8x64 led dot matrix sign board

Status
Not open for further replies.
M

maboelectro

Newbie level 6
Joined
Oct 29, 2007
Messages
12
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Visit site
Activity points
1,396
led matrix bc337

hi all

i have design 16*64 moving text led sign and i use multiplexing tequnic . the sign working good but i notice two problems .
first the brightness of the sign is low
second the brightness of columns that have few leds on is greater than columns that have mor leds on >
i know that my proplem is in driving circute but where excatly i dont know .
i use for drive 16 rows and 64 columns bc337 transistors and use cmos gates ic connected to 180 ohm resistors connected in turn to bases of that transistors .
and i use 12v/1a transformer for all and 7812/1a to drive led matrix and 7805/1a to drive control circute . notice : i use columns scan tequnic

thanks
 

led bc337 matrix

You have duty-cycle dimming problems. If you are scanning the 64 Columns your duty-cycle is only 1.5%. That's an LED which under normal circumstances is barely lighted. If you are scanning the 16 Rows your duty cycle is a little over 6%, which is a little better, but not great.

You overcome the lower effective power resulting from short duty-cycles by increasing the current to the individual LED's. Higher current, even for a short burst, will brighten the LED's.

Next the difference in brightness amongst different LED's is caused by not using a constant current source. You must have each LED in a row or column driven equally. That means to either sink or source each LED when ON with the same current limiting values. A good example of how it's done can be seen here:

**broken link removed**
 
Last edited by a moderator:
ledsign pdf


I guess you didn't like the answer in your previous same question thread.

Ok your logic is flawed, a 1/64 duty cycle is only suitable for an LCD (a very slow device) not LEDs they'll be very dim.
Here's the easy answer, to appear of normal brightness lets say 10ma at 100% duty your going to have to use 64x10ma or 640 ma per LED which won't be very healthy for the LED. Even a 1/16 duty (what you should be using) you'll still need about 60ma as the eye will respond to peak brightness. It'll be brighter but may not be as bright as you'd like but 60ma is about as much as most standard LEDs will tolerate (not at 100% duty though)
 

led matrixscan rows

dear blueroomelectronics

thank you very much for your reply and your notic.

first : you are absolutely right if i use row scan technique , but i use column scan technique , that means i have just one column will be active in any moment .
so if i try to over- drive all led in one column so the current will be 16*60 = 960 mA if all leds in one column were ON in same time .
but i think that my problem is in the duty cycle of column scan that equal 1/64
so even i over-drive the led to 60 mA >> the current through the led will be
60/64 = 0.937 mA .
so my analysis that if i want to get 20 mA in every led and with that duty cycle
the over-drive current will be 20*inverse duty cycle >> 20*64 = 1.280 A
what do you think ?

second : i already use transistors bc337 driven by cmos gates ICs through 180 ohm resistors .
thank you
 

16*64 led sign

Your column scan design will always be very dim.

60ma at 1:64 duty is still going to be dim. 20ma per LED at 1:64 is much worse.
No matter how you figure it that many LEDs are going to need a fair amount of current.
Properly done 60ma is more realistic for a 1:16 (it'll never be bright with 1:64) and we're back to the much higher current supply then you seem to want to use. 3.84A (64x0.6) That also means your column drivers have to pass 4A each and your row drivers a mere 100ma is fine.
20ma is the maximum CONTINUOUS current for a given LED, 60ma is PULSED current, some LEDs can even go up to 100ma pulsed.

Have you already built this thing?

Simple test next time, build a SINGLE LED circuit, run it at 1:64 duty and adjust the current till it appears bright BEFORE you solder down 1024 LEDs
 
Re: 16*64 led sign

dear blueroomelectronics

yes i already built this led sign and i try to improve its performance
but i dont know how ? i will post to my prif led matrix driver circuit avd i hope to help me in this .

thank you
 

Re: 16*64 led sign

hi maboelectro and all,
i think that scaning 64 colums to obtain 25Hz is not too mutch for britness, even if you scan the rows (16) you will not obtain a good britness ! so i suggest you two (2) solutions.
1 - use a double scan of rows, scan the the first 8 rows and the seconbd 8 rows at the same time, you will have 8*20ma = 160 ma, so it seems be too mutch for your LEDs, use 80ma for 1/25 s for each scan it will not destroys your LEDs, but not good for outdoor use, i use 70ma and all is ok from one year!.

2- use shift registers with buffers like HC595, one bit by one LED so you need 16*64/8 = 32*74HC595 ! with resistance of 150 ohms on each LED you will have a good britness.

you can use some intermediate solutions, so you will know what you need at what price ( the second solution is more expensive than the first).
excuse my english !
 
