Post is pinned.Post has attachment
This is a formal notice and statement of clarification:

I am an end customer doing independent verification and testing of devices and products. All statements I have made are factual and correct to the best of my knowledge and skill.

I am a volunteer with the Google Top Contributor Program, Nexus and Pixel Devices. The work I do here is independent of that. I also currently have no business relationship or interests with the USB-IF. I merely ask them questions as any diligent customer would.

Disclosures: I have no conflicts of interest to disclose.

Why am I posting this?

Some people at major organizations are getting up in arms over the analysis I have made revealing unsafe or poorly designed products. They have taken offense at the work Benson and I do, despite being protected speech and of direct service to the community.

As an engineer, I feel this work squarely falls under the Code of Ethics I agreed to when I passed my FE exam with the National Council of Examiners for Engineering and Surveying, and was certified by the National Society of Professional Engineers.

Furthermore, all my statements are backed up with hard data (freely available). If you have a disagreement with one of my conclusions, please responsibly bring it to attention through proper channels and I will endeavor to improve and correct that data.

I will pin this post at the top of this collection until further notice. I sincerely hope others will attempt to be responsible in their actions as well. Comments will be left open only for topics relating to the above. Others will be deleted.

I. Fundamental Canons
Engineers, in the fulfillment of their professional duties, shall:

1. Hold paramount the safety, health, and welfare of the public.
2. Perform services only in areas of their competence.
3. Issue public statements only in an objective and truthful manner.
4. Act for each employer or client as faithful agents or trustees.
5. Avoid deceptive acts.
6. Conduct themselves honorably, responsibly, ethically, and lawfully so as to enhance the honor, reputation, and usefulness of the profession.
Add a comment...

Post has attachment
I've moved #USB #TypeC writeups and studies over to a new collection. Please subscribe to "Writeups" if you'd like further #USBC deep-dives.

This is also to separate "product evals" from "testing methods" and "breaking news". The line is somewhat vague and arbitrary.
Add a comment...

Post has attachment
I've moved #USB #TypeC testing over to a new collection. Please subscribe to "Season 2" if you'd like further #USBC updates.

This is also due to the forced transition to the new Google Plus interface, which has much inferior Photo Album management tools. :(
Add a comment...

Post has attachment
This is a status update. After the loss of my primary computer due to a bad USB-PD power supply, and significant financial pressures, I think it is necessary to move to a "Season 2" of testing with different format and goals.

It is clear there are still very many bad products out there. People such as +Benson Leung and myself must continue to press the matter in the interests of public/manufacturer safety and awareness. However, as the recent article by +Glenn Fleishman I was cited in +Fast Company points out, it should not fall to volunteers to verify this sort of thing.

How Volunteer Reviewers Are Saving The World From Crummy—Even Dangerous—USB-C Cables

I plan to resume #USB #TypeC testing at the behest of colleagues like +Dan Ziegler, a fellow +Google Top Contributor Program volunteer in #ProjectFi, who sent me yet another bad cable he was using. When even fellow experts are having junk land in their lap, it is a signal much work is needed.

The lack of funding also means any impact I have is extremely limited. For example, Amazon filters reviews that are not purchased outright. Benson is able to have a greater impact since he is able to fund his reviews through employment with Google, whereas I cannot. Everything I've done has been personally funded, or donations and goodwill from the public such as yourselves. (+Plugable, +Total Phase, and +You.)

I've also had criticism from potential employers who look negatively upon my work here. They are concerned about statements that may conflate personal opinion with professional employment capacity (which is appropriate and correct), but also about an employee who follows proper procedures per the requirements of his Engineering license (which is a dubious grey area).

In defense of my actions, I point out unlike most every other profession associated with tech -- such as Computer Science/Engineering -- Mechanical Engineers possess a license that can be revoked for malpractice or ethical violations. Case studies are below. Please note the major choice of the individual at Stanford, and his initial disclosure steps; companies cannot claim to be ethical if they discriminate against hiring ethical employees.

