LED multiplexing tutorial.

Nigel ya might as well link to it directly
i just saw it , its pretty cool !!

Mike said:

I believe you're going to need "source" drivers for the LED anodes. The STP16C596 and the ULN2803A are both "sink" drivers.

If you can exchange your matrix displays for common anode versions then you could use the STP16C596 16-bit "sink" driver for the LED cathodes along with a UDN2981A "source" driver (instead of the ULN2803A) to drive the common anodes.

Mike

Click to expand...

Mike
Unfortunatley I'am stuck with the cc matrix's.so my stp16c596 is no good what would you suggest use two 74hc595 in place of it and still use the 2803a?

Nigel Goodwin said:

Well, this has sparked some lively debate!

OK, I'll explain further.

My original plan was to use source and sink drivers in order to push more current through the LED's - and the original circuit I started drawing incorporated that - but I thought it was getting a little away from the 'simple' philosophy of the tutorials.

While doing mental calculations of suitable resistor values I realised it's quite plausible using just the PIC to drive the displays, so I dropped the drivers and increased the values of the current limiting resistors to 150 ohms.

Due to the modular nature of my tutorials, a further thought was to optionally add driver boards between the PIC and the display - I was thinking of two seperate boards, a 'sink' board and a 'source' board - both of which could also be used for other purposes (again reinforcing the modular nature of the boards).

Click to expand...

Amen!

BTW, I've just assembled a quick program that puts all the LED's ON, and measured the battery current - it was only 62mA? - which seems rather low?, but the LED's seem plenty bright enough. This was on my meter at work, I wonder is the current ranges aren't reading properly? - it very rarely gets used on current, I'll measure again with my meter at home.

Click to expand...

that does seem low. Are you certain the duty cycle for each row is 1/8 (minus a little for blanking during row change)? based on your numbers, it should be about double that. with some voltage drop from the pic, I'd expect more like 100-110 mA.

philba said:

that does seem low. Are you certain the duty cycle for each row is 1/8 (minus a little for blanking during row change)? based on your numbers, it should be about double that. with some voltage drop from the pic, I'd expect more like 100-110 mA.

Click to expand...

Just measured it again, at home this time, still 62mA?.

Certainly the multiplexing was 1/8th the other day when I scoped it, I'll try it again and see what's going on!. Might not be till Monday now though.

OK, I've added two more sections to the tutorial, the first adds full text capability, and the second uses that to scroll a text message across the display.

philba said:

that does seem low. Are you certain the duty cycle for each row is 1/8 (minus a little for blanking during row change)? based on your numbers, it should be about double that. with some voltage drop from the pic, I'd expect more like 100-110 mA.

Click to expand...

OK, I've got the scope round it - with all 64 LED's lit, the PIC pins are only pulling the bottom of the LED's down to 1.65V above 0V, there's 1.75V dropped across the LED's, and 1.6V across the 150 ohm resistors. So this is about 10mA peak through the LED's.

As you say, adding sink drivers would easily increase that, but it's bright enough as it is.

LED capacitor

I was just wondering how the multiplexing of the LED matrix would affect the system power supply? Wouldn't such frequent high current draw(sink) feed noise to the Vcc? Have anyone experimented with that? I'm using an ADC in my system with multiplexed leds and wondering if I shouldn't add a parallel capacitor with each LED?

62mA for all LEDs lit up is about what I would expect. At any point there are only 8 leds going, which is about <8mA per LED.

The PIC output pin sink driver is a NFET with similar properties to any NFET that you would use in a PSU design etc, Rds-on and they are really rated for average power disipation so you can run quite a high current IF it is a low duty cycle.

Just because Microchip are conservative (and have to guarantee their products) they like to rate them with the convenient "absolute max current" figures. But in reality they are a power FET like any other type.

Juts watch out for the new "shrunken die" PICs have much smaller and weaker output FETs than the old large die chips. If you look in the electrical specs in the datasheet your PIC will be one of these 2 types;

Max current pin=25mA, max for ALL ports 200mA (old PIC with large FETs)
Max current pin=25mA, max for all ports 90mA (new shrunken PIC with weak FETs)

The old large die has bigger FETs on bigger silicon with more heat dissipation, larger tracks, lower Rds-on of the FETs themselves etc etc. But the new smaller die PICs are cheaper and use less power.

The PIC power supply should be fine to source 62mA (max) pulsed, i've got some 22uF tantalums for the PIC power pins to use for PIC direct multiplexing. Don't expect a 0.1uF cap to do the job.