Hi All

Let me say first that I havent used mosfets much before so all of my information is off the net

I have a 7 x 11 led array, each of the leds have a series resistor and the anodes of each column are commoned up as are the cathodes of each row.
In theory this will eventually give me addressable control of each led.

Because the rows and columns draw more than allowable current for the PIC micro that I am using (16F877A) I am switching the anodes only for now with IRF9630 p channel mosfets. All of the cathodes are simply commoned to the gnd rail for now.

I am using the micro to switch the gates of the p mosfets via a 800 ohm series resistor.

For the most part this is working but I have one row who's operation is inverted. I have hunted and hunted and cannot find what is causing it.
Have connected simple leds direct to the micro and it is working fine, have even swapped out the mosfet and still the problem remains. Any ideas?

All of the sources of these 7 9630's are connected together to a 5v power source and the drains are connected to the led columns.

Another thing I have noticed is that the mosfets seem to be switching on with a logic 0 from the micro rather than a logic 1, I will have to confirm this.

Thanks in advance for your help

Bruce

 

Using p-channel mosfets will require the signal from the PICmicro to be low to turn them on. The source is typically connected to +5V and the gate is driven by the micro. The gate voltage must be below the source by some specified voltage for the p-channel mosfet to turn on, i.e. the output must be logic 0.

You can invert the logic if you use n-channel and place them on the ground side of the LED. Here the source is at GND and the gate would have to be higher than the source by some specified voltage to turn it on, i.e. the output must be logic 1.

 

Are you removing the "inverted" row from the PICmicro circuit to verify that they are working? I would assume that they are not lighting at all while they are connected to the micro.

Maybe the LEDs in the row are placed in backwards. Can you verify or is it a closed package? Hopefully that is not the problem, but it's not out of the question.

 

Hi, Thanks for your suggestions stayathome.

I wasnt aware that the p mosfets needed a logic 0 to turn them on, not a problem I will just NOT my logic just before output.
I do have some N channel mosfets that I plan to use to 'turn on' the cathodes and will use them in conjunction with the p channel fets on the anodes.

In response to your other question I NOT 255 (00000000) and send it to the output port on my micro and all of the seven columns are on except the 'faulty' column, I then NOT 0 (11111111) and send it to the output port of my micro and all of the columns are off except for the 'faulty' column.

Very Puzzling.

 

While looking over my veroboard again I saw the tiniest dot. Turns out there was a small track between a couple of pins on the fet (gate and drain I think)

Thanks again for your help