In order to "Do The Right Thing" (publish analysis data and call out bad products) while facing the realities of business (not ruining my chances of employment), I will merely be publishing abbreviated data from now on. I will still be doing deep-depth tests, but they will merely be a +/- number rating per my "Suggested Peripherals for the Google Pixel" spreadsheet, which has proved to be well-received.

As a major change, I will set up a Patreon or GoFundMe account to fund random-sample testing of products at "point-of-sale" to the public, ala +Snell Memorial Foundation testing.

Manufacturers have approached me privately for #USBC pre-release testing, and I have obliged out of "noblesse nerdige". I don't mind giving my opinion, that is only responsible -- especially for "revised" products. But I am starting to have products arrive at my door unannounced, and that type of work clearly falls in the realm of paid employment, or professional consulting. Benson is a far better choice for that, since it is within the scope of his formal employment at Google. (Chromebook development.)

This ensures if I am ever hired at such a company, there will be no conflict of interest in continued testing. If anything, it would allow me to pivot it under their auspices and goodwill.

I am open to feedback or opinions on this change in direction and would welcome input. If you would like to share comments privately instead, please +mention me in a privately shared post. I will try to respond as best as the "Jingles" notification mascot will allow.

Nathan K.
Add a comment...

Post has attachment
You know that bad #USB #TypeC power supply that took out my desktop computer? It just took out my oscilloscope while measuring leakage current. This just keeps getting better and better. 😤 #USBC (Pinging +Benson Leung​​) This has all the hallmarks of a Class-Y capacitor failure.
Animated Photo
Add a comment...

Post has attachment
It was bound to happen eventually. A bad #USB #TypeC product finally took out my computer. My DESKTOP computer no less. Was logging and it got fried through the datalogging connection. Big shower of sparks out the back. PSU smells like victory. Autopsy pending. #USBC
Add a comment...

Post has attachment
70th (and capstone) #USB #TypeC analysis: Black Squid "USB3.1Gen 2 3a/60w" 1m cable [Model BSQ_USB_CC_1-1: v1, v2]. (The v1 is what +Benson Leung tested.)
tl;dr: "Very bad" cable (v2) and "Bad" cable (v1). Cease using both "when reasonable". Velesco, under their lifetime warranty, needs to compensate purchasers. These cables are not fit for their eMarked/advertised ratings.

As this may be the last public critique permitted in this series for a while, I want to stress the importance of the some details. Please consider reading this foreword:

This class of failure was only possible to discover with a $7500 piece of +Total Phase test equipment. You should not hold this data against anyone. If you should take any moral away, it should be why independent peer-review is absolutely critical and a well established part of research.

We should respect and appreciate the contributions of all researchers and engineers -- whether they work out of a garage, or a multi-million dollar lab. Manufacturers and resellers alike need to exercise diligence and act responsibility regarding validating their products prior to sale to the public, and followup if there are flaws.

We shouldn't be afraid of "doing the right thing" for fear of making mistakes -- or worse still, having those actions held against us. Do good, do evil, or do nothing. Only two are acceptable -- it's up to each individual to determine which.

(Edited 1/14: at request of Velesco) +Velesco is offering either refunds, or exchanges for the corrected model when they complete it. I want to highlight they are "trying to do the right thing". I still fault the fact product was not verified before sale to the public, but given the circumstances their response has been adequate. I reserve judgement until I see how they actually implement their fix.

(Edit 1/15: correction notice) As clarification of the timeline:
   January - > Benson's Amazon review
   April - > USB-IF logo issue addressed [V1 - > V2]
   December - > SI issues discovered
   January - > current safety issues discovered
Benson's initial feedback was unrelated to the issue, and Velesco did everything they could by notifying their factory to correct the USB-IF logo. The additional flaws were unauthorized changes made by their factory, per their comments below.

