Post has attachment
This isn't a full #USB #TypeC eval: just a quick warning. The dodocool 20100mAh 45W PD battery pack has fundamental problems, and does not work with a Google Pixelbook. Avoid it.

This report and image is courtesy of +Jason Stanton​​​​​​, a fellow +Google Top Contributor Program​​​​​​ volunteer. Thank him for the notice, and being the unfortunate test subject.

When plugged into a Pixelbook, Chrome OS will briefly display a charging drop-down, which will disappear and reappear intermittently. The battery pack itself will flash LEDs. As a result, it is non-functional.

I do not know the mechanism why, as I do not have either a Pixelbook, nor this pack. So I cannot post traffic logs. (Read: you're going to have to trust Jason and myself on our word.) I suspect the DRP recognition logic (Google uses TRY_SRC) is defective in this battery, possibly among other things. Judging by the fact dodocool says in the fine print it is "not compatible" with some very basic Type-C devices, I suspect they know this product doesn't pass muster either.

Caveat emptor. Again, look for data supplied by technical experts you trust, before purchasing electronics. Look for the data firsthand. And don't trust Amazon blindly either! (This battery has 5* rating at the time of writing.)

Add a comment...

Post has shared content
A fellow pro-user pointed out some #USB #TypeC weirdness with his +R Λ Z Ξ R laptop, and a "USB-IF Certified" power supply. Thanks to +andy o for the initial report. (TL;DR: Google/Razer chargers works full speed, USB-IF ones don't.)

Having just spoken with a professional associate of mine that attended CES 2018, and hearing about Qualcomm QuickCharge 4.0 (i.e. highly discretized USB-PD levels), I was reminded of this "input voltage soft-DRM" issue. I suspect it will become more prevalent as time goes on, since companies aren't listening to feedback like mine. I'd bet money on it.

If you've noticed this sort behavior from any other products, please let me know. As I demonstrated, at minimum, the Google Nexus 6P and Google Pixel 1 Mini exhibit this behavior too -- at 4.95v or below, they will charge erratically. As a reminder, USB-C allows tolerance of 5.5v-4.0v (at cable end for 5v level, Figure 4-37 in spec). Similar for other PD voltages.

This is part of the reason why I refuse to comment on products until I vet them personally -- even though I've stopped testing for the most part. There's simply too much inconsistency across brands.

Even if a charger is "to spec", devices may still behave badly due to manufacturer decisions about implementation and "defensive design", as opposed to "robust design".

(My official position: If it's allowed in the spec, devices/peripherals should work 100% -- it's not a matter of "quality" at that point, it's one of basic functionality.)


+Nathan K. Would you happen to know if the USB-IF certification includes confirming the actual max power delivery from chargers, or does it only deal with safety and following USB spec (but obviously not the published charger specs for wattage)?

I ask because I bought a couple of PD USB-IF certified Nekteck chargers, the NK‐CC‐PD45W [1] for the car, which advertises 45W on the C port, and the TC‐AC‐60W [2], which is a rebrand of the CE Link PDS75‐4UT01, also certified with a separate TID number (so it's been certified twice over, I guess), and which advertises 60W on the C port. Both don't output even close to the max.

Testing with my Razer Blade Stealth 2016 (Kaby Lake), which can take up to 56W (measured) from the likes of the old 60W Chromebook charger, or ~43W from its own charger (rated to 45W), I can only get about 19.8V/~1.8A on both. I'm not 100% sure it's the chargers though, because of the suspicious coincidence that they both output about the same max amps to the laptop. But at the same time, the laptop is happily taking ~56W and ~43W from the Chromebook charger, and both the Anker Powerport+ 5 (V1) and Razer's own charger, respectively.

I'm using both the Satechi and the Plugable USB-C meters. Any insight on what might be happening is greatly appreciated. I'm leaning toward blaming the chargers, seeing Nekteck's past, but the coincidence that they output about the same capped wattage being so different chargers bothers me nonetheless, plus I know USB certification is not flawless, but is it that unreliable? That would be pretty disappointing.



