08-07-2018 08:18 AM - edited 08-07-2018 08:19 AM
I am working with the Artix 7 (XC7A200T-1FBG484C) and am wondering how important de-sequencing is in terms of device reliability. I have the rails setup to de-sequence in the correct order but after observing the rails on an actual unit, the 1st rail that is supposed to come down (3.3V) does not have enough draw on it to fall below the other rails before they start to de-sequence. Is it imperative that this rail drop before the other rails? i.e. should I be adding extra load during de-sequencing or increasing the de-sequencing time?
The rails are 1.0V (VCCINT, VCCBRAM), 1.8V (VCCAUX, VCCADC), 3.3V (VCCO). Sequence up is 1.0V >> 1.8V >> 3.3V. De-sequence is 3.3V >> 1.8V >> 1.0V
The snapshot shows the 3.3V rail (purple), the 1.8V rail (yellow), the 1.0V rail (green) and a sequencer fault line (red).
08-08-2018 08:13 PM
From DS181, if the power sequence is not satisfied, "the minimum current draw and ensure that the I/Os are 3-stated at power-on" cannot be guaranteed.
The requirement is related to device reliability is "The voltage difference between VCCO and VCCAUX must not exceed 2.625V for longer than TVCCO2VCCAUX for each power-on/off cycle"
08-09-2018 06:08 AM
Thank you for the response. It makes sense that excess current could be drawn during shutdown. Fortunately, this shouldn't be an issue and may actually help bring a rail down faster.
As for device reliability, I do meet the allowed time per power cycle for VCCO(3.3V) - VCCAUX(1.8V) > 2.625V (300ms @ 125C). Are there any other concerns in powering down in this way?
08-09-2018 07:09 PM
I don't think there is any other concerns except for these two listed on the datasheet.
08-10-2018 01:36 AM
08-16-2018 07:54 AM
Thanks for the response. What do you mean by a glitch? Do you mean that the state of the outputs are not determined during the power-down, where they would be if the right de-sequence was followed?