Ultra96-V2 Errata

The first 3000 kits include a Quick-Start Card that states no LEDs are lit when powered off. This is not correct. For Revision 1, we added LED D17 which is lit whenever 12V is alive. This was added to remind users that they should not add pods or mezzanines to the board when plugged into 12V, even if the power is off (because it's not really off until you unplug the barrel jack!).

Switcher A of Pmic 0x14 may issue an over-current warning, especially at hot temperatures. This warning can be viewed in the Infineon PowIRCenter tool. We believe this is a false warning. Please ignore.

Some of the first 1000 kits may have a microSD card cage where the card insert and eject isn't smooth with the included Delkin card. This is NOT due to the Delkin card, which is entirely within spec. This is due to a lack lubricant inside the sliding mechanism of the cage. All boards are functional and passed our factory functional test. However, if you find your cage starts to malfunction, you may request a free SanDisk microSD card from your Avnet FAE. The SanDisk card is very marginally thinner, but it is enough to make a difference. You may also request an RMA through your Avnet sales organization and we will replace your board.

Hi Bryan:

The Quick Start card also references a Getting Started Guide that doesn't yet seem to exist on the V2 Documentation Guild site yet.  Probably just not posted yet?

Yes, working on it now. Hope to have it posted tonight. We are also working on a new design with the access point working, so that will then require a new download and updated document.

Bryan

A first version of the Getting Started Guide is now posted. We are working on updating the image to fix many of the bugs in the Webserver. Thank you for your patience. The document is posted here:

Ultra96-V2 Development Board | Zedboard

Bryan

Ultra96-V2 -- A power-down/reboot issue related to heavily loading the PL is being investigated. This was first noticed when executing the DPU at 666 MHz. Operating it at 500 MHz makes the problem go away. During operation when the failure happens, the board will spontaneously reboot.

This appears to be caused by a threshold being set to low on Pmic 0x13 CHD, which is the 0.85V, 3.0A, +VCCINT rail. When the current exceeds the Fault threshold, the Pmic disables that rail, causing Power Good to be deasserted. Since Power Good is tied to the MPSoC POR, the board resets. When the Pmic rail stabilizes, then Power Good is asserted, which deasserts POR, and then the board boots again. This is the intention of this design to prevent damage to a board if excessive current were drawn. However, our initial estimates using XPE and a heavily loaded PL test image appear to have been too pessimistic.

+VCCINT powers all of the following:

More information will be provided as we finish the verification of this issue.

Bryan

Hi Bryan:

DPU in the ultra96v1 operates at 650MHz without reboot.

But ultra96v2 does not.

Could you please fix this issue and try to make the power rail spec of v2 board as same as that of v1 ?

Yes, we are aware of that. The V2 board's power rail is specified exactly the same as V1. The difference is that the V1 power supply had nothing in place to signal when it exceeds the spec. The V2's enhanced power supply does. So, the new power supply comes with the additional benefit that your V2 board shuts down when exceeding a certain threshold. Since the V2 power circuit was over-designed to support more current, the fix is simply a matter of adjusting those thresholds in the Pmic. We have a proposed solution now, which is being tested. Thanks for your patience.

Bryan

Hi Bryan:

Any update for the pmic issue?

Is there any solution for it?