PS: On the surface these seem extremely good value because of certification and also because they come with a certified 5A 1m cable!
Add a comment...

Post has attachment
[Study] I was able to run a #USB #TypeC charging study on the +Made by Google "Marlin" (HTC Pixel 1 XL) using its stock 18w USB-PD charger.
tl;dr: This phone charges at 16w@13% => 14w@52% => 7w@77%, dropping exponentially from there. Total charge time 1hr 58min (from 4%) but REALLY is 2hr 44min, due to 105% "hidden reserve"! (Remember: charge rate is NONLINEAR. The first portion is the fastest.) Max thermals at 44.0° Celcius. Preferred voltage 9v@2a.

This is also why I have extreme reservations about YouTube tests merely using a timer and the battery screen. They don't (and CAN'T) catch "tricks" by manufacturers like this.

Data dump for Marlin (Pixel 1 XL)

This is the most direct comparison I have to the Pixel 2 XL "Taimen" done below, that blew up. I'm doing "context control" now.

This Pixel 1 XL study highlights some of the shenanigans companies do with "indicated battery percent" to trick you into thinking your phone is charging fast (or slow). Notice the phone charges well after showing "100%" on the display. Also, the phone bugs out around the 10% level... this seems to be a common problem with Android phones!

(Edit: Upon further consideration, this actually seems to be another "safety valve" to trick people into charging their phones before they get TOO depleted. It dwells at 10% for 5% worth of battery, giving some time for you to listen to the warning prompt.)

I fit a fifth order polynomial curve to the "Level vs. Charge Counter" graph to show that, at actual "full capacity" by measuring electrons no longer flowing into the battery, the "indicated %" would have to be 105% to be accurate.

I suspect HTC did this as a "trick" for two reasons: (1) to prevent people from charging their Lithium-polymer packs obsessively to 100% max capacity, thereby degrading lifetime [ideal cycle range is 20-80%] and (2) to trick people into thinking the phone charges a lot faster than it actually does.

After all, the last "5%" of the battery takes over 45 minutes to fill. So just block that part of the battery out from being seen. REMEMBER: charging is nonlinear -- the beginning part is fastest. So you can inflate the charging stats (and lifetime) of your phone by installing a large battery, and not using all of it.

(This is a similar principle to the one people used with old SSD: "overprovisioning" to allow for better wear leveling and lifetime.)

How does this relate to the Pixel 2 XL? Take a look at the graphs carefully. The "2hr 30min" I reported for that was when the battery physically stopped accepting current. (The "cell_current" polarity inverted.) LG didn't use the "hack" HTC used here. However, the HTC still does indeed charge faster (16/14w) and hotter (44°C) -- so some criticisms still remain.

In conclusion:

This is why I think a far more accurate metric to report to people is the "Zero-to-Sixty" (to-Eighty) (to-100) times, including charge rate in Watts. Especially given charging is highly nonlinear, and the first part is fastest.

And given what I have just described about "hidden reserve" capacity tricks to further battery life, (or lie about your charging speed to customers,) perhaps manufacturers should more openly disclose such "hidden reserves" in the above stat so as to not game the measurement.

If you are familiar with the gaming of (mobile|desktop) graphics benchmarks by cheating companies using hacks in their drivers, this is the closest I can come to a battery analogue. LG displays the real stats, HTC (and others) pad those stats. They may have legitimate ends -- but the means makes competition unfair.

I'll try to come up with a more "honest" metric of my own design. Currently, it would be the ""Watt Hour" Charge vs. Time - Normalized for Battery Capacity" graph I made in my [Right of Reply] to the media running away with my Pixel 2 XL study without context. That one is below.

It is VERY hard to convince people to compare line graphs (or torque curves) -- and much easier to make a blurb like, say, "15 minutes of charge for 7 hours use". A good compromise may be the "Zero-to-Sixty" or "0-25-50-75-100" quartiles.

