Dot Matrix improvements

Go To Last Post

19 posts / 0 new

Author

Message

theFloe

Level: Hangaround

Joined: Sat. Dec 14, 2002

Posts: 241 View posts

Location: Austria, Europe

Posted by theFloe: Sun. Nov 21, 2004 - 06:29 PM

Hi,

I've build a Clock using a 5cm high 5x7 Dot Matrix Display. The problem is that the LED's do not shine enough. So i thought to put a lower resistor into the circuit. First i used a 330 Ohm resisitor (about 10mA) and the i put a 150 Ohm (about 20mA). The confusing thing is, that the led's with 150 Ohm don't shine much brighter than the others with 330 Ohm. So it has to have something todo with my update routine. I played around with the timing values and could improve the bightness a little bit, but there should be more to get out of this display. Any ideas or any knowledge about this problem?

Hava a look at my web page -> http://www.tobiscorner.at.tf

Tags:

AVR Microcontrollers, megaAVR and tinyAVR

RoBSki!

Level: Resident

Joined: Mon. Mar 15, 2004

Posts: 705 View posts

Location: the Netherlands, Europe

Posted by RoBSki!: Sun. Nov 21, 2004 - 07:34 PM

I have the same problem. I cannot get bright light from it.
I have tested another display yesterday, and this one had bright light, so it depends on the type display. It's the sensitivity of the leds. Try a few brands displays.

snafles

Level: Wannabe

Joined: Sun. Jan 18, 2004

Posts: 66 View posts

Posted by snafles: Sun. Nov 21, 2004 - 08:44 PM

I think if you check out some data sheet on the LEDs, you may be able to drive them at a much higher voltage than normal. The amount of time one LED is active is so small, depending on your implementating, you may not even be giving the LEDs enough time to turn on. When I made a moving light display some time ago, I was running my matrix off 12V.
I was running 4bit PWM per bit, so on time was extremely low.

Make some duty cycle calculations, see what you come up with.

Vern

theFloe

Level: Hangaround

Joined: Sat. Dec 14, 2002

Posts: 241 View posts

Location: Austria, Europe

Posted by theFloe: Sun. Nov 21, 2004 - 09:47 PM

Thank you for your suggestions. I think i've to play a little more with the duty cycle. See http://tobiscorner.shellmaster.de/test.avi for my display... scrolling the time.

Hava a look at my web page -> http://www.tobiscorner.at.tf

bobgardner

Level: 10k+ Postman

Joined: Wed. Sep 4, 2002

Posts: 26609 View posts

Location: Orlando Florida

Posted by bobgardner: Sun. Nov 21, 2004 - 09:53 PM

Maybe the 5v is sagging when you use the 150 ohm resistors. Check that with a voltmeter?

Imagecraft compiler user

theFloe

Level: Hangaround

Joined: Sat. Dec 14, 2002

Posts: 241 View posts

Location: Austria, Europe

Posted by theFloe: Sun. Nov 21, 2004 - 10:04 PM

The 5 volts are still there. I could to some improvements with a new timing but there has to be more. The 330 Ohm & 5V in a constant to multiplexed row are much brighter. What timing values (refresh time for pixel & row or column) did you use?

Hava a look at my web page -> http://www.tobiscorner.at.tf

glitch

Level: Raving Lunatic

Joined: Sat. Jan 12, 2002

Posts: 8620 View posts

Location: Canada

Posted by glitch: Sun. Nov 21, 2004 - 10:21 PM

Duty cycle is a big thing here, in a scanned LED matrix. Scan speed has an effect as well, though to a lesser degree.

If you are scanning 40 columns (assuming common column config) your duty is 1/40 and if you use a resistor it will need to be approximately 1/40 the value it would be if you were driving the LED with a constant source. This should get you in the ballpark. Many scanned displays will not use a resistor at all, since they can guarantee that the average current seen by the LED will never exceed it's maximum. (you need to be careful here, as if the AVR freezes you can burn out some LED's because you will be feeding them too much current)

Slowing down your scan will brighten the display as well, but it will also start producing a flicker, if you go too low.

