PDA

View Full Version : Solid State Brake Light Flasher


jj_alukkas
08-02-2009, 09:47 AM
Hi,

I made a brake light flasher for my car using a 555 and a 12v relay to blink two pairs of bulbs alternatively. It works fine but makes a lot of noise each time I step on the brake and it is quite irritating to hear that tic-tic at a traffic signal for a long time. So I designed a solid state circuit of the same. I managed to design the +ve side driver but I have no idea what to do for the -ve side to make it like N/c and N/o of a relay and blink the lights alternatively. Each pair would have a power consumption of 30W(max) I guess.How would I go on for the other side??

http://forum.allaboutcircuits.com/attachment.php?attachmentid=11207&stc=1&d=1249206296

KMoffett
08-02-2009, 12:28 PM
It works fine but makes a lot of noise each time I step on the brake and it is quite irritating to hear that tic-tic at a traffic signal for a long time.
If that "tic-tic" irritated you, think how the driver behind you, at that long light, feels about your continuously flashing brake lights in their eyes...especially at night. ....I didn't like it much.

I did this once, but mine flashed three times when I pressed on the brake, and then stayed on steady.

Alternating would just require another TIP117 and a 2.7K resistor. Emitter to Brake Line +12, collector to the second bulb, and the resistor from the first TIP117's collector to the second TIP117's base.

Also, alternating brake lights might cause a lot of confusion for other drivers about your intentions if you are also using your turn signal.

And in some places this is illegal. ;)

ken

jj_alukkas
08-02-2009, 05:09 PM
If that "tic-tic" irritated you, think how the driver behind you, at that long light, feels about your continuously flashing brake lights in their eyes...especially at night. ....I didn't like it much.


Ya, you are absolutely right, but My car very rarely gets out at night and during the day, you see it only if you look at it. Anyway I plan to remove it cos it may take attention of drivers from far away. I just fitted it today didnt feel muach abt the over-effect at that time. Now I think its not good. Anyway I liked to know how to drive it.

It might be illegal here too but lots of people do this here and no one cares abt it. But the difference is that those people have small tail lights. So alternating wouldnt cause distraction. But my car has a big tail cluster as its a hatchback with one bulb on the top and other at the bottom which will cause problems. Thanks for pointing it out. Either I'll switch to pulsar ( few blinks and then on) or Ill remove it.

fanie
08-02-2009, 06:05 PM
Joseph,

Personally I don't like 555's, but the idea of the flasher is a good one. Put a diode in line to the 555 and a 100uF cap to isolate it from the 12V.


You can also design a circuit to flash the brake lights on ie 3 times then stay on using one 4093 gate as slow flip flop to step a 4017. The 4017 is a decade counter. The outputs switches the remaining three 4093 gates in parallel through diodes that drives a FET or relay to switch the brake lights. Bit of logics but you can figure it out. The IRFZ44 is a good choice FET for such an app.

KMoffett
08-02-2009, 10:10 PM
This is the one I made several years ago. The PCB was just for layout verification, I actually just point-to-point wired it on perf board. It fit inside the third brake light lense. Three flashes...then steady. :)

ken

jj_alukkas
08-03-2009, 12:44 PM
Personally I don't like 555's, but the idea of the flasher is a good one. Put a diode in line to the 555 and a 100uF cap to isolate it from the 12V.

Ya, that chip ,might appear antique but it is almost a universal solution for most circuits. I really started to rely upon them as I have built lots of them in the past and didnt like it in the beginning. But now abt 10 555's serve my daily uses since years. It havent ever failed or malfunctioned so I really like its robust perforance. Thats the 1st reason I use them.

I have a circuit with a 4093 which does all in one. Ill mostly switch to it.