[Right of Reply] to Pixel 2 XL Study
"Watt Hour" Charge vs. Time - Normalized for Battery Capacity [Google phones]

4 Photos - View album
Add a comment...

Post has attachment
A quick announcement. While at the #TCSummit, I had the opportunity to hear some criticisms of my study on the +Made by Google​​​​​​​ Pixel 2 XL. As a result, I unboxed it again and am running (right now, in the assembly hall) a follow-up discharge study.
tl;dr: Charging rate is only as relevant as phone efficiency. I am now measuring the "power hogginess" of the Pixel 2 XL in my 'average' use to confirm (or refute) Google's claim of "15 minutes charge for 7 hours".

(I apologize in advance to the speakers whom I am quite distracted from. :( )

In an earlier post (that sort of ran away from under me), I said the Pixel 2 XL only charged at 10.5W. This is true and correct, and I stand behind my data. I still think a faster charging profile could have been used. (Technological parity with competitors, NVMe, etc.)

However, I did indeed do the engineers a disservice by only discussing the charge rate, without citing the discharge rate. I would like to take the opportunity to correct that. If the phone is super-duper efficient, the charge rate doesn't matter as much.

If you pay attention to the fine print, Google's "15 minutes charge for 7 hours" claim used some "50th percentile male" questionable models.... such as NOT using "always on" display feature, a major selling point of the phone. (No reasonable person would disable that, it's a main feature!)

Pixel 2
2700 mAh battery
Up to 7 hours of go with 15 minutes of charge1

Pixel 2 XL
3520 mAh battery
Up to 7 hours of go with 15 minutes of charge1

1Based on use of the included charger and a mix of talk, data, and standby use with always on display off. Actual results may vary.

This same problem affects crash safety testing using dummies. So I am running that test myself, to see how that claim lines up with "my" reality.

(Basically, the "50th percentile male" is an artificial construct that DOES NOT EXIST in reality. So crash tests based on that "average person" do not reflect when real-world people who are tall, short, skinny, or rotund get injured in accidents.)

The data isn't out yet. I'm literally running the test RIGHT NOW in the conference room. I'm making this announcement blind.

But I hope people pay attention to this, with the same intensity they listened to my notation about the charge rate. If the Pixel 2 XL sips current with as little use as Google claims, this is only good news for everyone.

I am 100% confident in stating I think the Pixel 2 XL is the BEST Android phone on the market right now, bar none. That isn't even up for debate. The hardware, features, and format are great. The only question is how marketing claims match up to reality. That is a relatively minor issue.

Data forthcoming.

Add a comment...

Post has attachment
Here's an anecdotal story of some of the sillier #USB #TypeC troubleshooting I do in meatspace. I got a kick out of this, and I hope you do too. (Props to +Russ Buchmann.)
tl;dr: If you are running Android 8.1 Beta on a Pixel 1 XL, and your phone ceases to charge: it's your phone, not the charger! The USB-PD chipset may be bugged.

While at the +Google Top Contributor Program #TCSummit , Russ recognized me in the crowd. He literally walked up with a dead Pixel 1 XL (and OEM charger + cable) in hand. It refused to charge for more than 2 seconds at a time. He just got off his flight, needed to call someone, his battery was at 2%.... not an ideal situation.

Luckily I am a bit of a screwball. I always carry around my Surface Pro and Twinkie derivative, a +Total Phase PD analyser. So while balancing my Surface next to a water cooler in the hallway, we logged the PD traffic to figure out why his phone wasn't charging. (And also compared it to another Pixel 1 XL, to isolate if it was the charger, cable, or phone.)

Problem summary:

Since USB-PD power supplies are "smart", they can reboot, change voltage, etc as needed. They're basically mini-computers. And if the communication traffic with a device appears corrupted, the charger ITSELF will start rebooting to try to clear the fault.

In this case, Russ' Pixel 1 XL (running Beta) is sending some REALLY shonky USB-PD traffic. The chipset isn't acknowledging packets (GoodCRC) correctly, leading to the charger thinking the connection is broken, so it continually reboots every 1-5 seconds.

Worse still, even if the negotiation proceeds smoothly by sheer luck -- there are random and spontaneous "Hard Reset" commands issued, causing the same problem. Yikes.

