New connector type/hardware annoyances

[Jan Axelson]

The USB 3.0 Promoter Group has announced that development has started for a type C connector that works in either orientation (no right-side-up):

http://www.usb.org/press/USB-IF_Press_Releases/Type-C_PR_20131203_Final.pdf

A reporter has asked me what other hardware annoyances remain. Any thoughts?

[Barry Twycross]

The orientation is definitely an annoyance, its been immortalised in cartoons. (I wish I could find the original of this, its been copied a lot).



I'm not convinced changing the connect is worth the win for orientation. Interesting it was Apple which popularised the flipable plug.

Other annoyances are commonly power. The spec has been abused by devices for so long, and the proliferation of brick standards.

[Jan Axelson]

What irks me about the current USB plugs is that there are no obvious visual or tactile hints to tell you which way is "up" (assuming vendors are consistent with socket orientation). For vertically oriented sockets, I don't know if there is a convention.

I'm not surprised Apple had it first.

[Barry Twycross]

I look at the face of the plug, to see where the plastic bit is in the plug. By spec, that should be on the bottom. (The spec is the trident icon on the plug should be up, but the trident icon doesn't show up well.)

I found the spec section, Sec 6.5.1 USB Icon Location, that says "Receptacles should be oriented to allow the Icon on the plug to be visible during the mating process." For horizontal ports, that means up. For vertical ports, all best are off.

That's also not stopped ID departments from designing enclosures where the motherboard went in upsidedown, so the motherboard connectors were upside down once it was in the enclosure. We didn't notice that when we only had boards to work with, once it was in an enclosure it was too late to fix it.

[Jan Axelson]

Yes, thanks! the plastic bit is much easier to see than the trident.

I've thought of another annoyance that's hardware-related: PCs and other components that don't recover properly from energy-saving states so you have to leave them powered up.

[Barry Twycross]

I'm not surprised that low power states don't work, I rather don't expect them to work and I'm pleased when they do. I've spent sufficient time debugging power transitions to know how difficult a problem it can be. Particularly amusing are issues which only happen within a certain bit time window relative to suspend or resume. The last one I debugged happened about 1 in 60,000 so much have had an 8 bit time window (for high speed).

[Jan Axelson]

Yikes.

I need to adopt your attitude on it!