control 8x8 matrix with 74HC595

[kikoou]

I'm sorry because I didn't mention that I'm using 20 matrix 8x8 to create my project. Thats why I need to use 74HC595 and thats why I need to use 3 pins from port C and 8 pins from B - no more.

[Lemosek]

Hello,
At may last post I suggest to see EX_EXPIO.c file.
You mast connect :
hc595 PIC
14 pin B2
11 pin B1
12 pin B0

Then if You include #include <74595.c>
You have nice function:
write_expanded_outputs (&data);
Variable data is 8 bit output state of 595.
Next You mast implement shift register to drive one column like 10000000 01000000 00100000 ...
So if you drive first column you must set 595 output (8 bit row) for this column.
This is simple

Best regards sorry for my English
R.L.

[Mike Walne]

OK. You've moved the goal posts for us, but it does not change what I and others have said before. You still have to learn how to code in C.

I'm assuming your schematic for one matrix has become:-

[John P]

Depending on the size of the digits and how brightly you want them to be lit, you might need to add a source driver equivalent to the 2803, so both the anodes and the cathodes would have power drivers.

If each digit is an 8x8 matrix and there are 20 digits, overall that's 1280 LEDs to drive, and as Mike Walne said, most economically you'd probably want to wire it as 32x40, but then you'd get a 1/32 duty cycle, and that isn't much. If that won't work, you'd have to think of alternative ways to wire it.

But with the 595's connected to the processor's SPI port, I don't think there's much danger of running out of time.

[Mike Walne]

Like John P hints there is a danger of you not having enough power, and like I said there is also a risk of you running out of processor time.

I don't want to waste time addressing irrelevant issues. Before I eleborate please answer the following questions:-

How many pins have you got available for your display?
Are you proposing using a string of 20 off 595's to drive your anodes?
What do you want your average lit LED current to be?

Mike

[asmboy]

Let me add that the code you submitted does NOT look like it owes its ANY of its provenance to the CCS framework l - rather - it looks like a lifted/translated MIKRO/HITEC/etc compiler source - barely adapted to follow a CCS header.

[kikoou]

thx everybody because you inspired me to do a little program to light the letter P. But my problem now is to send my data on 595 data pins and how can I use correctly the 595. My program is:

[Mike Walne]

You've moved the goal posts yet again. You started out by asking about coding for one digit. We answered your questions, so you then moved to 20 digits. Several of us have responded so you've returned to one digit. You are getting very infuriating.

The code you've posted does nothing, virtually all lines are commented out!!!!!!

I really was trying to help you. You posed a problem, I considered it and required answers to further questions. You've totally ignored me. I will do the same to you until I get a response. If you do not answer the questions in my previous post I will treat you with the same contempt that you are treating me.

Mike

[temtronic]

Yup, amazing how the 'project' grows in leaps and bounds...
The simple,easy,FAST solution is to use 'distributed processing',21 PICs..1 master 20 slaves. Each 'slave' controls one 8by8 LED 'unit'.
Circuit Cellar Ink did a PIC article on them a decade(+-) ago. Heck probably by a PIC for not more than a 595 these days.
Simple to cut code for,easy to hookup,expandable to any reasonable number of 'units'.Makes power supply considerations a joke(use a PC supply).Someone out 'there' , probably makes/sells such units,but really the code is simply 20 copies and one controller.Figure a weekend for the clones and couple of days for the master....2-3 days to debug, heck whole thing 'up and running' in less than 2 weeks.

[asmboy]

a most interesting thread .

I suspect that any one of the five professionals who have weighed in on this could do a fine job of designing the support hardware and code to do this
task. And that in the process at least 2 and probably 3 workable approaches would be evident. But as this gives every indication of being a school learning project , the level of starting knowledge seems too low - and the goal too broad and fluid - for anybody to take seriously.

perhaps the poster hopes to one day make his or her living doing electronic engineering of this sort - and i sincerely wish the best of luck in that-
but i believe that everybody who has answered so far - would be embarrassed to have shown so little effort while asking so much
of others. Especially if this is schoolwork.

just my 2 cents

[Mike Walne]

Kikoou

You were TOLD on the SAME day as your first post that the CCS example shows you, with a fully coded example, exactly HOW to drive a '595 for this project.

I will elaborate AFTER you show me the courtesy of answering these two questions.

(1) Are you proposing using a string of 20 off 595's or arranging the 8 * 8 LEDs in a 4 * 5 array?
(2) What is your proposed average lit LED current?

Arranging 20 off 595's in serial string is the simplistic method, a 4 * 5 array requires a little more thought.

I am assuming that you do have actual hardware in front of you, and that this is not just some simulation exercise. You can get a feel for the current using one of your 8 * 8 matricees, a 5V power supply and a handful of resistors. Connect any cathode to the supply negative, and any anode to the supply positive via a 3k3 resistor. Look at the lit LED and decide if its too dim, OK or too bright. Adjust the resistor value until OK, then deduce / measure the current.

I may even explain how the '595 driver works but you MUST do your bit FIRST. You're insisting on the '595. You may have to consider temtronic's suggestion after I've outlined some of the hurdles that litter the path to your target.

For a newbie you are in grave danger of alienating all the guys on the forum who are in a position to help you.

I, for one, am rapidly running out of patience.

Mike