I suggested to Russ trying a Vbus Hot A-to-C cable to see if he could force it to charge, since there is no PD traffic and 5V is always-on, but I was unable to do logging internally since he had to attend to other matters. (Later he said it didn't charge when connected to a computer.... so the phone may be going back.)

I've attached sample logs. I don't expect to investigate this issue too deeply as Beta is... well... Beta, but it's interesting to see how by sheer dumb luck, having the right person, with the right tools, at the right time, can uncover some really interesting findings.

Also, if any other TC's are having phone problems, I'd be more than happy to help troubleshoot. Just be warned there may be some mad science involved, and I may be voiding some warranties. >:)

#USBC #TCSummit2017
2 Photos - View album
Add a comment...

Post has attachment
[Study] This is my final study on the charging of the +Made by Google Pixel 2 XL. Charging while "Screen On" and loaded (as if you were running GPS in a car).
tl;dr: Please avoid using the Pixel 2 XL heavily while charging. You run the risk of overheating your phone. Charging will actually continue -- the input will go up to 11W (from 6W) -- but only 2W will be directed to the battery. And thermals will spike to unhealthy levels, 48°C or more. I actually had to abort this test, due to concerns about temperature and "BATTERY_HEALTH_UNKNOWN".

To all: this is normal. The fact the Pixel 2 XL lets you run it hard, is a good thing. It means Google didn't put the thermal throttling training wheels on.

"With great power comes great responsibility". My advice is nothing more than the time-old "Don't charge your phone when it's hot".

The only thing even remotely questionable is the BATTERY_HEALTH variable. I do not know why that would ever go to 1 under "normal" operation.

(This is a followup of the test I described here, where I noted the Taimen (Pixel 2 XL) would limit battery charging rate to 6W if the screen was on.)
[Previous "Screen On" test, 0% CPU load]

In the comments, fellow users (rightly) inquired as to how the phone would behave if it was loaded with the screen on, as opposed to 0% (idle). I already discovered the ceiling battery charge rate (with screen on) was 6W -- and now wanted to investigate the charging behavior of the floor.

(Would energy be redirected to the CPU? Would the allocation of wattage to the battery increase or decrease? What about CPU and display? What would it do thermally?)

To run this test, I simply loaded a GPS and some other tester apps to simulate dashcam recording to replace a GoPro. (Not an atypical use in a car.) However midway through the test, I noticed thermals spiked to 48°C and I became concerned.

Moreover, I noticed the internal logging scripts started reporting "Battery Health" as "1". This was extremely dubious, since according to Android spec documents that represents "BATTERY_HEALTH_UNKNOWN". The phone had become confused about its own internal sensors due to the situation.

Using some math, I was able to subtract the "Battery wattage" draw from the "AC wattage" input and determine about how many watts were going into the battery charging circuit itself, and how many were going into the "Other/CPU wattage" category. This is how I determined out of the 10-11W AC input, only 2W is routed to the battery after a time. (The rest, 8-9W, is routed to the CPU and screen and burned off as heat.)

Data attached, including the point at which I noticed the questionable Battery Health field and aborted the test.

In conclusion: based off this data, the Pixel 2 XL WILL allow you to charge even when using it heavily, with a USB-PD car charger (like the Google Store sold +Belkin 27W F7U026) , albeit at a really low rate of only 2W.

However, the price you pay is unconstrained thermals that may risk harm to your battery health and longevity. So please don't do it!

3 Photos - View album
Add a comment...

Post has attachment
[Right of Reply] My recent study on the Pixel 2 XL's charging seems to have run away from under me. I wanted to push this out there to clarify something I think a lot of people (including Google) are missing -- if you'll pardon the rant.

tl;dr: Nobody sells cars advertising how many "miles per 30 second fillup" it gets. They advertise tank capacity, miles per gallon, and horsepower. (And torque curves if you're lucky.)

[Data dump for the Graphs below:]

Why do I bring this up in the context of phones? Because I feel product marketing teams have gotten far and away from their job of reporting plain, practical, comprehensive stats to consumers. Review sites are guilty of a variation of this, too.

