×
INTELLIGENT WORK FORUMS
FOR ENGINEERING PROFESSIONALS

Log In

Come Join Us!

Are you an
Engineering professional?
Join Eng-Tips Forums!
  • Talk With Other Members
  • Be Notified Of Responses
    To Your Posts
  • Keyword Search
  • One-Click Access To Your
    Favorite Forums
  • Automated Signatures
    On Your Posts
  • Best Of All, It's Free!
  • Students Click Here

*Eng-Tips's functionality depends on members receiving e-mail. By joining you are opting in to receive e-mail.

Posting Guidelines

Promoting, selling, recruiting, coursework and thesis posting is forbidden.

Students Click Here

Jobs

LED clock too dim

LED clock too dim

LED clock too dim

(OP)
I'm building a digital clock using LED's in a 7 segment pattern. Each segment has around 35 LEDs and the height of a complete display is around 12 inches. i'm using two displays each, for hours, minutes and seconds. A PIC 16f877 sets the number to be displayed and a 74HC238 demultiplexer selects each display. Only one display is on at any given time, to save power. Problem is the LEDs are not bright enough. I tried using transistors to increase the current but it doesn't help, because its switching between each display so fast. Any ideas???

RE: LED clock too dim

Pull the data sheets, look at the current for the LED and the amount of current you can sink with the driver.

RE: LED clock too dim

If the overall design is power limited, then perhaps that is the problem.  If I understand your description (35 LEDS per segment, 7 segments per digit, 6 digits), then you have as many as 1470 LEDs to run (actually slightly less on at any one time since '88:88:88' isn't a valid time).

Multiplexing and scanning the digits can reduce wiring, but it does not save power because the average current sets the brightness.

RE: LED clock too dim

Remember that if you are using the LEDs in a pulse mode, you can boost the current far past the "rated" current, which is the max average current.


In other words, if you are running the LEDs such that they are only on 10% of the time (10% duty cycle) you should be able to run them at a much higher current. While the rules here have certain variations, with a 10% duty cycle you *should* be able to operate the LEDs at 100 ma with no problems.

NOTE: Your mileage may vary

Read this:

Quote:

LED Lamps may be operated in the pulsed mode. The absolute maximum ratings of LEDs have been determined theoretically and also by extensive reliability testing. Forward current, power dissipation, thermal resistance, and junction temperature are all interrelated in establishing absolute maximum ratings. In the pulsed mode, maximum tolerable limits should not exceed the LED junction temperature that would be reached by operating the LED at specified maximum continuous forward current. This correlation is obtained by establishing combinations of peak current and pulse width for various refresh rates and maintaining the maximum junction temperature as reached by operation at maximum continuous current.

Quoted from:
From: http://www.gilway.com/html/appl-ledlamps.htm

Andy

RE: LED clock too dim

If there really are more than a thousand LEDs, then I hope that they're wired in series - at least somewhat - so as to reduce the overall current requirements (while increasing the voltage requirements).  Doing so would make the overall design more practical.

RE: LED clock too dim

Hello,

What about super bright LEDs?

Tofflemire

RE: LED clock too dim

What's the cycle time on the multiplexing?  While going fast is required to minimize flicker, I think that going too fast will reduce the apparent intensity.  The cycle time on the display should be set for around 60 Hz.

TTFN

RE: LED clock too dim

Intensity is more about duty cycle duration and not speed (after speed passes integration to eliminate flicker, which is usually about 50 to 60 Hz depending on contrast ratio).


A 25% duty cycle at 100Hz should appear as bright as 25% at 1000 Hz.


On to the issue at hand, looking more closely at the application as stated at the beginning of the thread:


If I understand, you have a total of 6 digits, and each digit is 7 segments, and each segment is composed of 35 LEDs, which we'll assume are common LEDs with a 1.7 v drop each.


VE1BLL makes two very good points. One is that for a given brightness, you're not going to save power by multiplexing. If you multiplex you will have to increase the current to each segment to get the same brightness. And running the 35 of each segment in series would be the best design for a host of reasons.


