrays and the the transistor boards then the 4017 circuit. I'm doing the red and blue first then the yellow second.

 

[NascentOxygen]

Did you mention something about your circuit running the LEDs at 20ma? Aren't they rated for 30?

Nominally 20mA, but rising a bit if the car voltage is higher than 12v. If you want electronic devices to give reliable service, you don't push them to their absolute limit.

After passing through the LEDs, the current then goes through the BC547. The BC547 does not have a very high current rating. 3 strings of nominally 20mA is what I planned for.

Can I assume the LED switching is now behaving perfectly? Which of the changes you made fixed that stepping up in brightness?

What company manufactured your 555? What is the full component number on it?

[strike]"Firefox can't establish a connection to the server at twitter.com."[/strike] EDIT: I had twitter blacklisted. https://www.physicsforums.com/images/icons/icon11.gif

 

[NascentOxygen]

Since you are constructing this now, I'll summarize things:

One pair of 547's will drive 3 strings of LEDs with each string made up of:
(i) 4 RED or YELLOW LEDs and a 220Ω ½ watt, or
(ii) 2 BLUE LEDs and a 390Ω ½ watt

So each pair of 547's will comfortably switch 12 REDS but only 6 BLUES.

You haven't yet confirmed that 390Ω with 2 BLUES gives an optimal brightness for your needs. Maybe it is still too bright, or perhaps not bright enough?

 

[Freddythunder]

Whoof, new jobs got me busy!

I have started constructing the LEDs to boards, but hadn't connected anything together yet. But what you said above confuses me so I'm going to have to reaffirm things with you. I don't want the LEDs to burn out so I want to make sure that I'm doing this right.

So, with the above, are you saying that each pair of transistors will light 12 red or yellow LEDs but only 6 Blue LEDs?? Is 10 Blue LEDs too much current for them? Can I use the 2222's for the blue ones? I'm going to need 10 LEDs on each section no matter the color so they are physically the same size.

I did try 4 blue LEDs with 100Ω and I think I did 2 with the 390Ω. I could not tell the different in brightness, both configurations were really bright. As for filming, in my more previous tests, I just ran 12V through 2 blue LEDs with a 270Ω resister and they looked blue to the eye, but almost white on camera. That LED calculator thing told me that I can run 14V at 20mA with 4 blue LEDs and a 68Ω and 2 blue LEDs with a 390Ω. Can I do that with a different circuit?

 

[NascentOxygen]

Whoof, new jobs got me busy!

Good.

I don't want the LEDs to burn out so I want to make sure that I'm doing this right.

The LEDs aren't at much risk, it's the low-power BC547's that are being pushed hard.

So, with the above, are you saying that each pair of transistors will light 12 red or yellow LEDs but only 6 Blue LEDs??

Yes, if you use the blues with 390Ω.

Is 10 Blue LEDs too much current for them?

Yes. But if the camera is happy with the blues when you change to using 560Ω resistors, then you can safely power 10 blues, arranged in 5 strings each with 560Ω. Try that and see.

Can I use the 2222's for the blue ones? I'm going to need 10 LEDs on each section no matter the color so they are physically the same size.

As I pointed out previously, the 2222 has a higher current rating, and a single 2222 could switch half of the blues plus half of the reds all together, a total of 40 if I'm not mistaken, which is ultimately what needs to be done, isn't it? Your call.

I half expect that with your blue LEDs the camera will always perceive intense white dots with a bluish halo whatever the current. To film a proper blue you may have to buy LEDs encapsulated in blue-coloured or cloudy plastic, not the clear stuff. (But I'm not certain of this, it needs to be tested.)

A wild thought: maybe a frosted dome on the LED would diffuse the light better? Have you enough blues that you could sacrifice one or two to an experiment? Use fine sandpaper to evenly roughen one LEDs hemispherical cap and turn it into a uniformly matt surface and see what the camera thinks of that alongside an unmodified sample for comparison. Also could try a smear of white toothpaste over the end of a LED to diffuse the light into a less intense but larger spot. (If it works well, go looking for some cloudy plastic to glue onto your blue LEDS.)