And KMoffett (http://forum.allaboutcircuits.com/member.php?u=16292), thats almost ditto to the circuit I have. Thanks a lot !!

jj_alukkas
08-18-2009, 03:48 PM
I breadboarded Ken's circuit using CD4093B and it works, but the lamp I tested with was a 12v 3W bulb and it doesnt have much contrast between the On and OFF levels. Is it due to the bulb I am using or will it change when I wire it with the 10W brake bulb. I dont intend to have a full off and On, the current circuit setup is excellent, just I need the light to be a bit more dimmer at the low levels, not completely off either. What to do??

http://forum.allaboutcircuits.com/attachment.php?attachmentid=11216&d=1249250685

jj_alukkas
08-18-2009, 03:50 PM
And by the way Ken, I only recently figured out that the flasher ckt I had saved 2 yrs ago from some website was designed by you. Cheers for you man, keep up the great works.

KMoffett
08-19-2009, 12:09 AM
The output should be able to drive the lamp from full off to full (well -2.5 Vsat) on. So, you want the pulses to dim the bulb...or...do you want the bulb normally dim with full-on pulses?

ken

jj_alukkas
08-19-2009, 08:10 AM
Sorry, it was a wiring mistake.. In the replacement of the driving transistors, I accidently swaped the E & C of BC547. Now what I only need is how to alternate 2 sets of bulbs. The current setup works fine with a 15W 12V bulb. I am using TIP127, which being a 65W darlington pair, will it be able to drive, say 30-40 W easily. It will need a heatsink, but will it be in stress? Also I like to alternate the upper and lower lights during the flash, so I plan to connect another TIP127 to this one with the 2nd ones Emitter to +12, collector to second bulb and base to 1st Tip's collector through a 2K7.. will it fine?? Also is there a slight modification possibilty to make the 2 TIP's stay on after the pulsing period instead of only the first one alone??

KMoffett
08-19-2009, 01:41 PM
For full-on/full-off of higher wattage lamps you might want to drop R9 and replace Q2 with a low Ron, p-channel, power MOSFET. Much lower power dissipation in the MOSFET than the darlingtons. The thermal time lag in the bigger lamps may do you dimming.

As for alternating lamps, ending with both on...maybe an gated oscillator (4093) driving a 4017 counter. Diode OR-ing the counter outputs so outputs 0/2/4/6/8 drive lamp1, outputs 1/3/5/7 drive lamp2, output 9 drives both lamps1, lamp2, and kills the oscillator. No schematic...just an idea.

If you absolutely need dimming, I would go with a small microprocessor driving MOSFETs with PWM. Then you can do any sequencing and dimming you can imagine. ;)

Ken

jj_alukkas
08-19-2009, 04:30 PM
The lighting setup is completely fine as well as its timing, I just want another set to alternate. Its ok if it wont remain lit after the blink, but is the connection for the second TIP I stated correct?? MOSFETS's would have been more efficient, but this thing lights only for a few seconds, so its ok I think.

KMoffett
08-19-2009, 04:36 PM
is the connection for the second TIP I stated correct??

Yes. It should work.

ken

SIcam
08-19-2009, 05:34 PM
There is a Federal Requirement for the flash rate and duty cycle of a turn signal.

FMVSS 108 details the requirements.


If a turn signal is included in the headlamp assembly, it shall be operated at 90 flashes a minute with a 75 <plus-minus> 2% current ``on time.''

If you alter the Federal requirements on a vehicle and are in an accident the changes that do not meet the requirements could be used against you.

Solid State blinker flashers are needed over thermal flashers due to the low current draw on some LED lighting.

KMoffett
08-19-2009, 05:41 PM
SIcam,

Did you check JJ's location? I don't think he'll be concerned about our Federal requirements. No clue as to what the Indian government mandates are. From our discussion I think JJ is using conventional incandescent lamps, not LEDs.

Ken

jj_alukkas
08-19-2009, 07:34 PM
Did you check JJ's location? I don't think he'll be concerned about our Federal requirements. No clue as to what the Indian government mandates are. From our discussion I think JJ is using conventional incandescent lamps, not LEDs.


Yes Ken, you are right, we have no problems here with such lights.