Let's talk about duty cycle again. With 6 digits, and at least one ON and the other 5 OFF at all times, and when on each one is on for the same period, then you have a duty cycle of 16.67%. So if you are happy with the brightness of an LED at 10ma constant current, in this configuration, multiply that by 6 and operate the LEDS at 60ma. Since they are only on for 1/6th of the time, they will still only consume an average of 10ma.


Second, in running the 35 leds of each segment in series, you lower your parts count considerably (when running LEDs in parallel you should use a separate limiting resistor for each).

Now, with 35 in series, and assuming a 1.7v drop per LED then you have a total drop of 59.5 volts. Your voltage into the LEDs needs to be at least 59.5 volts, but will need to be much more if we are to get the 60ma per LED we want for good brightness.

35 LEDs running at 60ma pulse is 10ma average, accounting for the 16.67% duty cycle. If all 7 segments are lit, that's a total peak of 420ma, or 70ma average. With 6 digits, and all segments lit (assuming that each digit is on one at a time, with a full 16.67% duty cycle, as you indicated) then your max average current draw is 420ma for the display. (It's useful to note that if you have all digits and segments on (with a 100% duty cycle) but set the current for each segment at 10ma (instead of 60ma), you'd still only be drawing 420ma total). You could vary the pulse WIDTH (i.e. shorten the duty cycle) to eliminate the need for a series current limiting resistor keeping the average current in bounds for the LEDs you're using, depending on the voltage you'll be driving them at.

Regards,

Andy

RE: LED clock too dim

(OP)
Ok guys, I took your advice. I brought some new transistors, BD 679 to drive the LEDs anode and BD678 to drive the cathode. They can handle up to 4A with a gain of 750. I'm using a 1K resistor on the  their bases. This should fully saturate them. A 10 ohm resistor limits current to about 0.5A.

Forget 35 LEDs x 7 segments x 6 displays = 1470, I just want to get ONE LED to work correctly for now. I powered up the new circuit with the BD's, no difference. It seems like these transistors don't make a difference. I'm using a 7805 regulator (1A type) for power. 1A should be enough to drive ONE multiplexed LED.

Could it be a software problem?? Highly unlikely, cause my normal 0.5 inch 7 segments work perfectly. Should i revamp the entire circuit ? More importantly would this solve the problem ? thanks guys I really appreciate your advice    

RE: LED clock too dim

The gain of the transistor doesn't really matter if you're saturating it.  If you're trying to use them as simple on/off switches, move to FETs... that will reduce the on resistance and reduce power dissipation.  You're using a resistor to set the LED current, so there's not much point in using transistors.

Or, use the transistors to set the current and forget about the extra resistor (assuming the transistors can handle the current draw and power dissipation).

You haven't specified if the LEDs are wired in series or in parallel, but from your verbiage it sounds like they're wired in parallel.  If they're in series, the 5V regulator isn't going to give you enough voltage to drive that many LEDs in series.  The whole thing is confusing, as you mention a 10 ohm resistor limiting current to 0.5A, but that makes me think you have a single resistor in series with a number of paralleled LEDs... that's bad.

Post a schematic... a pic is worth a thousand words.

Dan
Owner
http://www.Hi-TecDesigns.com

RE: LED clock too dim

Prenavein - you haven't said what voltage your using. As macgyver suggests, post your schematic...

Andy

RE: LED clock too dim

(OP)
GUESS WHAT ??? Its working ! When those LEDs illuminated, I couldn't stop smiling. You guys are my mentors, thanks.

I swapped the NPN's and PNP's around. I added BC 547's from the 74HC238 to drive the bases of the PNP's. I've only used about 15 LEDs so far cause my power supply goes up to 30V. I'm sure if I add another 15 LEDs or so with the correct voltage, it will be bright enough. Strangely though, if one looks closely, the LEDs are illuminated slightly when they supposed to be off. I'm sure its a software problem, an error i probably made when i swapped the BD's.