Please take this as a cautionary tale of keeping strict QC metrics in place to prevent "quality fade".

The "v1" cable is on loan from +Benson Leung. I plan to return it now that this evaluation is complete.

The "v1" cable was the very first eMarked cable Benson evaluated -- and is my very last. You can identify it by the presence of a incorrect "SuperSpeed Trident" logo on one plug, and "BSQ" on the other. It is advertised and sold as "Meets USB 3.1 specification Rev 1.1. Super speed (up to 10 gbps) and fast charge up to 3A (supports USB 3.1 Power Delivery PD)." [See screenshots.]

[TotalPhase report for for BSQ_USB_CC_1-1-3] (v1)
[TotalPhase report (50% SI) for BSQ_USB_CC_1-1-3] (v1) <- Gen1 speed check

[TotalPhase summary sheet for 6 separate tests]

From observation and testing of the unit, my findings differ from Benson's. He re-visited this cable prior to myself, and we now share an opinion regarding the data. It has a number of safety and compliance failures:

(1) It fails DC Resistance (DCR) testing at 3A. The per-pin resistances are in excess of USB-C mandated minimum thresholds. This makes it questionable for data and charging purposes (since differential data and power share ground).
(2) It fails Signal Integrity (SI) testing at USB3.1Gen2 10gbps speeds. (It marginally passes at 5gbps). This means data will suffer from an excessive amount of corruption. It also suggests Black Squid's manufacturer used inferior "Gen1" cable stock.
(3) It has eMarker errors that render it noncompliant per the spec. It is labeled "Vconn required" as opposed to "Vconn for identify only". This will result in the cable wasting 25mW of energy when connected to "smart" devices. (Potentially including battery-powered Apple iPhone 8 or iWatch 3.) Product ID is blank. Hardware field is blank. Firmware field is blank. [Vendor ID is filled in 0x2109 = 8457|VIA Labs.]
(4) This is in legal violation of USB-IF licensing agreements. It uses the Trademarked USB-IF "SuperSpeed Trident" Gen1 logo, without having permission and violating design guidelines. It should be a "SuperSpeed 10 Trident" Gen2 logo, only with permission of the USB-IF. They could be sued for this and have their products embargoed at customs (were the cable still in production).

This "v1" cable was supposedly pulled  from sale some time after Benson's initial evaluation. I am not at liberty to speculate how or why. However, as recently as January 3rd, 2017 I was STILL able to place an order for it. Please see the attached receipt, screenshots, and packaging materials.


This new "v2" model I alone have tested. You can identify it by the presence of a "Squid" logo on one plug, and "Black Squid" text on the other. It is advertised and sold as "Meets USB 3.1 specification Rev 1.1. Super speed (up to 10 gbps) and fast charge up to 3A (supports USB 3.1 Power Delivery PD)"

The "v2" has even more significant safety and compliance errors that make it blatantly unsafe. On a process level, Black Squid apparently did not use this OPPORTUNITY to re-evaluate their production line from scratch when they removed the "USB-IF Logo" and resumed sales. In the process, their factory seem to have unilaterally actually made the cable WORSE.

[TotalPhase report for for Blacksquid BSQ_USB_CC_1-1-3] (v2)
[TotalPhase report (5A DCR) for BSQ_USB_CC_1-1-3] (v2) <- 5A current check

[TotalPhase summary sheet for 6 separate tests]

Per my observation and testing with a +Total Phase Advanced Cable Tester Level 2, manual examination with a Google/+Plugable "Twinkie" sniffer, TotalPhase USB-PD Analyzer, and custom microcontrollers and Python scripts of my own development:

