How long... LED on time without resistors?

[Heckler]

Hey Group,

I am building an 8x8 LED matrix device. I am planning on NOT using resistors to limit current to the LED's. In one example I looked at it appears the on time is ~1.5 mS.

Therefore, I was wondering... How does one determine the amount of time that you can safely turn on the LED's without damaging them? I do not have the 8x8 LED modules yet, so I do not know whether they are common anode or cathode. I plan on using one PIC pin to control a ROW and one pin to control a Column.

I realize there are several factors at play here. Such as... supply voltage, number of LED pixels on at any given time (up to 8), etc.

In my case I plan on using 3 volt supply.
I have not determined whether to scan the display rows or columns. One would have only 1 LED lit at a time and the other would have from 1 to 8 LED's on at one time. (I would also be interested in suggestions on the correct way to scan the display?)

Thanks

[mackrackit]

At ~3 volts resistors are not required for the "normal" LED.

[Heckler]

Well that was EASY!!

What if I want to use a 5 Volt supply for testing?

I thought about just ramping up the "on" time until the brightness matches the brightness of a properly current limited LED of the same type.

[Archangel]

Well that was EASY!!

What if I want to use a 5 Volt supply for testing?

I thought about just ramping up the "on" time until the brightness matches the brightness of a properly current limited LED of the same type.

Hi Dwight,
I will give a metaphor to answer the question: how long can I hold a lit cigarette to my skin without a heat shield in between?
It is exactly the same scenario, you are abusing the components both PIC and LEDs, how long before complete burnout? I guess it depends upon the particular batch of parts. I "think" they do sell LEDs with internal resistors, or at least they did. FYI they also sell transistors with internal bias resistors too . . . call them digital transistors. I know it is NOT Uncommon for designers to overdrive components particularly when using PWM, but IMHO it is not the best practice. If this is a home project, and not life critical then who cares right? Then do it, if/when it fails you will fix it.


EDIT:
Here is a link to Mouser's search for 5 volt LEDs: http://www.mouser.com/Optoelectronics/LED-Indication/Standard-LED-Through-Hole/_/N-75pv5?P=1z0x8c7&Keyword=LED&FS=True

This link will allow you to select from whichever voltage and current values you like: http://www.mouser.com/Optoelectronics/LED-Indication/Standard-LED-Through-Hole/_/N-75pv5?Keyword=LED&FS=True

[rsocor01]

Hi,

No, you shouldn't use 5V in the PIC with a normal LED. You should use 3 or 2.5V for a normal LED. All the PIC digital pins have a small internal resistance that varies with the current drawn from the pin. A "Current Drawn" vs. "Internal Resistance" graph is ussually provided in the datasheet in the Electrical Characteristics section. That small internal resistance is what helps that the LED doesn't get fried.

Robert

[mackrackit]

Read this, it gets interesting...

http://www.picbasic.co.uk/forum/showthread.php?t=11575&page=1

[rsocor01]

Read this, it gets interesting...

http://www.picbasic.co.uk/forum/showthread.php?t=11575&page=1

Hmmmm, there was a heated debate in there :D. However, like DT mentioned in that thread, you can drive a regular LED (using a 3V supply) from a PIC without any problems, but if you connect that same LED directly to the power rail it will burn and it can burn your fingers (I say this from experience :o).

The reason it works on the PIC is the internal resistance of the pin in the PIC. For example, the following graph is taken from the 16F84A datasheet from the AC/DC Characteristic Graphs section.

5284

By calculating the slope in the dashed line it can be determined that the internal resistance for VOH = 3.0V is equal to ROH = 130 Ohms. Now, in the case the output is low, VOL = 0 then the VOL-IOL graph is

5285

Again, by calculating the slope of the curve it can be determined that the internal resistance ROL = 36 Ohms.

So, in the case that you are using a regulator of 3.0V you will have an equivalent PIC resistance of 130 Ohms; therefore, there is no need for an external resistor.

More detailed information about port characteristics can be found in the next book in chapter 3.

http://www.amazon.com/Designing-Embedded-Systems-Microcontrollers-Second/dp/1856177505/ref=sr_1_1?s=books&ie=UTF8&qid=1300073607&sr=1-1

[Heckler]

Thanks to everyone who replied...

The project is a very NON-Critical Boy Scout electronics merit badge project. I know I need to teach proper LED current limiting to the Scouts (and I will), as well as OHM's law, capacitors, inductors, transistors, resistors, etc. (oh yea, and Software concepts)

In this case due to time limitations and dollar limits on project costs... I am trying to keep the project as simple as possible. They need to be exposed to soldering and basic electronics. I want to "wet" their appetite for programmable microcontrollers. The kit will be an 8x8 LED matrix with a PIC 2-AA batteries and that's it. It will end up being a neckerchief slide that scrolls a scouting message. I may include two pushbuttons to include a simple game.

Similar to these excellent projects...

http://www.instructables.com/id/Mini-8x8-LED-matrix-Christmas-display-with-built-i/

http://tinkerlog.com/howto/64pixels/

[ardhuru]

Along with the other parameters you listed that would matter, its also important *how much time the LEDs stay off* between their 1.5 ms on state; the time they get to catch their breath, so to say. As long as this duty cycle is small enough the design should work fine even at 5 volts. I say this on the basis of a 4 digit 7 segment display I have integrated in a product; no resistors, running 22X7 the past 8 odd years, and no failures...yet.

Regards,

Anand

[ardhuru]

running 22X7 the past 8 odd years, and no failures...yet.


Sorry... "24X7".

Anand