Re: 16*64 led sign



Hi,
I didn't get the first point can you elaborate it.
In fact double scan is impossible unless we split the led matrix into two 8*64 matrix.
 

16*64 led sign

25Hz is way too slow for an LED, it will flicker. 50Hz per LED is more like it.

As for using HS TTL it can sink 20ma (the HC is 35ma) per pin BUT only 75ma for the entire package so it's not an ideal solution. A ULN2803 is an ideal buffer and you can get your 60ma (must not be DC) per LED nice and bright.

The OPs problem is he built it without testing a prototype or understanding the problem. Wanting something to work won't make it work.
 

Re: 16*64 led sign

You are right blueroomelectronics for 25Hz, i used 40Hz for an 8*80 led dot matrix and there is no blinking.

Excuse me master_picengineer , i can't explaine more cause of my english vocabulary is not so rich !!! i will try to do so, you are right to say "split the led matrix into two 8*64 matrix", so we can use 8*74595 buffered shift registers to drive, for example 64*BC547 transistors, for the high 8 rows, and the same for the 8 low rows and scan at the same time the high and low 8*64 matrix, the rows will be drived by 16 ouput port of a uc for example or by 2 other 74595 wich drive some darlington power transistors, and this will take less output of the uc, all depend of what we need.
 

Re: 16*64 led sign

hi all
dear blueroomelectronics may be your version of eagle is older than mine i use eagle 4.09 version and it opens this file .
dear beny thanks for your reply , your idea is very good , but let us assume that i use double scan in the same time and there are two rows will be scanning and all the leds of there tow rows was in the ON situation so the current of one row is 70 * 64 = 4.480 A so tow rows 2*4.480 = 8.960 A , now if i use 3A transformer can this transformer give me this peak current even for a short time .
before i forget thanks xorcise for your reply

regards
 

Re: 16*64 led sign

I also don't have eagle and so I can not see your schematic in order to try and help you.

Previous suggestions to scan rows on two 8x64 matrices produces 1/8th (12.5%) duty cycle and plenty of brightness. If you drive the Output Enable lines on the driver ICs using PWM signal with period equal to row scan rate you end up with a simple brightness control with fade-to-black capability.

Good luck on your project. Mike
 

Re: 16*64 led sign

hi all

excuse me blueroomelectronics you right and i will post this file again , and i am sorry about this prif sch , i use tow shift registers 4094 to drive the tow rows and four chips of 4514 to drive the 64 columns .
first : dont constrate on the control circuits but on the circuit of driving the led matrix is it ok ? to get the same brigtnees to all of the leds , cause xorcise tell in his reply that i should use resistors in every column to geeting constant current source .

second : i have asked beny that let us assume that i use double scan in the same time and there are two rows will be scanning and all the leds of there tow rows was in the ON situation so the current of one row is 70 * 64 = 4.480 A so tow rows 2*4.480 = 8.960 A , now if i use 3A transformer can this transformer give me this peak current even for a short time ( duty cycle of scanning 8 rows ) what do you think ?

regards
 

Re: 16*64 led sign

It's a bad design, but I already knew that. You need PNP row drivers. How did you come up with the design?
Why are you using 12V for the LEDs?
Where are the LED current limiting resistors?
A single 4514 (4 to 16) should be for the row drivers not the column drivers.
The design wont work, why didn't you test the design first?

Time to start again, nothing will easily fix your design.
 

Re: 16*64 led sign

ok blueroomelectronics i will start again by using tow 8*64 led sign to get 16*64 led sign , and i will take in account all of the your notices .

but you dosnt reply me on the second question about the ability of the 3A transformer to give me about 9A for a short time ( duty cycle ) .

thanks
 

16*64 led sign

Transformers arn't my thing, when you design something design for worst case not expect something to work well under stress.

I was looking at liteon 5x7 displays, they specify 32ma @ 1:16 duty so you could build a 16x64 but remember 16 is the for a 4514 and high current darlington PNP drivers
5V is plenty for the LED supply; 12V just means more parts as you would need to buffer the PNP drivers.
As you will need CONSTANT 2.1A (not Peak unless you like shimmering displays) 32ma x 64 which is not a difficult design. Your darlington PNP transistors should be rated for at least 3A
Lastly the darlingtons will drop 1.2V the BC337 0.6V and the red LEDs 1.8 so you'll need a 43 ohm resistor for LED current limiting @5v.
 

Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…