(1) It fails DC Resistance (DCR) testing at 3A. The cable total resistance -- on a brand-new cable -- is in excess of USB-C mandated minimum thresholds. This makes it questionable for data and charging purposes (since differential data and power share ground).
(2) It fails Signal Integrity (SI) testing at USB3.1Gen2 10gbps speeds. This means data will suffer from an excessive amount of corruption. It also suggests Black Squid's manufacturer used inferior "Gen1" cable stock. In fact, these are some of the worst SI data eye diagrams I have ever seen.
(3) It has eMarker errors that render it noncompliant per the spec. It is labeled "Vconn required" as opposed to "Vconn for identify only". This will result in the cable wasting 25mW of energy when connected to "smart" devices. (Potentially including battery-powered Apple iPhone 8 or iWatch 3.) Product ID is blank. Hardware field is blank. Firmware field is blank. Vendor ID is blank. The "latency" field is an illegal/reserved value of "0000" (no length).

The most serious error:
(4) This cable is electronically marked as being engineered and capable of supporting 5A/100W capacity. Remember, this cable doesn't even meet 3A thresholds, let alone 5A. This is an absolute bust.


This means if you use this cable to connect your Apple 87w (20v/4.3a) power supply to your Apple Macbook Pro 2016 15" with TouchBar -- the cable will lie to the power supply about its capabilities . It will result in a substandard (if not electrically dangerous) situation where the charger will try putting out 5a through a cable that is not capable of it.

In fact, I replicated this situation using my Google Pixel. Please see the attached traffic dump taken with a TotalPhase USB-PD Analyzer in just such a setup. Note how the Apple charger asks the cable for its capabilities, and it responds "5a". So the Apple 87W charger is then tricked into advertising its full 20v/4.3a PDO, as opposed to limiting itself to 20v/3a.

Please, NEVER use this cable with your Macbook Pro. At best it will perform poorly, at worst it may cause hardware damage.

All in all, I am extremely disappointed in the performance of this cable -- especially the second iteration. I must advise users cease using this cable "when reasonable" and file for a refund/exchange. Screenshots of said guarantee have been included from cached webpages.

I greatly appreciate the peer-review and mentorship Benson has provided for these studies. As with this and the Anker cable previously, independent verification and replication of results is one of the few avenues to truly ensure good data. It is only though open discourse, sharing of material and methods, and good faith can issues be resolved.

[Plus] Analyses BlackSquid BSQ_USB_CC_1-1-3
12 Photos - View album
Add a comment...

Post has attachment
69th #USB #TypeC analysis: +CHOETECH "USB-IF Certified" USB 3.1 Gen 2 Type-C Cable [Model B01N8U9075 / A3002].
tl;dr: BAD cable. Another case of USB-IF doing a poor job of enforcing manufacturer-level quality control post-Certification.

This sample was provided by​ +Francisco Vigil from my queue of "USB-C Stuff to Analyze". So thank him for the test!

What I detail is based solely on findings from the "Fulfilled By Amazon" sample purchased by Francisco Vigil. The data is right here if you wish to verify. I was prompted to share this data (after attempting to work with Choetech) as a result of findings from +Brian Lowry regarding hardware damage to his #madebygoogle Pixel phone.

Also see +Benson Leung's comments dated December 3rd here:

[TotalPhase report for for Choetech B01N8U9075]

[TotalPhase summary sheet for 6 separate tests]

This cable failed signal integrity on EVERY run on the +Total Phase Advanced Cable Tester 2 at 10Ghz at default settings. This does not mean the cable doesn't meet compliance criteria -- it just means the cable failed a test every single other good cable in my section has passed.

This 1 meter "Gen2 10gbps cable" performs qualitatively WORSE than a 2m "Gen1 5gbps" Plugable Thunderbolt3 cable. Please see the attached image comparing two eye diagrams. This either means the Plugable is capable of 10gbps at 2m (extremely unlikely) or the Choetech cable has very poor cable stock (very likely) and operates at thresholds that +Intel would refuse to permit in their own cables.

There are major eMarker errors that by definition make this noncompliant. The CertStat field is blank. Vconn is marked as being required. Vendor ID and Product ID are blank. Therefore this is another demonstrable "BAD" cable certified by the USB-IF for sale to the public.

