Driving LED Dot Matrix Bright?

[arrow]

Hi

I am trying to light up a LED dot matrix. I am using a Microchip uC, and extending the outputs with the 74HCT595. What I find though, is that the LEDS are very dim. (7 LEDS are driven by roughly 20mA).

Can someone please recommend how to get the LED's to glow brightly?
I cant source the high current 595 where I am.

Any suggestions would be most welcome.

Regards
a.

 

[Nigel Goodwin]

Hi

I am trying to light up a LED dot matrix. I am using a Microchip uC, and extending the outputs with the 74HCT595. What I find though, is that the LEDS are very dim. (7 LEDS are driven by roughly 20mA).

Can someone please recommend how to get the LED's to glow brightly?
I cant source the high current 595 where I am.

Assuming you're multiplexing the LED's?, 20mA shared between 7 leds is less than 3mA each! - you need to provide 140mA for the 7 leds to give an average of 20mA to each one.

 

[arrow]

Hi Nigel

Yes I am multiplexing the LED's. And yes there are 7 LED's per column that is approximately 3ma to each LED.

Could you suggest a way/ circuit that I could boost the current to each column?

Thank you
a.

 

[Nigel Goodwin]

Hi Nigel

Yes I am multiplexing the LED's. And yes there are 7 LED's per column that is approximately 3ma to each LED.

Could you suggest a way/ circuit that I could boost the current to each column?

A driver transistor!, either bipolar or FET.

 

[Mike - K8LH]

If you're using the 'HC595 to 'sink' current to a column (common cathode), you might consider switching to a Micrel MIC5841 IC instead... It's a serial to parallel 8-bit 500-ma sinking driver (example below)...

It might help if you tell us about your configuration... Is it several 7-segment displays or 5x7 or 8x8 matrices? Common cathode or command anode columns? Are you 'scanning' columns or rows? LED duty cycle?

Good luck with your project... Regards, Mike

 

[mstechca]

Here's an interesting thought:

LED's in parallel tend to light brighter than LED's in series. when you think about it, you can treat them like resistors.

resistors in series add up. resistors in parallel give lower values.

 

[Nigel Goodwin]

Here's an interesting thought:

LED's in parallel tend to light brighter than LED's in series. when you think about it, you can treat them like resistors.

NO!!! - this is yet another STUPID suggestion, you can't treat LED's like resistors, and you shouldn't put them in parallel.

 

[arrow]

Hi

I use the 74HCT595 to source current. The columns are the anode, and the PIC pins are the cathode.

I have 7x5 dot matrix; 12 units (will be expanded to about 20).

I would appreciate any suggestions as to how to increase the brigthness.

Regards
a.

 

[Nigel Goodwin]

Hi

I use the 74HCT595 to source current. The columns are the anode, and the PIC pins are the cathode.

I have 7x5 dot matrix; 12 units (will be expanded to about 20).

I would appreciate any suggestions as to how to increase the brigthness.

As I said above, you've got to substantially increase the current, I would suggest driver transistors on both rows and columns (PNP to source, and NPN to sink) - a PIC won't handle the currents required either.

 

[arrow]

Hi Nigel

Yes I think I understand. Can you actually suggest a chip that would do that for me (source 140mA)?

Thank you
a.

 

[Nigel Goodwin]

Hi Nigel

Yes I think I understand. Can you actually suggest a chip that would do that for me (source 140mA)?

You might try looking at ULN series chips?, personally I'd probably use discrete transistors.

 

[arrow]

Hi Nigel

Great thank you.

Would you have a part number for an 8 pint ULN device?

If I were to use discrete components I would need something like 100 of them.

Just in case could you recommend a part number for a discrete component as well?

Thank you
a.

 

[Nigel Goodwin]

Hi Nigel

Great thank you.

Would you have a part number for an 8 pint ULN device?

They are not something I normally use, but have a look at under 'Power Source Drivers' and 'Power Sink Drivers'.