And Yes, im using incandescent lamps, just an extra ckt to the current tail cluster. I'd love to make an LED one, but since the tail set integrates the normal types, I left the plan as I would have to crack open the clear lens. Anyway thanks Ken, Love your design.

jj_alukkas
08-19-2009, 07:36 PM
Solid State blinker flashers are needed over thermal flashers due to the low current draw on some LED lighting.

Actually I use them despite the heat loss since it irritates me with its sound, otherwise I would only prefer relays.

SIcam
08-19-2009, 11:20 PM
Sorry, didnt realize you were in another country.

What does (+5.5) at the end mean?

jj_alukkas
08-20-2009, 05:54 AM
5 hrs 30 mins more than Greenwhich mean time(GMT) . I guess US is abt -6 hrs. So a total of 5.30 + 6 = 11.3 hrs difference b/n indian and US time.

jj_alukkas
08-20-2009, 03:42 PM
If I invert Pins 3 and 11 using transistors and OR them using diodes, will I have 2 lights alternating during the flash and both lit up after the flash period? I found this to use fewer and simpler components than to invert Pin 12 and NAND it with pin 3. Also, how do I create a NOT using a single NPN transistor? I know the connections, but biasing resistor values seems to be wrong according to my ideas.

KMoffett
08-20-2009, 04:46 PM
Can you post a schematic of what you have and want to do?

Ken

jj_alukkas
08-20-2009, 05:40 PM
This is your circuit I found online, same as the one you posted on this thread.
D was your original output and what I need along with D to drive my second bulb is the last one.
In words, during the flashing period, the 2 bulbs will alternate. After the flashing , both the bulbs will light up together. I need your help on how to make an inverting gate (NOT) using a transistor and wiring both the inverted outputs to the final bulb after adding them (OR).

http://forum.allaboutcircuits.com/attachment.php?attachmentid=11631&stc=1&d=1250789939

KMoffett
08-20-2009, 07:39 PM
I think this will work.

Ken

jj_alukkas
08-20-2009, 08:15 PM
Thanks Ken, I will test and reply sooon.

KMoffett
08-20-2009, 09:57 PM
I did that very quickly...and now I see a problem with it. Don't bother trying it. I'll take another shot at it later.

Ken

KMoffett
08-21-2009, 12:43 AM
OK, a little more thinking. It does get more complicated. I think this works...at least in my head. :)

Ken

jj_alukkas
08-21-2009, 07:49 AM
I tested the ckt you posted before the above one and found a minor problem of Q4 drawing too much of the collector current from Q1, and as a result Bulb 1 stays lit and Bulb 2 blinks as it should. So I added a 100k resistor before D2 and solved the problem.
Now it works good with your previous circuit.

Is there any trouble in following the previous one? I'm happy with it. Thanks a lot Ken. Also I would like to have your permission to publish this ckt on my website.

KMoffett
08-21-2009, 09:44 AM
Yes that was the problem,but since I didn't have a way the bench test it after I saw that problem, I opt'ed for the second circuit. Happy It's working for you.

Ken

jj_alukkas
08-21-2009, 12:14 PM
Thought I'd upload the redesigned circuit.

http://forum.allaboutcircuits.com/attachment.php?attachmentid=11655&stc=1&d=1250856861

jj_alukkas
08-21-2009, 12:16 PM
Ken, which CAD soft do you use for illustrations??

KMoffett
08-21-2009, 02:04 PM
ExpressPCB http://www.expresspcb.com/

It's pretty intuitive and easy to make new schematic or PCB components...and it's free. :)
The down side is you are locked into their PCB house. But, their "mini board" price is OK for me.
I frequently use the PCB layout program to create a layouts for one-off, point-to-point wired, perf-board projects. The schematic and PCB are linked so you can find layout errors. If components don't have 0.1" (12.5mm) spacing, just create new ones that do.

Ken

Søren
08-22-2009, 01:55 PM
Hi,

If you modify the oscillator to run in a single gate, you free one for making a complimentary output - like this:
http://That.Homepage.dk/Img/BrakeFlasherMod.png

jj_alukkas
08-22-2009, 03:53 PM
Hmm.. seems interesting. Thanks, but I'm planning to settle on the last one as I dont have time now for benchtests. Will try it another time.

