I had a very peculiar situation on a PCB I am working on, and the only possibility I can think of is that my assembler installed the wrong chip on my board. I'd like to present it to the forum to see if anyone has another idea.

I was testing USB on my board, which includes a TVS chip at the input - pinout below: enter image description here

I had pin 1 tied to D-, pin 4 tied to D+, pin 2 & 3 to GND, pin 6 & 5 to Vbus. Pins 3 and 6 are NC, but I've used this same wiring previously and it worked no problem, and I'd prefer not to leave them floating.

When I went to test the USB communication I noticed a short from D- to VBUS and also from D+ to GND. After probing a couple copies of the board and noticing the same thing, I desoldered the chip from the board. The short on the board & the USB cable was removed, and USB worked flawlessly.

I probed the removed chip itself and saw shorts from pins 1 (IO) - 6 (NC) and 3 (NC) - 4 (IO). This explained the shorts on the board. But I can't understand why there are shorts here given the "not internally connected" statement in the function column of the table.

Is there any reasonable guess for what happened here, besides the manufacturer installing the wrong chip?

  • \$\begingroup\$ Next time it might be helpful to also provide a link to the full datasheet. Takes more of the guessing out of it as to which revision you're looking at and such. \$\endgroup\$ – Asmyldof Mar 31 '16 at 23:04
  • \$\begingroup\$ Can you not tell if it's the wrong or right chip from the markings? What about part orientation? Is pin 1 right? Also, if it's a NC then there's no such thing as floating. \$\endgroup\$ – DigitalNinja Mar 31 '16 at 23:04
  • \$\begingroup\$ Good point on the datasheet - here it is: bit.ly/1q2YjUh \$\endgroup\$ – Jim Apr 1 '16 at 3:49
  • \$\begingroup\$ Also, the chip seems to be oriented the correct way, although the markings are unclear. The NC's were also connected in a reference design so I took the word for it. \$\endgroup\$ – Jim Apr 1 '16 at 3:50

It could be the wrong chip mounted.

It could also be that Maxim as they sometimes do, decided to make an updated version with through-signal connection, with maybe another letter following it and this caused this behaviour.

Many manufacturers who make a TVS chip like this also make one with that connection you measured intentionally, so that you can have the signal go in one side and come out another and have extra freedom to wire up ground and pour it underneath for better shielding/characteristic-coupling.

This is half-similar to "mounting the wrong chip", but with the caveat that it's closer to "small mistakes can happen". It is the reason that generally "NC" pins are advised to be kept "NC" at all levels - You don't want to "inherit" ownership of this mistake if you can prevent it.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.