 

[Mike - K8LH]

I suspect you will want to 'scan' the seven columns for an overall 14.3% LED duty cycle... Overdrive the LEDs for adaquate brightness... I suggest using seven N-Channel MOSFET row drivers (or NPN Darlington, or....) capable of sinking the 100 LEDs in each row and P-Channel MOSFET (or PNP transistor, or UDN2981A) 'source' drivers for the columns...

I suspect you realize you'll need a 3 or 4 amp power supply to drive the 100 LED rows in your twenty 5x7 displays...

Good luck with your project... Regards, Mike

 

[Gorgon]

Would you have a part number for an 8 pint ULN device?

Hi,

ULN2003A has 7 drivers
ULN2803A has 8 drivers

If you need 100 drivers you need to think about segmenting the driver system. You will not have enough time to generate a flickerfree display with enough light.

TOK

 

[mstechca]

Here's an interesting thought:

LED's in parallel tend to light brighter than LED's in series. when you think about it, you can treat them like resistors.

NO!!! - this is yet another STUPID suggestion, you can't treat LED's like resistors, and you shouldn't put them in parallel.

I didn't make it a recommendation, it is just a thought. It was something I thought of to make the LED's brighter.

I hope no one used my thought as written.

If you do intend to use it, make sure there is a resistor in series with the arrangement or the LED(s) will malfunction.

and why can't you treat them like resistors?
Isn't that why two LED's in parallel light brighter than two LED's in series?
I mean, think about it. Let's pretend that each LED has one ohm of resistance. Two in series makes 2 ohms, which means current is in half. Two LED's in parallel will be about 1/2 an ohm. this means current is doubled.

the current is porportional to light intensity. That's why LED's can be treated like resistors. Just make sure you don't make the assumption that an LED has a LARGE resistance. In fact, it is safer to make the assumption that an LED has a SMALL resistance (under 1 ohm).

 

[Nigel Goodwin]

and why can't you treat them like resistors?

Because they are NOT resistors, and don't behave in any similar way to resistors!.

Isn't that why two LED's in parallel light brighter than two LED's in series?

They don't, and you SHOULDN'T place LED's in parallel.

I mean, think about it. Let's pretend that each LED has one ohm of resistance. Two in series makes 2 ohms, which means current is in half. Two LED's in parallel will be about 1/2 an ohm. this means current is doubled.

Doesn't apply, because they are NOT resistors.

the current is porportional to light intensity. That's why LED's can be treated like resistors. Just make sure you don't make the assumption that an LED has a LARGE resistance. In fact, it is safer to make the assumption that an LED has a SMALL resistance (under 1 ohm).

IT'S SAFER TO TREAT AN LED AS AN LED, UNDER NO CIRCUMSTANCE EVER TREAT THEM AS A RESISTOR - THE ONLY RESEMBLANCE BETWEEN THEM IS THE NUMBER OF WIRES!.

Doubling the current WILL increase the brightness of an LED (obviously, and is what I've said all along), but you do that by decreasing the value of it's current limiting resistor. Adding another LED in parallel (which you SHOULDN'T DO!) would split the available current between the two LED's, giving the same brightness split (unevenly, which is why you shouldn't do it!) across the two LED's.

 

[justDIY]

with the uln2803, what do I do with the COM connection?

the one example I found, was using the 2803 to drive relays, and they had the COM connected to V+ via a 30v zener... I'm planning on driving LEDs, like the original poster... so do I need to work about the kick-back protection (that COM and the internal diodes) are providing?

 

[arrow]

Hi

I have connected the ULN2803 into the circuit: each output of the 74HCT595 I connect to the input of the 2803. I connect the GRND of the 2803 but not the COM.
I put in a +5V on the 595, and measure what is comming off the 2803. I see about 500mV on the output of the 2803. (Clearly I cannot drive a single LED using this).

What am I doing wrong? How should I be connectin the 2803 to get the 500mA current I need to drive the LEDs?

Thank you
a.

 

[arrow]

Hi Nigel

I have just tried to source the ULN2981 with no success.

Can you please recommend a part number for the discrete transistors that will do the same job? (Source 140mA and driven by the 74HCT595).

I appreciate your help.
Regards
a.