KMoffett
08-22-2009, 05:02 PM
Soren,

I like the way you freed up the extra gate for me. ;)

JJ,

Sorry I couldn't leave it alone. :)

Ken

jj_alukkas
08-22-2009, 08:12 PM
Hmm.. This looks better, should I give it a try??

KMoffett
08-22-2009, 08:36 PM
I haven't bench tested it, but i think it should work. Give it a shot.

ken

jj_alukkas
08-23-2009, 06:34 AM
Will do a bench this evening as it seems easier to solder on a common PCB with fewer components.

jj_alukkas
08-23-2009, 04:34 PM
I benched it and works good. One problem is that the duty cycle isn't 50%. Around 60-70%. Dont have a scope to measure. Still it works good as a light flasher. Also R5 and R4 values both need to be 100k as the flashing is too quick and the 50k doesnt provide enough room for adjustment and the major variations occur after 100k. I used the components from the previous circuit and now I have 3 resistors and a transistor left over. You guys are great. I think I can now solder this ckt. Or any more mods??

I like the way you designed this circuit. I never thought that an oscilator could be made with an RC and a NAND gate. How do you calculate the frequency? I've been into electronics since a few yrs and I learned a lot but the only one thing I can't manage is transistors. I can't get to design a circuit with transistors as I jump into some sorts of error somewhere. Do you have any idea where to learn about this thing. The only way I learn abt them is when I experiment with them. Rest everything in electronics seems ok and I love to use IC's and design with them. Anyway keep it up guys.

KMoffett
08-23-2009, 06:06 PM
Schmitt triggers have hysteresis, that's why the unequal on/off times. I think there is a way to get an ~50% duty cycle with the one Schmitt NAND gate oscillator. I think you have two sets of feedback resistors/pots. A diode is placed in series with each set. the diodes point in opposite directions. The sets are paralleled. That way you can set the charge and discharge times independently.

Component values due vary a lot with the specific circuit. Just use what works. ;)

Ken

Søren
08-23-2009, 07:19 PM
Hi,

How do you calculate the frequency?
Study the 4093 datasheet (http://www.national.com/ds/CD/CD4093BC.pdf) - formula is on page 4 and the numbers you need is on pages 2 and 3.

For a quicker (somewhat less precise) approach, you can use:

fq = K/(R*C)

Where:
fq = Frequency in Hz
R = your resistor in Ohm
C = your capacitor in Farad
K = 1.2 (sometimes given as 1.3)


To get a M/S ratio of 1:1, perhaps try another 4093 (preferably from another manufacturer), or adjust the load on its output to get V_tH and V_tL symmetrical around Vdd/2.

You can use one of the six inverters in a 40106 the same way, as long as you don't need to gate the oscillator on/off.

Søren
08-23-2009, 07:35 PM
Hi,

Schmitt triggers have hysteresis, that's why the unequal on/off times.

Actually, it's not the hysteresis, but a thing happening when the threshold voltages aren't symmetrical around Vdd/2 (0V reference assumed).


I think there is a way to get an ~50% duty cycle with the one Schmitt NAND gate oscillator. I think you have two sets of feedback resistors/pots. A diode is placed in series with each set. the diodes point in opposite directions.Just adjust the load until you satisfy the above :)

The surefire way of getting 1:1 is of course using double the frequency and then divide it by 2 in a flip-flop.

jj_alukkas
08-24-2009, 09:55 AM
I think you have two sets of feedback resistors/pots.

Hmm.. no, I test for each frequency by varying the preset, if it seems good, replaces it with a resistor. I use no more than 1 preset at the same time.

To get a M/S ratio of 1:1, perhaps try another 4093 (preferably from another manufacturer), or adjust the load on its output to get V_tH and V_tL symmetrical around Vdd/2.

Thats right, cos the first time when I powered it up, I used LED's from the TIP's and found that one LED would only dim, not go off at all. Then I loaded them with bulbs and everything became normal.