Writing code is like having sex.... make one little mistake, and you're supporting it for life.

big bang

Level: Wannabe

Joined: Tue. Dec 25, 2001

Posts: 69 View posts

Location: Antwerp / Belgium

Posted by big bang: Sun. Nov 21, 2004 - 10:21 PM

Hi,

A while ago (about 2 years) I made a 7*12 dot led matrix moving message display
controlled by a AT90S8515. I was just playing around and curious. 8)

Because I couldn't find 7*5 dot matrix displays with a high light output directly I made the matrix myself. I had a few hundred medium bright leds in stock left so no problem there.
I think these leds had an output of 45 to 70mcd @ 10mA and 24° beam width. The display was clearly visible at a distance of about 20m at daylight with a display of only 50*90mm.

When you scan a matrix like that, your led current has to be higher then 10mA to get a bright output. Because the matrix is 12 dots long, each led can be driven with a duty_cycle of max. 1/12 so the led current must (theoretically) be 12 times higher.
For this, always check the datasheets of your dotmatrixdisplay or leds.
Normally they give a max. continuous and a max. pulsed current at a certain duty_cycle.

In my case i used a 47ohm resistor (and a transistorswitch) and a 5VDC supply to drive the leds which gives a current of about (5V-1.7V)/47=70mA.

Remember the higher the number of dots to be scanned, the lower the duty_cycle and the lower the light output.

For bright displays I rather tend to build them using discrete leds. I use 7*5 dot matrix displays for 'normal' purposes. If you use leds you can pick your favorite color (even multicolor) and the needed light output. There are leds available with an output of a few 100 to 1000's mcd @ 20mA. (And for very extreme cases you can use 35 diode lasers lined up in a 7*5 matrix to burn someone's eyes out :twisted: )

Hope this helps :lol:

Regards,

- - big bang - -

Take a good look around. Admire our world, our galaxy, the whole universe. Then you just feel there must be a
'Theory of everything'

Maybe, someday, the human race is ready to discover it.

big bang

Level: Wannabe

Joined: Tue. Dec 25, 2001

Posts: 69 View posts

Location: Antwerp / Belgium

Posted by big bang: Sun. Nov 21, 2004 - 10:24 PM

oeps,

Glitch is a little quicker... :lol:

Regards,

- - big bang - -

Take a good look around. Admire our world, our galaxy, the whole universe. Then you just feel there must be a
'Theory of everything'

Maybe, someday, the human race is ready to discover it.

glitch

Level: Raving Lunatic

Joined: Sat. Jan 12, 2002

Posts: 8620 View posts

Location: Canada

#10

Posted by glitch: Sun. Nov 21, 2004 - 10:35 PM

lol,

I did forget one thing though. Frequency also has an impact on the resistor value calculation, basically the slower the scan, the higher the resistance required. Though for a moderately sized display, and an overall refresh rate of 15-30Hz the above should hold (meaning a 1.2Khz scan rate for a 40 col matrix).

Writing code is like having sex.... make one little mistake, and you're supporting it for life.

big bang

Level: Wannabe

Joined: Tue. Dec 25, 2001

Posts: 69 View posts

Location: Antwerp / Belgium

#11

Posted by big bang: Sun. Nov 21, 2004 - 10:42 PM

Me again...

theFloe wrote:

What timing values (refresh time for pixel & row or column) did you use?

Column refresh rate: 2ms
Complete display refresh: 12*2ms=24ms or 40Hz.

Regards:

- - big bang - -

Take a good look around. Admire our world, our galaxy, the whole universe. Then you just feel there must be a
'Theory of everything'

Maybe, someday, the human race is ready to discover it.

theFloe

Level: Hangaround

Joined: Sat. Dec 14, 2002

Posts: 241 View posts

Location: Austria, Europe

#12

Posted by theFloe: Mon. Nov 22, 2004 - 06:18 AM

Thank you! I'll short all resistors this afternoon! Let's see if the display burns or turns into bright light ;-)

Hava a look at my web page -> http://www.tobiscorner.at.tf

RoBSki!