Cars, like phones, have metrics "under the hood". We know they are important and look for them: Horsepower. Torque. 0-to-60 time. Braking distance. Fuel capacity. Miles per gallon, city and highway. Passenger seating.

Yet with phones, fans ignore all that and snap straight to incredibly nuanced metrics (that are meaningless in isolation) -- while marketing departments attempt to handwave all those numbers into a single blurb (that is a meaningless abstraction).

Both groups need to focus on those individual variables IN-CONTEXT, and how they interact. And customers need to be proactive and demand them.

For example, everyone knows the "City/Highway" MPG rating on the EPA stickers is only a rough estimate. As is "range". We intuitively know everybody drives differently, so actual MPG varies per person

Yet with phones, either that information is never published... or we get a blind promise from the manufacturer. Then we're correspondingly upset when we find out we don't get "8 hours SoT" like the marketing materials said.

(Another example: Porche claims its Panamera hybrid gets 30mpg highway. But we both know a sportscar with 380bhp will NOT be driven like Ms. Daisy, nor get that number. We would be upset if Porche sold the car ONLY with the promise it could get 600 mile range, without mentioning the rest of the facts.)

Furthermore, those "abstract metrics" sound to me like an excuse. If one brand can pull off a feature like real 15W charging, you need technological parity to be competitive. Simply because you can discern some artificial metric to compare your product, doesn't make that particular feature any less missing.

Along those lines, my data is only good analysed in combination with the other "under the hood" variables .

* Battery capacity in Watt-Hours (Joules)... not mAh. (Since cell voltage is a factor and ranges from 3.3 to 4.4v)
* (Typical) Efficiency of processor and screen in Watts. (Who cares if you have a mondo 10,000mAh battery if it's powering a plasma TV display and SD810.)
* Charging rate given in "Zero to Sixty" (to Eighty) (to 100) format. (Just like cars, phones might get to 60% really fast, but lag on the higher gears.)
* Honest charging numbers given in Watts. (Not people like me finding out the truth experimentally.)
* Phone size. (Your new super-slim battery may be stuck charging slowly due to physics limitations of anode/cathode.)
* Longevity of said setup. (Even Tesla admits SuperCharging kills batteries. PUBLISH the wear percent after given cycles! Or enable toggles to let people chose, rather than choosing for them.)


To demonstrate this, I put together two graphs compiling ALL the charge rates of the phones I have tested. I've even normalized the data for "true %" using the internal Coulomb Counter (instead of artificially calculated "indicated %"), and obtained the battery ratings from iFixit teardowns.

* Nexus 6P (3.82v 3450 mAh = 13.18 Wh) (Tri-stage 5v/3a charging = 15/12/9w)
* Pixel 1 Mini (3.85v 2770 mAh = 10.66 Wh) (Tri-stage 5v/3a charging = 13/11/7.5w)
* Pixel 1 Mini (Dual-stage 9v/2a charging = 13/7.5w)
* Pixel 2 XL (3.85v 3520 mAh = 13.6 Wh) (Mono-stage 9v/2a charging = 10.5w)
* Pixel 2 XL (Dual-stage 5v/3a charging = 10.5/8w)
* Pixel 2 XL (Mono-stage 9v/2a "screen on" charging = 6w)

Most users would immediately assume the Pixel 1 was the "best" phone merely because it charges in the shortest time. Others would say the Nexus 6P was "best" because it has the fastest circuit rate in Watts.

This is a fallacy.

Look at the first graph, "'Out of 100' Charge vs. Time". This shows how fast all phones charge from 15% to 100%.

Then look at the second graph, "'Watt Hour' Charge vs. Time". This accounts for a single "under the hood" variable by adjusting each line for battery size.

* Then factor in "battery longevity" -- that the Nexus 6P's charging was perhaps TOO fast for the Li-Ion cell technology to handle, and there were tons of reports of battery degradation.
* Then factor in "battery capacity" -- and suddenly the Pixel 1 Mini's "rapid charging" seems a whole lot less rapid. The Pixel 2 XL overtakes it and manages to shove more Joules down the pipe.
* Then factor in "processor efficiency" -- and although a Nexus 6P and Pixel 2 XL have roughly the same Watt-Hour battery capacity, the Pixel 2 XL's processor and display are leagues more efficient.