See my comments on the Targus "USB-IF Certified" cable here for the same criticisms:

Cable Termination Type
00b = VCONN not required.
01b = VCONN required
Cable Plugs that only support Discover Identity Commands SHALL set these bits to 00b Vendor ID
Manufacturers SHALL set the Vendor ID field to the value of the Vendor ID assigned to them by USB-IF. Product VDO
Manufacturers SHOULD set the USB Product ID field to a unique value identifying the product and should set the bcdDevice field to a version number relevant to the release version of the product. Cert Stat VDO
The Cert Stat VDO SHALL contain the XID assigned by USB-IF to the product before certification in binary format.
"Test lab engineers are required to confirm that the XID used in the cable under tests’ E-marker matches that which the USB-IF assigns. (If it doesn't match...) the E-marker tests are considered a Fail." Shall/Normative
Shall and Normative are equivalent keywords indicating a mandatory requirement. Designers are mandated to implement all such requirements to ensure interoperability with other compliant Devices.

Please also note +Brian Lowry 's report here regarding the cable failing immediately after purchase, sparking sounds, and causing errors to appear on his computer charge indicator:

"I purchased this cable and it in fact seems to have failed right out of the box. Furthermore, every time I try it on my Pixel, I cannot then seem to charge the phone with ANY other C-C cable until I reboot the phone. It also produces a discomforting crackling noise when I unplug the cable."

"Choetech is sending me a replacement cable as they are sure this must just be an isolated defect and seem adamant that the USB-IF certification is itself the pudding that proves. It does seem to work fine for others, including +German M. on his Pixel XL. But my laptop indicated a problem with it beginning with the very first use ('X' next to the battery indicator and no charge) and others have reported possible electrical crackling on disconnecting from a live power source. I'm also now consistently replicating the issue with my Pixel where I cannot use ANY known working C-C cables after I try the Choetech 10Gbps (with no success), until I finally reboot the phone. Only then will the phone again accept power from any of those cables."

End ramifications?

This cable will be less energy efficient and burn (5 volts)^2/(1k ohm) = 25mW of power it otherwise shouldn't on some "smart"/PD devices. Some devices may refuse to work with it if they demand a Certified cable. Some devices may get confused because of the 0000 VID and 0000 PID. (This is a non-valid entry.)

**You** be the one to tell Apple that you're draining their iPhone 8/iWatch 3 battery budget by 25mW due to a firmware bug with your cables, and see what happens!

I do not recommend this cable. It raises serious doubts about the USB-IF's commitment to their own stated requirements.

[Plus] Analyses Choetech B01N8U9075
7 Photos - View album
Add a comment...

Post has attachment
34.1st #USB #TypeC analysis: AUKEY "24W" USB-PD Dual Port 12v car charger [Model CC-Y7 "v2"].
tl;dr: PROVABLY BAD. Also the previous "v1". Both do not honor the emergency-stop "Hard Reset" command. And send VDMs inappropriately. This USB-PD controller chip is no good. AUKEY rolled out a second item number, I thought they fixed it, they didn't.

[Original post on AUKEY CC-Y7 "v1"]

This one I bought myself. I noticed AUKEY rolled out a new product number. (new v2) B01NAICO3G (old v1) B01LGO9LE4 and I had hoped +AUKEY Official / +Aukey Aukey  would listen my and +Benson Leung's warnings and fix their charger. Or at least the safety text printed on the side/in the PDO's so they line up.

Instead, I find they made NO changes. And in fact, this v2 -- now that I have the +Total Phase PD-Analyzer -- I can PROVE has USB-PD chipset failures that render it operably dangerous with expensive "smart" equipment. (Which is exactly what you'd use a USB-PD charger for.)

First, it issues "HARD RESET" commands as a hack to reset its port partner during the negotiation process. However, AUKEY completely fails to implement the required safeties.

