LPC43xx USB V_BUS 5V tolerance errata, removed on latest errata sheet

[lpcware]

Content originally posted in LPCWare by Zuofu on Sun Oct 27 15:26:32 MST 2013
On the LPC43xx series, USB_VBUS is only 5V tolerant if the 3.3V supply is connected to VDDIO. Therefore, connecting USB_VBUS directly to the 5V line on the USB bus is a problem if the 3.3V supply takes some time to power up (for example, if it is powered by a switch-mode power supply). This issue is mentioned in document ES_LPC43x0 revision Rev. 1.2 — 1 February 2012 under section 3.7 (page 8) as USB 0.1, which recommends a workaround using a voltage divider or load switch. However, the latest version of the same document ES_LPC4370/50/30/20/10 Rev. 6 — 21 October 2013 no longer mentions this issue, nor does this issue seem to be mentioned anywhere in the latest datasheet. I can confirm on a LPC4337-FT100 chip that connecting the 5V USB power directly to USB_VBUS when the 3.3 V line takes time to start up is a bad idea, it will cause a latch-up issue where the chip gets very hot.

Is there a reason why this issue was removed from the errata sheet? Is there a more in depth discussion of this in the proper documentation somewhere? Attached are 2 revisions of the errata sheet for the LPC43xx series. I think not mentioning this issue anywhere in the current official documentation will cause a lot of problems for people (it has for me). This is especially the case as the KEIL MCB4300 only has a protection diode on USB_VBUS (which works because when using USB power, the 3.3V rail is provided by a linear power supply with low start up delays).

[lpcware]

Content originally posted in LPCWare by Hydron on Thu Oct 31 14:06:54 MST 2013
I don't need the on-the-go function, so a simple voltage divider works fine for me.
I'm just quite annoyed at NXP for yet another gotcha with this chip, especially given that if it had been kept in the errata sheet I'd have known about it much earlier, rather than finding it by luck in a tiny footnote. Using this chip in a design has been frustrating to say the least.

[lpcware]

Content originally posted in LPCWare by Martin84 on Thu Oct 31 04:25:46 MST 2013
Hi Hydron,

the more complex circuit is used for on-the-go, because the VBUS Pin must be able to detect the right vbus-level in device mode. With the voltage divider the vbus level will be distorted.

It would be interesting to have a special category like "Hardware issues" in this forum where such small issues are mentioned and discussed for all the LPCs together. So a developer can search for this hardware issues without reading the software issues.

Regards

Martin 

[lpcware]

Content originally posted in LPCWare by Hydron on Tue Oct 29 21:39:50 MST 2013
I had exactly the same issue, I had almost finalized a PCB when I saw the small footnote about this pin. Absolutely stupid design given the intended use for this pin, and very easy to miss the tiny footnote.
The footnote also doesn't mention that a voltage up to ~3.6V is safe, so I used a circuit more complex than the voltage divider suggested in the errata sheet. Whoever removed it from the errata sheet deserves a kick up the ass.
/rant

[lpcware]

Content originally posted in LPCWare by Martin84 on Tue Oct 29 08:29:07 MST 2013
Hi Zuofu,

I know what you mean. For the 4357 there's also nothing mentioned in the errata sheet. The only hint on the USB0_VBUS Issue is in UM10503 Page 626 (Remark:).

In this post "USB 0 circuit on LPC4330/LPC4337" there's a solution given by NXP Support.