I'm using this IC.

This is its Hardware Checklist.

I'm using DVDDH as 1.8V (which is VDDIO)

My question is, on page 55 of the datasheet, Table 6-4, for the LED1 and LED2 pins,

it is mentioned as "All digital output pins, expect LED1 and LED2? The output characteristics are not specified in the Datasheet. "All digital output pins, except LED1,LED2"(p.55)

What is the meaning of this?

Also, If Buffer type changed from CMOS Output to others at DVDDH=1.8V, could you please also tell me the buffer type

The Hardware design check list states that when *DVDDH = 1.8V, a MOSFET should be connected. However, Vol and Voh of LED1 and LED2 were not mentioned.

If using RJ45 jacks with integrated LEDs and 1.8V VDDIO, a level shifting is required from LED 3.3V to 1.8V. In this case, a BJT or MOSFET or a level shifting device can be used.


2 Answers 2


1.8 V is not enough to drive a LED; even red LEDs have a forward voltage of 2 V. This is why the datasheet tells you that you need level shifting, and not just a buffer.

Comparing tables 6-4 and 6-5, it appears that the LED1/2 pins have the same drive strength as other digital outputs. A LED current specification for 1.8 V is missing only because it would not make sense.

  • \$\begingroup\$ Suppose, we have put a voltage level shifter btween PINS and LEDS. But we are not sure the VoH of LED PINS could satisfy the ViH of the voltage level shifter. Any idea on that \$\endgroup\$
    – user220456
    Commented Oct 13, 2022 at 8:19
  • \$\begingroup\$ "DVDDH (digital I/O) = 1.8V, IOH (min) = 13mA All digital output pins, except LED1, LED2" mentioned in the datasheet is for LED current ratings not applicable for VoH minimum voltage. What does it mean when it is said, "mentioned in the datasheet is for LED current ratings not applicable for VoH minimum voltage." Can you please clarify ? \$\endgroup\$
    – user220456
    Commented Oct 13, 2022 at 8:28
  • \$\begingroup\$ @Newbie As I said, it's likely that the specifications of the other digital outputs apply. \$\endgroup\$
    – CL.
    Commented Oct 13, 2022 at 12:13
  • \$\begingroup\$ Still unable to understand. Could you please clarify a little bit more in simpler terms? \$\endgroup\$
    – user220456
    Commented Oct 13, 2022 at 16:43
  • \$\begingroup\$ Just see table 6-4. \$\endgroup\$
    – CL.
    Commented Oct 14, 2022 at 13:55

Let's have a look at the tables in question, from the datasheet:

enter image description here

What they're saying is that the pins aren't useable as LED outputs by themselves when DVDDH is 1.8V, so they don't give a Ioh/Iol rating in that condition. Why? Because LEDs need higher forward voltage, while the I/O pin isn't tolerant of a higher voltage.

In other words it's an ordinary I/O pad, not a true open-drain pad. They don't come right out and say it, but if you use DVDDH at 1.8V and pull up the LED to 2.5 or 3.3V, the I/O pad will leak through the LED to the higher voltage. That's why needs a driver / level shifter when running at that lower voltage.

(I had this argument at a previous company. 'Engineer' thought it was ok to drive a 5V-tied LEDs with a 3.3V buffer IC. They wondered why the LEDs continued to glow dimly when they were supposed to be 'off'. After about an hour of whiteboard time explaining leakage, I introduced them to the 2N7002 n-FET.)

They clarify this further in Section 9.0:

enter image description here

The subtle detail they're glossing over: they count on the LED's forward voltage in the 'strap low' case for that 1k pull-down to work. After the strap is read, then the pin gets driven stiffly high or low, overriding the 1k.

If you need to use 1.8V DVDDH, what would work as a 'level shifter' to drive the LEDs? A straightforward way is to use an actual IC that accepts 1.8V input levels and outputs 3.3V levels. Example: SN74AVC2T244

I've also used bidirectional devices like the LSF0102 for the same purpose, as well as discrete n-FETS (possibly the cheapest / most available solution.)

Further reading: https://www.ti.com/lit/an/scea030b/scea030b.pdf


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.