 

[Freddythunder]

I will test the blues with the higher resistance, uh, resistors right now to see how bright they are. Otherwise, I will buy some 2222's online, I still need to buy more LEDs. I will sand down one of the blues and see what happens. I actually have to sand down the sides for them to fit side by side on the circuit boards so I already have some fine sandpaper with my current-hobby kit.

I don't think the toothpaste would work because I plan to enclose this whole thing in plastic & plexiglass. Plus I live in Phoenix so the toothpaste would probably just melt off. You'd be surprised what melts out here... In October...

So with the 2222's could I have one 2222 for the 40 blue and 40 red ones that all blink together and another one of them to alternately blink the other 80 red and blue ones? I'd eat the $6 I spend on the (50) 547s if I only have to wire up 4 transistors for the whole red/blue side! That'd be great! Yes that's what I want to do. Then I can have all the components on one board and run two wires to each segment. Yep.

Thanks

 

[NascentOxygen]

The toothpaste was an improvised test! If it showed promise you could use a dab of white primer.

So you have returned to using 80 red and 80 blue? https://www.physicsforums.com/showpost.php?p=4032050&postcount=21

Let me know how the blues with lower current go. If 80 proves too many for a 2222, we can probably use a BD139. I'll have another look at the figures ...

You have tried filming the reds and yellows with the resistor values I offered earlier, so are certain they don't need to have drive current reduced?

 

[Freddythunder]

Ahh! Test toothpaste, that makes more sense. I still have yet to shave one down with sandpaper but I wanted to let you know, NO I'm not at 80 LEDs. I have no idea why I cannot keep that straight in my head. It's 80 total, 8 segments of 10 because they proved bright enough on the test - so it's half of the LEDs in that animation I made.

On that note, one 2222 would run 20 red and 20 blue, then another 2222 would run the other 20 red and 20 blue. Whew. Sorry about that.

 

[NascentOxygen]

Don't get the 2222, it will get too hot. You'll need the higher-power BD139.

 

[Freddythunder]

Change of plans! [2 months later] I think that I will instead be using 556 for each 'panel' of LEDs (going through the darlington transistors). I figured out (after scouring the interwebs for the schematic) a way to get the blinking effect that I want by having two astable multivibrators one quick one slow and the slow one turns the quick one on and off. That way all the panels will be blinking independently which is probably better anyhow.

Plus I'm thinking as things break, they will be easier and cheaper to fix down the line. I remember my dad telling me that 555's are pretty sturdy and pretty hard to blow up.

Let me know what you think, can I use a 556 to make the blinking, take the output of the second multivibrator and run it through two of these 547 transistors I now have a bunch of and run 10 LEDs per circuit?

 

[NascentOxygen]

Welcome back! I wondered about the sudden disappearance.

Change of plans! [2 months later] I think that I will instead be using 556 for each 'panel' of LEDs (going through the darlington transistors). I figured out (after scouring the interwebs for the schematic) a way to get the blinking effect that I want by having two astable multivibrators one quick one slow and the slow one turns the quick one on and off. That way all the panels will be blinking independently which is probably better anyhow.

The blinking function can be achieved using 555s, yes. But how do you propose to have the yellow blocks light up sequentially?

Plus I'm thinking as things break, they will be easier and cheaper to fix down the line. I remember my dad telling me that 555's are pretty sturdy and pretty hard to blow up.

Any electronic component can be destroyed by abuse or misuse. 4017's and diodes are not expensive.

 

[Freddythunder]

Thank you, it's good to be back! I will still have to build the yellow blocks with the 4017 but need the red and blue first to start filming. My step dad has THE car to use in the film project but won't have it forever so I have to step up my timeline.

You are correct about the parts being cheap however, I don't like how the 4017 are static sensitive and how the whole thing could die from a shock from a finger. Plus I think knowing me and my building style I would like to have them in modules (each block of LEDs) so I can make extras for other parts of the car and have backup segments on film day. If that light ar doesn't work when filming then there's no point to filming.

Price wise the 556 is .40 so it'll keep cheap anyway. My step dad suggested I build the lightbar into a four foot fluorescent light fixture so I'm at the point of building. I think I trust the 556 more too.

 

[Freddythunder]

I've combined the 556 circuit with the led and 547 transistors array you suggested and let it run for a while. All seems good so I'm going to start building. Wish me luck.