The data sheet seems to show that the IO expander is capable of supplying sufficient current ONLY during the write pulse. Otherwise, the current supplied from the expander is not enough (~100 uA). At first glance I thought that each output port provides around 25 mA.
I connected LEDs (15-20 mA Forward current, green and red with a 2V and 2.1V Vf) to the IOs and expected to get adequate brightness. That, however, didn't work.
Hi Aleksandr,
The data sheet seems to show that the IO expander is capable of supplying sufficient current ONLY during the write pulse.
The write pulse is on the I2C bus. When you save only supplying enough current during a write pulse, I don't understand how this would effect the output of the IO expander?
Otherwise, the current supplied from the expander is not enough (~100 uA).
Correct, there is a built in internal pull-up current source with strength of 100uA. If you are using LED's, you may need additional pull-up connections. An example is shown in the typical application in the datasheet.
Note that the P-port output is responsible for setting the output node to GND by pulling it to GND through its push/pull driver circuitry. Depending on the value of the pull-up resistor, in this case 2kohm, the GPO may experience more or less current draw. Careful calculation of this 2kohm pull-up is required to avoid stressing out the output pin of the GPO causing it to draw more current than the absolute maximum condition (50mA).
Do you have a schematic to show how you are connecting your LED's?
Regards,
Tyler
I'm looking at the block diagram and this is not clear why we have 100mA current source here?
As I understand this correctly the writing pulse (one clock pulse) affects the output.
This is gonna be a battery powered device where I'm driving anodes of these bi-color LEDs. So the TCA9535 looks as a good option. For this alternative however, the unused I/O pins should be outputs for avoiding pull-ups and pull-downs on it, right?
