TXS0102: Level shift for MDC/MDIO

Hyoungrae Ryu

Prodigy 30 points

Part Number: TXS0102
Other Parts Discussed in Thread: TXS010, TXS0104E

Dear TI

I want to control MDC/MDIO.

I want to design a circuit with TXS0102 or TXS0104.

I have two questions.

Q1] Please check the attached spec below and answer if it is possible to circuit design.

Ref. guide MDIO=1.5k p/u

Q2] Where does the MDC signal belong to the data-rate of TXS010* spec?

In my personal opinion, MDC is 125ns, so does it belong to Push-pull?

[Attach#1. Ethernet switch MDC/MDIO datasheet]

[Attach#2. TXS0102 data rates]

please check it

5 months ago

0 Michael Ikwuyum 5 months ago

TI__Mastermind 27940 points

Hi Hyoungrae,

1. The TXS device includes internal pull-ups. Hence, external pull-ups are not necessary.

2. Push-Pull, thanks.

Best Regards,

Michael.

0 Hyoungrae Ryu 5 months ago in reply to Michael Ikwuyum

Prodigy 30 points

Dear Michael

Thank you for your response.

I'd like to know if the MDC(125ns,8MHz) clock works as well.

For example, If the MDC corresponds to an open train, it will not work.....,

And, The MDIO of the data sheet is guided with 1.5k resistance,

A is an internal 10k resistance, so it seems to be very different.

Shouldn't I add 1.8k resistance to the outside?

please check it

0 Michael Ikwuyum 5 months ago in reply to Hyoungrae Ryu

TI__Mastermind 27940 points

Hi Hyoungrae,

Yes, supporting up to 24Mbps push-pull is of no concern as 8MHz = 16 Mbps. However, not recommended for open-drain i.e up to 2 Mbps.

Also, external pull-up < 10K is not recommended and the device will work okay, please help see section 8.3.5 of the datasheet for more information, thanks.

Best Regards,

Michael.

0 Hyoungrae Ryu 5 months ago in reply to Michael Ikwuyum

Prodigy 30 points

Dear Michael

The MDIO appears to be an Open-drain.(because of P/U resistor)

If then, Is the MDC signal Push-pull or Open-drain?

Am I getting it wrong?

please check it.

0 Jack Guan 5 months ago in reply to Hyoungrae Ryu

TI__Mastermind 22181 points

Hi Hyoungrae,

Yes you are correct, MDC is typically driven with push pull driver since it is a fixed direction signal and can support a higher data rate then MDIO. It is briefly discussed in this one pager: Translate Voltages for MDIO

Regards,

Jack

0 Hyoungrae Ryu 5 months ago in reply to Jack Guan

Prodigy 30 points

Dear Jack,

The TI's MDC/MDIO guide is TXS0102 IC.

However, I have to design with 4-ch Level shift IC, going to design it as TXS0104 IC.

Is it okay to design it?

0 Michael Ikwuyum 5 months ago in reply to Hyoungrae Ryu

TI__Mastermind 27940 points

Hi Hyoungrae,

Yes, this is okay, thanks.

Best Regards,

Michael.

0 Hyoungrae Ryu 5 months ago in reply to Michael Ikwuyum

Prodigy 30 points

Dear Michael

When testing with TXS0104ERUTR, I checked the operation as follows.
The MDC signal is operating normally(6.25MHz), but the VCCB's MDC I/O level is insufficient(about 2.5V).
What could be the cause? and I also see Over-shoot problem.

please check it

[Circuit]

NLSX5014 --> TXS0104E change(pin to pin)

[Waveform_TXS0104E's VCCA MDC/MDIO]

[Waveform_TXS0104E's VCCB MDC/MDIO]

0 Michael Ikwuyum 4 months ago in reply to Hyoungrae Ryu

TI__Mastermind 27940 points

Hi Hyoungrae,

The overshoot is most likely due to the device's interaction with the impedance line. You may use series resistors on the I/O to help reduce if a concern.

VOH not reaching VCCB is typically due to a voltage divider network on the line i.e. from pull-ups and pull-downs connected or from excessive capacitive loading / long traces. You may confirm by disconnecting peripherals connected to observe how VOH increases.

Best Regards,

Michael.

+1 Hyoungrae Ryu 4 months ago in reply to Michael Ikwuyum

Prodigy 30 points

Thank you.

I'll check for additional problems.(Based on the your answers)

Logic

Logic forum

TXS0102: Level shift for MDC/MDIO