I initially wanted to wire the LED's in parallel and use a PC power supply ( 30A on the 5V line). However, I now can see the benefits of wiring it in series. Which brings me to the problem of power supply. I need around 70V DC with 4A at least. I have 220V A.C. from the wall and i haven't worked with A.C. alot, so PWM seems challenging. I was thinking of stepping up a DC voltage, but then current becomes a problem. Is this a good idea ? I drew a schmatic, but i don't know how to post it.   YAY !!!! it works. he he he.

RE: LED clock too dim

1) The LED's are probably slightly illuminated due to leakage through the driver transistor. Not seeing your circuit, it's hard to say, but if a logic "1" drives the leds ON, then your logic "0" may not cclose enough to ground to fully shut off the transistor


2) Wow - 220 - well, simplest is to use a step down transformer.  You ought to be able to find one with a secondary of 56V AC, which full wave rectified and filtered with a big cap would give you around 72V DC.

http://www.toroid.com/standard_transformers/rectifier_transformers/rectifier_transformers_dual.htm



A

RE: LED clock too dim

(OP)
Hi there! i have a quick question. I connected 12 red LED's in series with a 100 ohm resistor to a 30V supply, no transistors. Its looks really bright. But according to my calculations, 30v - (12*1.7) = 20.4V. therefore 30v -20.4v=9.6V.  9.6v/100ohm= 0.096A. Thats the current that should flow throuh the LEDs, however i'm only getting 20mA. Where is the extra resistance coming from?? I thought when LEDs are foward bias resistance their is negiable.

RE: LED clock too dim

No, leds drop around 2v fwd biased. Maybe thats why your original setup was dim.

RE: LED clock too dim

2V might even be low.  Many of the superbrights run directly from two 2025 cells, so nearly a 3 V drop.

TTFN

RE: LED clock too dim

The drop depends very much on the composition of the LED. Old style red GaAs had a nominal drop of 1.7V - however, other colors have different drops, and the superbrights generally have much larger drops.

Measure the drop and see what it is.

I've generally found that red and yellow superbrights have a drop of around 2.1 V. Blue, green and white around 3V to 3.2V (remember that white is just a blue LED with phosphor on top).

Based on your 20ma, I'll venture the average drop is about 2.3V. If that's the case, you can pretty much toss the resistor.

Also, how are you measuring the current? The drop across the 100 ohm resistor?


A

RE: LED clock too dim

(OP)
Well, i pulled out one leg of the resistor and connected my ammeter in series. I measured the volt drop Of each LED, it 1.71V. Its not the high bright type, normal GaAs- model BL-B5134 bright red. I want to get 60mA through them when connected directly. So that when i multiplex them by 6, the average current will be 10mA.   

RE: LED clock too dim

If your ammeter is set to a sensitive range, it will introduce resistance into the circuit which may be large enough to affect your measurements. All multimeters have this problem. The better ones use lower burden resistance, so they have less influence on the circuit.

Try measuring the true resistance of the 100R resistor when it is out of circuit, then measure the volt drop across it when it is in circuit and apply Ohm's Law. By measuring the voltage across the resistor, your measurement makes virtually no difference to the operating conditions of the circuit. Post the results.

----------------------------------

If we learn from our mistakes,
I'm getting a great education!

RE: LED clock too dim

(OP)
I have made an amazing discovery thats gonna change the electronics industry forever. He he he ! I finally understand what you guys have been telling me over the past week. Myndex, i read over your first post carefully, all the answers are there.

Firstly, if i had 10 LEDs in series, that would create a volt drop of 17V. If i supplied 17V to these LEDs, there would be no need for a current limiting resistor, as these LEDs would bias themselves. If i wanted to increase the current through them, i would have to INCREASE THE VOLTAGE. Aha! At 25V, 10mA should flow through them, with no current limiting resistor. More voltage = more current. Of course using a resistor is a good design practice when high voltage allows more than 20mA through the LEDs.