So long story short -- don't discount the Pixel 2 XL simply because its charging circuits are limited to 10.5W. There's a boatload of other stuff not accounted for.

Instead, be concerned that marketing departments aren't publishing the hard stats that matter. Ask why features that are present on their competitor's phone aren't on their phone. And be critical of review sites that aren't keeping an eye on all the OTHER variables that matter.

That includes me. Take my data only with a healthy dose of salt, and in context.

2 Photos - View album
Add a comment...

Post has attachment
[Study] This is data in preparation for an important post. I repeated the +Made by Google​​ Taimen (Pixel 2 XL) study in two anomalous conditions: (1) forcing 5v/3a, and (2) "screen on".
tl;dr: (1) Charging at 5v/3a exhibits dual-stage charging. Capped at a slightly lower 10.5W@35% => 8W@75%. Total charge time 2hr 54min (from 5%). Max thermals at 31.6° Celcius.

(2) Charging with the screen on REDUCES charge rate to 6W max! A literal case of "a watched pot never boils". Single-stage, capped at 6W@85%. Total charge time 4hr 12min (!!) (from 5%). Thermals constant at 31.7° Celcius (w/ zero CPU load). This is bad news for using your phone while it's plugged in, but has its reasons.

Taimen/Pixel 2 XL 5v Charging Data Dump:

Taimen/Pixel 2 XL 9v "Screen On" Charging Data Dump:

I'll briefly address the elephant in the room: slow charging while "screen on" is likely 100% intentional and common. For two reasons: (1) touchscreen insensitivity while charging and (2) thermal concerns of CPU dumping waste heat into the battery while charging.

Whitepaper on (1):
Notes on (2):

It's not ideal. Other phones have found ways of addressing this technical challenge. But it's here, and we have to deal with it. "That's all I have to say about that".

Charging at 5v/3a was simulated by using an Innergie 45w PD1.0 charger. This has an excellent voltage output, but old so is missing the 9v PD level. Therefore the phone is forced to use 5v. (It won't use 12v.) The Pixel 2 XL does show dual-stage constant current charging, so clearly there was some optimization done by LG. But it is still slower than 9v/2a charging by a fair margin. (Also note the amplified sawtooth current pattern due to the lower conversion voltage, there's some user theories about this.)

Charging at 9v/2a with "Screen On" was simulated by using "Wakey" app to force the screen to stay awake with a wakelock at 50% brightness. There was ZERO additional CPU load.

I am somewhat concerned about the cap at 6W. For two reasons:

First, you literally have to plug in your phone and walk away from it if you want it to charge in a reasonable timeframe. Even if you just want to browse the web or respond to SMS messages.

Second, even if you bought that expensive $50 27w USB-PD car charger (cough), if you so much as have GPS on screen or Spotify playing, your phone will not make full use of your fancy new charger's capabilities. This really sucks, since that's exactly when people (like me) need to charge their phones:

When I'm driving between job sites (or to a friend's house), with navigation on, listening to music, in one of the FEW places with a Type-C charger -- my car. And my phone is already low from the day.

Data attached above. Important post to follow.

#USB #TypeC #USBC #ShowMeTheData
7 Photos - View album
Add a comment...

Post has attachment
Someone pinged me asking if it was normal to see the "indicated battery percent" on the Pixel 2 XL go down from full, even when still connected to the charger.

Here's data suggesting the Pixel 2 XL:

(a) doesn't have a "battery disconnect" circuit, and slightly drains the battery after full. (Internal charge counter direction is reversed.)

(b) the internal "full charge" variable, and charge counter, recalibrate themselves upon full charge. (This MAY be a mechanism by which Google keeps track of battery wear.)

Graph attached. This data is indeed present in my previous spreadsheet, but I truncated it from the line graphs since it was extraneous. Feel free to use the ADB scripts to run the test for yourself, and come to your own conclusions.

Bonus graph: Here's a linear regression showing the "percent charge per minute" during rapid constant-current charging. (Technically the "charger counter" variable is more precise than "indicated %" -- but indicated percent doesn't need special scripts to be displayed.)