The surefire way of getting 1:1 is of course using double the frequency and then divide it by 2 in a flip-flop.

Duty cycle is not of much importance in this app as the more weird it blinks, the more outstanding the circuit becomes. If it was some precision application we would have used precision timers. So it isnt really a problem right?

KMoffett
08-24-2009, 10:49 AM
I think you have two sets of feedback resistors/pots.
Hmm.. no, I test for each frequency by varying the preset, if it seems good, replaces it with a resistor. I use no more than 1 preset at the same time.

This is to set the duty cycle. These are paralleled, replacing the single "preset" timing resistor.
---/\/\/\--|<---

---/\/\/\-->|---
Because of the diodes' orientation, the upper resistor sets the charge time and the lower resistor sets the discharge time. So you can tweak your duty cycle. They will slightly affect the frequency, as you change one resistor value or the other.

Ken

jj_alukkas
08-24-2009, 02:26 PM
Alright, let me see..

jj_alukkas
08-26-2009, 02:29 PM
I did a quick work and made a small board...

http://forum.allaboutcircuits.com/attachment.php?attachmentid=11776&stc=1&d=1251295595

KMoffett
08-26-2009, 02:37 PM
Pretty and compact!...but does it work? ;)

Ken

jj_alukkas
08-26-2009, 05:00 PM
Ya it works fine and is cool cos of the heatsink. Without it, the TIP wont survive the 4A load. Thanks to you guys. I need to get a box to mount this thing in. I have limited it to 3 flashes to minimize the stunning effect.

KMoffett
08-26-2009, 05:15 PM
Congratulations on a successful project.

Ken

jj_alukkas
08-30-2009, 12:47 PM
It seems I have a small doubt to clarify. I tested the circuit with 2 10W bulbs and now before fixing it permenantly into my car, I find that the darlington is getting very hot with the heat sink on my car's circuit. One heat sink a bit more hotter. One bulb pair draws a bit over 3A and the other around ~4.15A totaling about 7.5A. I used the TIP127 taking into account that it can handle 60W and 5A. With the current load and the heating, will I need to replace the TIP with higher ones ? And also is the heatsink in the image ok to handle a hot TO-220 chip or will I need a bigger one? i needed to know it before I fix it permanently. Hot heatsink is not an issue for me as long as it doesn't burn out.

One other doubt I have is that the wire I used to tap the brake line temporarily, was not a good one. When I connect the 2 bulb pairs (7.5A) directly together through this wire, it heats up lightly, but when I used the circuit, it melted the plastic coating of the wire after about 30 seconds. Why didn't the same happen when i connect them directly? I have double checked for short circuit and the circuit carries a 10A fuse, so that wouldn't be an issue. Could the wire gauge used in the circuit be the problem also for heating the TIP?

jj_alukkas
08-31-2009, 04:43 PM
Well, I replaced Vcc and output wires with 1sq.mm industrial grade wire and now all most everything is solved except my doubt about the heatsink. One of the TIP heats up like soldering iron and even smells like it!, should I go for a TIP 137/147 or a larger heatsink? Its ok even with the current one, but I doubt if it would burn the TIP. If I swap the outputs, the other TIP heats too much. One of my load is obviousely heavier. I ran it for 1min continuous, nothing happened other than getting too hot, would it be a problem? One another possibility is that I ran the test with the engine off, so voltage at around over 11V.

P = VI

V(normal) x I(normal) = V(low) x I(high?)

Could that be a reason that since the supply voltage is low, the bulbs are drawing more current than it should and would become ok when the engine runs. I would only be using it on a running engine.

Other than all this the circuit works good with no problem.

