AM3358: MDIO Voltage range

John Duquette

Part Number: AM3358

I'm bringing up the MDIO peripheral on a new AM3358 design. I'm seeing the wrong voltage levels on MDIO_CLK:

I'm expecting 3.3V_TTL levels (VOH < 0.4V, VOH > 2.8 per the datasheet to support VIL <0.8V, VIH >2.0V) but I'm seeing a swing from 0.9V to 2.2V. VOL is too high, and VOH is marginal.

This is with MDIO_CLK completely disconnected from the circuit the PHY. Is it possible the MDIO peripheral is getting mis-configured and putting out a different voltage level? The timing looks correct (1 MHz clock, 1Mbps MDIO) so it doesn't appear to be a wrong peripheral on this pin.

(edit of original post to test other thread)

over 2 years ago

0 John Duquette over 2 years ago

Intellectual 710 points

Another interesting tidbit. The MDIO_CLK signal starts properly at 3.3V, but then droops with this 1.3V signal on the output:

0 John Duquette over 2 years ago in reply to John Duquette

Intellectual 710 points

I'm guessing this is a feature compatible with 802.3 Clause 45 levels that is undocumented in the AM3358 datasheet. I've found that even with this odd clock level my PHY is responding to read requests.

0 -DK- over 2 years ago

TI__Mastermind 27720 points

Hi John,

Both MDC and MDIO are connected to dual-voltage I/O's (3.3V/1.8V). Is it possible that you have VDDSHV5 configured as 1.8V?

0 John Duquette over 2 years ago in reply to -DK-

Intellectual 710 points

Good question, but no, all VDDSHV* are tied to 3.3V. I had attached a scope plot showing the signal level drooping from 3.3V after a power up, but now I don't see any images in the attachments; that's odd.

0 peaves over 2 years ago in reply to John Duquette

TI__Guru 55265 points

The AM335x IO select their operating voltage based on the voltage applied to their respective power rail when reset is released. Therefore, you must make sure the IO voltage does not change after reset has been released.

Regards,
Paul

0 John Duquette over 2 years ago in reply to peaves

Intellectual 710 points

Interesting idea, Paul. Channel 1 (yellow) shows MDC starting at 3.3V and then cycling from 0.9 to 2.2V. Are you stating that if VDDSHV5 was at 1.8V when reset was released then later rose to 3.3V, it could explain this?

Would this apply to all IO on VDDSHV5? Checking the levels on another GPIO might confirm this, I think?

0 peaves over 2 years ago in reply to John Duquette

TI__Guru 55265 points

I'm not able to say what will happen. I can only tell you that the IO supplies (VDDSHVn) must reach one of the two valid voltage ranges defined in the Recommended Operating Conditions table before reset is released and must remain at this voltage while reset is not asserted.

Regards,
Paul

0 John Duquette over 2 years ago in reply to peaves

Intellectual 710 points

Thanks for the feedback. Once I stared checking the supply sequencing thisproblem resolved itself; this was coincidentally at the same time I switched oscilloscopes and probes, so either I had a bad setup that was fooling me, or I have an intermittent connection somewhere. I'll start a new case if this comes back up.

Processors

Processors forum

AM3358: MDIO Voltage range