Level: Resident

Joined: Mon. Mar 15, 2004

Posts: 705 View posts

Location: the Netherlands, Europe

#13

Posted by RoBSki!: Mon. Nov 22, 2004 - 06:39 PM

When I short a resistor, one (red) dot turns yellow kinda colour. :twisted:

theFloe

Level: Hangaround

Joined: Sat. Dec 14, 2002

Posts: 241 View posts

Location: Austria, Europe

#14

Posted by theFloe: Mon. Nov 22, 2004 - 07:10 PM

Thats looking like toooo much current. I haven't shoted the resistors at my clock because during programming or other failures the Matrix LED's can light randomly. So without resistors the leds will burn.

Hava a look at my web page -> http://www.tobiscorner.at.tf

big bang

Level: Wannabe

Joined: Tue. Dec 25, 2001

Posts: 69 View posts

Location: Antwerp / Belgium

#15

Posted by big bang: Mon. Nov 22, 2004 - 10:40 PM

Just added some info:

check
Kingbright dot matrices
and
All Kingbright products
and click "All products"

This gives a good impression that real bright dot matrices are not so easily found.

Regards,

- - big bang - -

Take a good look around. Admire our world, our galaxy, the whole universe. Then you just feel there must be a
'Theory of everything'

Maybe, someday, the human race is ready to discover it.

voxel79

Level: Wannabe

Joined: Mon. Nov 10, 2003

Posts: 66 View posts

#16

Posted by voxel79: Wed. Dec 29, 2004 - 08:10 PM

Ok I'v got to ask...

I'm working on a 72x16 leddisplay, and it works fine, at least to some level. Everything is bright and all but I have trouble with 'ghostleds'.

I use uln2803s to drive the kathodes and bd436s that are triggered to saturation by a bc337 to drive the anodes.

leds intended to light up are bright (and i mean BRIGHT, still seing white dots when I close my eyes), but allmost all leds above and below the lit led, especially the ones close to it, are lit too, visible but only very little. it must be my transistorsetup that leaks current somewhere. I've been thinking and trying and burning stuff for weeks now and I can't eliminate the sideeffect. If anyone has been successfull with getting leds light up bright without the ghostledstuff, could you share the idea of how to properly drive the anodes?

Many thanks from me and my poor transistors to the one who can help me ;)

John_A_Brown

Level: Raving Lunatic

Joined: Sat. Sep 20, 2003

Posts: 6192 View posts

Location: Underneath the Cotswolds escarpment, England.

#17

Posted by John_A_Brown: Thu. Dec 30, 2004 - 09:30 AM

Does your code blank the LEDs when switching? This is a typical sympton of failure to blank the display when switching rows (or columns, depending on your multiplexing scheme).

Four legs good, two legs bad, three legs stable.

voxel79

Level: Wannabe

Joined: Mon. Nov 10, 2003

Posts: 66 View posts

#18

Posted by voxel79: Thu. Dec 30, 2004 - 10:27 PM

john,

No I did not but that is not necessary. the 595-ic is a 8bit serial-in parallel-out latched shift register, with seperate shift an latch controllines. I do use the latch signal to first load all new data in the shiftregisters (driving the ledcolumns) and after updating the shiftregisters of the ledrows, i tell the latches to take the new data.

Evenso I tried blanking out first but it does not change. and allmost all leds are lit in a vertical row, meaning if the topled would be lit then the middle led or even the bottom led would be very slightly lit too.

John_A_Brown

Level: Raving Lunatic

Joined: Sat. Sep 20, 2003

Posts: 6192 View posts

Location: Underneath the Cotswolds escarpment, England.

#19

Posted by John_A_Brown: Fri. Dec 31, 2004 - 09:15 AM

Can you post a schematic(or a fragment, we don't need the whole 72 x 16)?
Maybe a bit of code as well?

Four legs good, two legs bad, three legs stable.

Jump To

Search form

Dot Matrix improvements

Tags: