Shared publicly  - 
 
My fifth #USB #TypeC analysis is a great product I am very happy to try: the Google Universal 22.5W Dual Port USB Type-C Charger [P/N ADP-23AW BH]
tl;dr: Great charger, highly recommend! Super fancy. Comes with looooong cord!

https://store.google.com/product/usb_c_dual_port_charger

[Please see this post for why I am unlikely to do further Amazon-based product reviews:]
https://plus.google.com/102612254593917101378/posts/Ha5L7qAR8x8

I finally was convinced by an associate to order one of these after exemplifying the "Buy cheap, Buy twice" (or thrice) maxim. Thankfully I was rewarded with an amazing charger that does everything it states on the tin, and more!
(or, How I Learned To Stop Worrying And Trust The Benson.)

First thing that sticks out about this is it is absolutely tiny. It is a 22.5W adapter yet somehow is almost as small as the Huawei 6P 15W one. So if you are looking for a travel adapter (2 ports almost mandatory) this is the one for you. Do not pass go, do not collect $200, just go and buy this one right now. Seriously.

Now to the fun bit: the current autoassignment. I was initially skittish about this because of reports on Reddit that it only charges from 3A on one port, etc. However, now that I know more about the USB-C Type-C CC protocol, and have the specialized tools necessary to test, I hope to explain it to everyone else so they are more comfortable with the idea.

https://www.reddit.com/r/Nexus6P/comments/3qmdte/official_dual_port_225w_usb_type_c_charger/
https://plus.google.com/+BensonLeung/posts/4xq4EDjXMw8

The speed at which a devices "knows" it can charge is controlled by a resistor in the charger. (Google "voltage divider".) Normally 10kR (ohms)=1.69v which means 3A,  22kR=.94v => 1.5A, 56kR (REMEMBER ME?) =.417v => (500/900mA). It's up to the device to slow itself down.

Google didn't want to have a dedicated 1.5A port and dedicated 3.0A port, because then people would have to swap after 1 device was fully charged! Instead they added brains into the charger to see when a device -- despite being GIVEN 3A -- only takes <1.5A. Or, conversely -- when GIVEN 1.5A, if there was no pre-existing load -- upgrade it to 3A.

Neat, huh? :) Check out the video I took of it in action! This is super-cool.

I don't think any other charger maket has learned how to do this trick yet. That's why we're getting chargers with fixed ports that brown out under load, or don't fast-charge. But according to my USB-C 35W load meter this maintains a decent voltage all the way to 3A under load. Considering how poorly other compact chargers I've tested fared, this is really impressive.

As for cons: there is one. The way the algorithm works: the bottom port will "preempt" the top port.

Meaning if you have a Nexus 6P drawing 3A on the TOP port chugging along, and you plug in something like a Watch that draws 1.5A in the BOTTOM port, the bottom port will hijack 3A, even if it doesn't use all of it.

The Nexus 6P/top port will then be bumped down to 1.5A. Once the Watch draws less than ~1.2A, and stays there (i.e. finishes charging), the bottom port will go back to 1.5A and the top will return to 3A to finish charging the 6P.

This means if you use one of these adapters in the bottom port to plug in an old USB-A device that uses BC1.2 "1.5A" charging, you're going to have a bad time:
https://store.google.com/product/usb_type_c_to_usb_standard_a_adapter

It's an edge case, but one I know people like myself with A-devices and that cord may fall into.

That said this is one of the best chargers I've had the chance to review and plan to use it as my everyday charger. I strongly recommend looking into getting one if you want a quality, safe, and really frickin' neat charger

#USBC

Edit: Technical values such as ratings, vRp-USB trigger thresholds are listed in the photo album. Tables don't fit well in text! USB D+/D- lines are shorted internally and floating, BC 1.2 signaling. A small burst of 5v is noted on Vbus on connect, cause unknown. Capacitance?
22
4
Avuton Olrich's profile photoBenson Leung's profile photoNathan K.'s profile photo
6 comments
 
I bought the Google 60W Chromecast type-c charger for my Dell XPS and my Nexus 6P; I've been hearing some strange noises coming from it, I'd love to read a review of it, if you were so inclined.
 
+Nathan K., the 1.8M cable that comes with that charger is actually a very nice cable, by the way. It's a custom designed cable with some very interesting features that allow it to have about 1/2 of the limit for IR drop on Vbus and Gnd as required by the Type-C specification.

You won't find a review of that cable from me because I work too closely with that team at Google, but it's a nice one.
 
+Benson Leung Thanks again for the nuanced engineering details! The thing about good engineering is, if you've done your job right, noone will ever know! So thank you for giving that shoutout to your team.

(Again, this "$40" charger comes with *really nice* $20 cable! Proof:)
https://store.google.com/product/usb_c_usb_c

For reference what that means, here is post where Benson describes IR drop in detail:
https://plus.google.com/+BensonLeung/posts/gd1C2b4io7C

(I have some anecdotal photos collected from Reddit of the implications, but I'll leave that an exercise to the reader. Needless to say, *don't cheap out on a $2 cord for a $700 device!*)
 
+Nathan K.​​, also note that early on the 6P had an issue where it didn't detect the changes in CC voltage after initial connect, but that was fixed in recent 6P software releases. I pointed you to the kernel commit. 
 
+Benson Leung  Just following up with the reference link since search doesn't work. (Translation: if you have issues, *update your Nexus firmware!*)
https://plus.google.com/102612254593917101378/posts/QgAtds2qLjk

Ah, I thought that was a subtle reference to an "anomaly". If I may speak openly for the benefit of readers, correcting silicon-level bugs (like in the TIUSB320 controller) is notoriously difficult, and deserves public acknowledgement. (This is similar to the type of problem Intel had with their Pentium chips math processor.)

As an example, this is the exact thing I had to deal with when making a custom library for the MCP2515 CAN controller. See this for reference:
http://ww1.microchip.com/downloads/en/DeviceDoc/80179g.pdf

Translation: "[MCP:] Oh, by the way, even though you bought this chip because it has 3 mailboxes, if you use any two of them at one time all the mail inside them is shredded. KTHXBYE!"
</facepalm>
Add a comment...