 Sometimes i get caught up in the complex stuff and tend to forget the basics.
Is there any formula to calculate voltage, for the required current.     

RE: LED clock too dim

It's not a good idea to not use a resistor for current control. The leds will only be stable over a very small voltage range. If 1 should fail it will take out the rest in the line, as well as overloading the power supply.

RE: LED clock too dim

prenavin,

This is a VERY poor design practice... you need some manner of current control.  The graphs shown on spec sheets are nominal values, not guaranteed... the graph may show a 20mA draw at 'X' voltage (assuming you can even read those graphs to any real degree of accuracy), but the chances of all of the LEDs matching those specs is beyond slim towards none.

Dan
Owner
http://www.Hi-TecDesigns.com

RE: LED clock too dim

And don't forget that the current is proportional to eVd/kt

TTFN

RE: LED clock too dim

hey irstuff, how did you get that post that formula?

RE: LED clock too dim

That's one of the TGML commands, using the brackets and sup and \sup

TTFN

RE: LED clock too dim

Prenavin - connecting your ammeter in series adds resistance, thus you really can't accept that as a remotely accurate reading. This is due to the burden resistor in the ammeter.

Either measure the voltage across your current limit resistor and use that to calculate the current, OR use a non intrusive (clamp on) ammeter.

B+K has a nice DC clamp on milliammeter.


Andy

RE: LED clock too dim

cbarn: When LEDs fail, they open - if all are in series, then you get zip. Using a resistor in series IS the proper design practice.

BAD design practice is using a single resistor to limit a plurality of LEDs when those LEDS are all in parallel. (one resistor for a bunch of parallel LEDs spells disaster as the variations in the LEDs will likely lead to one "running away" and frying, which will then open, causing the current to increase across the remaining and pow pow pow. Kinda like semiconductor dominoes....)


Macgyver: Yes indeed. - the rated specs are subject to the fluctuations inherent in all semiconductor technologies.

To improve reliability, derate by a factor of two - thus if the max average current is 20ma, design for 10.


As a matter of practicality though one can dispense with the current limiting resistor if one has the means for varying the duty cycle to limit current. What's ultimately important is keeping the junction temp within it's nominal limits.

I mention this as Prenavin's states that the LEDs are only on about 16% of the time. He could operate the LEDs at 60ma, and the average current would be only 10, yet appear plenty bright.





Andy

RE: LED clock too dim

(OP)
Hi guys! I tried a new LED, its a L53LSRD made by KingBright. These babies are so cool, 8 mcd at only 2mA. They go all the way up to 30mA and the view angle is a whopping 60 degrees.
I was about to scrap the multiplexing idea and drive the LEDs on all the time, but the new LEDs have solved my problems. I hope they are cheap though, cause i need to buy over a thousand.
The old BL-B5134's that i was using could only handle 7mA max. at 5 mcd. That goes to show, DATASHEETS ARE VERY USFUL.
Back to my experiments !!!. I'll keep you posted on the latest developments.  

RE: LED clock too dim

Myndex: LEDs can fail high forward voltage or short (I have a few HPLMP4700's in my drawer with this failure mode). LEDs are not guarenteed to fail open circuit. For refrence, perhaps take a look at Luxeon LED failure mode RD25.pdf
We have used current control for our LED signage fault regulation (but it wasn't cheap!)

RE: LED clock too dim

mstock: Interesting - I've never had one fail like that, and I've been working with LEDs awhile (I design LED light up toys). Can you post a link for the PDFF file? I'd like to read it.


Cheers


Andy

Red Flag This Post

Please let us know here why this post is inappropriate. Reasons such as off-topic, duplicates, flames, illegal, vulgar, or students posting their homework.

Red Flag Submitted

Thank you for helping keep Eng-Tips Forums free from inappropriate posts.
The Eng-Tips staff will check this out and take appropriate action.

Reply To This Thread

Posting in the Eng-Tips forums is a member-only feature.

Click Here to join Eng-Tips and talk with other members!


Resources