KMoffett
08-31-2009, 06:40 PM
JJ,
Can you post a schematic, with values, of the circuit you're currently using.
Ken

jj_alukkas
08-31-2009, 07:54 PM
Same last circuit you posted. Components changed are:-
1. Transistors BC547
2. Darlington TIP127 for more current handling
3. Diodes 1N4007 It was what I had in hand.
4. And a resistor in the second oscillator from 10k to 100k to adjust frequency.
Load first output around 3.2A and second around 4.2A
Tested only with engine off, little over 11v maybe (didnt check)
Circuit is good and functional with light loads 21+10W bulbs tested at 13V with mild heating only on the 21W load (1.5A) from my test battery.
When tested on car 4.2A loaded TIP heats verymuch, 3.2A loaded one heats less.
With wires reversed, the pattern reverses.
Tested on car for abt 1 min with both loads, heating problem only, no burning.
I think it calls for a single TIP replacement or a bigger heatsink.

What would you suggest for this Ken?

http://forum.allaboutcircuits.com/attachment.php?attachmentid=11867&stc=1&d=1251748435

KMoffett
08-31-2009, 08:10 PM
Which transistor, Q2 or Q4, is generating the most heat?

ken

KMoffett
08-31-2009, 10:11 PM
I "think" I see the problem.
http://www.fairchildsemi.com/ds/TI%2FTIP127.pdf
The datasheet for the TIP127 shows Vce(SAT) at 2V@Ic=3A with 12mA base current and 4V@Ic=5A with 20mA base current. The base resistors R9 and R11 are only supplying about 4mA to turn on the transistors. The transistors are not being fully turned on, so a lot of power is turned into heat. Try changing R9 and R11 to 560Ω 1/2W.

Ken

jj_alukkas
09-01-2009, 07:59 AM
Oh man, you are great. I didnt even think about it. 0.5W's are not in stock, I'll get them in an hour and try again.

Q2 heats up too much if the 'heavier' load is connected to it and at that time, Q4 heats like a normal power transistor. If the loads are inversed, the pattern reverses.

jj_alukkas
09-01-2009, 04:42 PM
Ken, thanks a ton.. The TIP's now heat as it should as far as I know from my experience using TO-220 case devices. Also the biggest noticeable change was that with the 2k7, the bulb wouldn't light up similar to my test bulbs and the flashing was a bit dull, so I though it was due to the thermal latency of the bulb or overload. But now with the 560R, I understand why it happened. It flashes clean and merges as a single one as if nothing has happened. Anyway, a project ended well. I moulded out a PVC case and installed them onto my car. Anyway here is the final circuit. Thanks again Ken, you are simply great!

http://forum.allaboutcircuits.com/attachment.php?attachmentid=11886&stc=1&d=1251820960

KMoffett
09-01-2009, 04:47 PM
Thanks for the opportunity to play! ;)

Ken

jj_alukkas
09-01-2009, 04:56 PM
Hihi.. You're welcome..

dking61
09-24-2009, 05:50 PM
The flasher circuit looks real good for light bulbs. What modifications need to be made to do the same thing using LEDs? There isn't nearly the current requirement and I doubt that a heat sink will be necessary.

KMoffett
09-24-2009, 05:59 PM
Basically nothing, if you're using 12V automotive LED lamps. If you want to just use LEDs you have to add an appropriate dropping resistor in series with the LED to replace the light bulbs.

ken

bountyhunter
09-24-2009, 09:19 PM
Hi,

I made a brake light flasher for my car using a 555 and a 12v relay to blink two pairs of bulbs alternatively. It works fine but makes a lot of noise each time I step on the brake and it is quite irritating to hear that tic-tic at a traffic signal for a long time. So I designed a solid state circuit of the same. I managed to design the +ve side driver but I have no idea what to do for the -ve side to make it like N/c and N/o of a relay and blink the lights alternatively. Each pair would have a power consumption of 30W(max) I guess.How would I go on for the other side??

http://forum.allaboutcircuits.com/attachment.php?attachmentid=11207&stc=1&d=1249206296

That's the same basic circuit I use on my motorcycle. Without protection, it would live long because voltage transisnets will kill the 555 timer (guess how I know). I added protection, schematic attached.

bountyhunter
09-24-2009, 09:21 PM
The flasher circuit looks real good for light bulbs. What modifications need to be made to do the same thing using LEDs?.
The one I posted works equally well with either LEDs or incandescent bulbs.