Max7219 LED Clock ghost image

[muthumani]

Created six digit LED seven segment display with 8 leds per segment. Totally used 8*7*6 = 336 LED's. Connected all 'A' segment, 'B' .... 'G' and connected to MAX7219.
Used DS1307 (RTC) and read using 89s52. The problem im facing is if i enabled set segment 1 with value 1 means im able to see 1 in all the other seven segments also.
Connected 1k resistor to base of transistor (BC148 for driving segment) and 470k to base and grounded (This is because 4 leds are serially connected, so need to drive +12V).
Connected 1k resistor to base of transistor (SK100 for enabling segment) and 470k to base and +12V.

Please suggest me to come out of this ghost image issue... Thanks

 

[betwixt]

The MAX7219 normally directly drives the LEDs and only needs the common LED connection driving for high current applications. Can you post the schematic please. It sounds like you have the drive signals inverted so only one digit turns off at a time instead of only one turning on.

Brian.

 

[D.A.(Tony)Stewart]

Your Diode Mux is not mux'ing.. i.e. each digit should be enabled exclusively and no crosstalk from digit to digit. Check that your Anode driver is turning off. Use a moderate pull down R of 10K across the PNP open collector of common anode digit driver thats pullup when enabled and floats when disabled.. Ensure the digit is not selected with stray cable or circuit capacitance. when not selected.

 

[muthumani]

Sunny i replaced 470K resistance with 10K but still facing the same problem.... Any idea about how MAX7219 refresh the segments ?. Is there any other technique to pullup ?..

 

[D.A.(Tony)Stewart]

As I recall you have digits pulled up and segments pulled down (CA LED) so you have the R in the wrong direction... YOu want to dis-charge the Digit CA to ground, when not selected. or a mid V+ if avail. So that's 10KΩ from Anode to ground.
If that still fails, something is faulty.
Normally you don't need this.. but you must be mux'ing fairly quick for the miller cap to store enough to hold those segments at 10% of nom. current... judging by the dim ghost.

(did I see this same question on the stack site?) those amateurs...

 

[muthumani]

Attached a portion of the cicruit...

 

[D.A.(Tony)Stewart]

OK My assumptions were wrong... YOu are using a CC (common cathode driver for common mode LEDs with no means of biasing the transistors off properly. They only spec Sink current on the segment outputs and Source current for the cathode digit out... and you are using 12V PNP biased at 5V off.and 0V on...


What voltage do the LED's need with 8 LED's per segment? what P/N? or color? current vs V specs?
Why does schematic show only 4 in parallel with 5 with segment? Should that be 2p4s array of LEDs per segment?

 

[muthumani]

In the circuit unfortunately i put 5 leds. Actually used 4 LEDs. 4 leds X 3V = 12V. Paralleled two sets of 4 leds. Used Red color LED from Everlight.
Should that be 2p4s array of LEDs per segment? => Yes.
After changing to 10K i wrote data 0x00 (Blanking) to digit 2,3,4,5 and 6. Wrote data from 0 to 9 with a delay of 1 second to digit 1. As a result i was able to see all digit liting with the data written to digit 1...!!!!....