If you have documented proof of faster charging than m = 1.0497% per minute, please let me know. There's some conflicting reports about this.

Also please remember charging rate is only as relevant as phone efficiency. If a phone charges at 2W, but is super-duper efficient, during normal use it will beat out a 36W charging phone that is a power hog. I suggest giving Google some slack for the moment.

(I lack sufficient usage data to comment on the Pixel 2 XL's "power-hogginess".)
2 Photos - View album
Add a comment...

Post has attachment
[Study] I've completed another #USB #TypeC charging analysis on the +Made by Google "Taimen" (aka Pixel 2 XL) using the stock 18w USB-PD charger.
tl;dr: The reports of "slow" charging are 100% true. Charging is functionally capped at 10.5 watts... decreasing exponentially from that after 65% charge. Total charge time 2hr 30min (from 15%). (Remember: charge rate is NONLINEAR. The first portion is the fastest.) Max thermals at 32.6° Celcius. Preferred voltage 9v@2a.

Taimen/Pixel 2 XL 9v Charging Data Dump:

Remember that video I shared earlier saying "needs better data"? This is the level of integrity I expect. Wireless ADB was used in conjunction with a +Plugable Twinkie and +Total Phase PD analyzers to capture data -- from both external sensors, and ones internal to the phone itself.

Of note: the HTC Pixel 1 Mini used multi-stage Li-Ion rapid charging. Two Three stages of constant-current, followed by a final constant-voltage stage.

However the LG Pixel 2 XL uses a SINGLE stage of constant-current, followed by constant-voltage. This seems far less "optimized".

[Charging study: Pixel 1 Mini @ 9v/2a Apple 87w]
[Charging study: Pixel 1 Mini @ 5v/3a OEM 18w ]

This suggests to me Google or LG is trying to avoid strain on the battery in order to maximize longevity. Rather than opting for performance, they are being extremely conservative with the charging current and temperature. (Given the battery degradation problems reported on the Nexus 6P, I can't blame them.)

Note: HTC Pixel 1 Mini peaked at 40°C 44°C and 13W@45% => 7.5W@85% for comparison. Plus it had a smaller battery that filled faster. The Pixel 1 XL, on the other hand, had a bigger battery with different, faster logic. Hence "more optimized".

This seems a bit of a waste of Type-C "15w/18w" rapid charging to me. Especially when it hogs a whole (9v@3a=) 27w from USB-PD power supplies like the Apple 87w due to the "PSU hog" bug I reported earlier.

It will not even use the full 18w from the included Google power supply under ideal conditions! (Note: this isn't a critique of the PSU. It is still useful in certain situations.)

I feel Google could have done additional MCBF studies and picked a better combination of charge rate/method, that allowed for faster rapid charging while at the same time preserving cell cycle lifetime and thermals.

This appears very rough to me, especially combined with the "NAK" responses to DISC_ID and DISC_SVID. It is just weird. Even the Pixel 1 Mini responded to those... (see my study on the Microsoft Display Dock HD-500.) Maybe it is related to "juice jacking" protection in Oreo? However the same Pixel 1 "PSU hog" bug being present in the Pixel 2 makes absolutely no sense... that should have been patched out long ago.

Study: Pixel 1 and HD-500 Dock DISC_ID issue:
Exact timestamp of video showing HTC Pixel 1 and (buggy) DISC_ID
Pixel 1 Mini:
Lumia 950XL:
Pixel 1 "PSU Hog" bug:
Pixel 2 XL "PSU Hog" bug:
Note in posted data from internal sensors, the internal current draw never exceeds 2a. This means it can never use the 9v@3a it requests from the Apple 87w power supply.

(Calculations based off the internal cell voltage and current measurements tell me conversion efficiency is not a major issue.)

Pictures attached of testing setup. Scripts and formatted XLSX data shared at the Drive link above.

TODO: Repeat study at 5v@3a by forcing a PD1.0 power supply.

4 Photos - View album
Add a comment...
Wait while more posts are being loaded