"Hard reset" is the electronic equivalent of an "E-Stop" or "Emergency Stop" for mechanical equipment. It's basically a giant red button that if SHTF someone presumably presses to abort. In motorsports it's the "lollipop" that cuts electricity and gas to the engine located near the bumper. (No engineer worth his salt won't ask "where's the E-Stop" as the first question with a piece of machinery he is unfamiliar with.)

During a Hard Reset, a source is MANDATED to reduce voltage/current to Zero. Please see the attached highlighted pages from the spec. Also see the experiment I performed where I loaded the charger through this phase -- and how the AUKEY does not honor the electronically transmitted emergency abort command. (In fact, it is the one initiating it -- so it is worse.) I maintain a 3A draw throughout the process.

Would you want to charge you $2,500 Macbook Pro 2016 on a charger that ignores this type of command?

Second, if it fails to find a USB-PD partner, or the partner doesn't respond for whatever reason, the charger sends a single Structured Unstructured VDM at the end. Presumably this is buggy code or a backdoor firmware update command for the unit. However, this is again a straight up violation of the spec. In this case, "if a USB-PD packet is sent in the forest, and nobody is around, yes it still matters". Again, it signifies deeper USB-PD errors with this chipset.

Before I was marginally tolerant of this charger since there were few better options. That is actually why I ordered this second unit, because I was LOOKING for a good one. However this type of error is unjustifiable. If this fries your Mac, you've been warned.

[Plus] Analyses AUKEY CC-Y7 (v2)
12 Photos - View album
Add a comment...

Post has attachment
68th #USB #TypeC analysis: iMuto 16750mAh Taurus Smart Power 3-port battery + A-to-C cable [Model X5TC].
tl;dr: GOOD battery pack, BAD cable. Pack is great: has decent overcurrent protection, minor charging bugs. Cable noncompliant:  has right Rp, but forgot to connect the shield wires!

This sample was provided by​ Adam Baxter​ from my queue of "USB-C Stuff to Analyze". So thank him for the test!

This charger uses a very similar chip to the RavPower RP-PB043 "dumb" Type-C DRP charger. It even is "always on" like that one. This also has a LCD display and GREAT overcurrent protection. So I'm assigning them the same rating. They're "equivocal" aside from capacity.

This charger also has "sawtooth" IR drop compensation. Unfortunately it is insufficient to compensate for the cable, so causes some weird behavior with the Pixel (see logs), but at least it is there. Better than nothing.

The Type-C port has BC1.2 signaling. Oddly, when I probed the charging behavior, it cannot RECOGNIZE D+/D- signaling. Also, it tries drawing current at a 900mA level without enumerating, as opposed to 100mA - > enumerate - > 500mA like it should. Also, the D+/D- output signaling is "2.5v".  This isn't quite 2.7v Apple 12w/2.4a signaling, nor 2.0v Apple 10w/2.1a signaling. It's some weird low value. But since that is a backup to a backup, I waive that item - BC1.2 DCP is good enough. Those are reasons it is just "OK" and not "GREAT".

Another major issue is the cable is missing the shield connection. Meaning it may cause significant grounding errors if used, or data integrity issues. The Rp is fine at 56k, so it won't draw 3A, but this is still "bad".

The rest of the details are documented in pictures, since I'm short on time. All in all this makes for a fine portable battery pack to jump-start your Type-C phone at 3A. I'd recommend the RavPower if you want QC3.0 Type-A output or high-cap, and the iMuto if you want it small.

I still wish I could have tested the +Kanex Products GoPower before handing off/wrapping up this project. Kanex never responded to my e-mail requesting a sample. I don't like the fact they are recommended based off a secondary reports. I may reduce that rating since I will not be able to validate it.

PhotoAnimated Photo
[Plus] Analyses iMuto X5TC
29 Photos - View album
Add a comment...
Wait while more posts are being loaded