SN65HVD72: About Recommended Operating Conditions

Tachibana-san,

It can be difficult to quantify this due to the number of variables involved.  To give some reference, though - some of the lower-voltage transistors used for internal logic circuit would be the most susceptible to effects of elevated VCC, and these transistors' component-level FIT (number of failures per billion operating hours) could be expected to increase from 0.000467 (at VCC = 3.6 V) to 0.0394 if VCC were operated at 4 V for 100% of the device's lifetime.  (The FIT would continue to degrade exponentially at this rate with further increases in VCC.)  As the percentage of time exposed to the high voltage reduces, the FIT improves: for example, with 10% exposure the FIT reduces by about 1000x.  With 1% exposure it would reduce by about 1000000x.

The overall device-level FIT (per ti.com) is 0.1, meaning 1 failure per 10 billion operating hours.  This is relatively low for most applications, and so in most cases even moderate increases may still result in a FIT that is within a given application's targets.

If the customer is still concerned about tolerance to higher VCC levels, I would recommend evaluating THVD1410.

In case you are not familiar with FIT terminology, please reference:

https://www.ti.com/support-quality/reliability/reliability-terminology.html

Regards,
Max

Tachibana-san,

I understand your confusion, since you are correct that the same concerns would apply to any IC.

There were a couple of reasons for this note.  One is that while this device was designed for 3.3-V operation, many other RS-485 transceivers are designed to operate at VCC = 5 V.  We wanted to make sure it was clear that the device was not designed for long-term operation at 5 V, since we have seen some confusion around this topic.

Another reason relates more to the change history.  Before this version of the datasheet, a previous version was published which lowered the absolute maximum VCC to 3.6 V.  This was because a review of the design identified some transistors in the IC which (per our internal component specifications) were not rated for 5-V operation.  This change was difficult for customers, though, since some applications may not always control VCC well to keep it below 3.6 V.  Based on this feedback, we performed a more in-depth analysis and found that the degradation at higher supply voltages was actually not so severe.  Based on that, we were able to re-revise the datasheet to increase the limit to 5 V.  However, as not to give the impression that there was zero reliability impact at all to higher voltages we added the footnote (which doesn't modify anything beyond the standard definitions of the recommended/absolute maximum limits but perhaps calls better attention to the concern).

Regards,
Max