Root Access

bwDraco

00:09

It looks like the iPhone 6 and 6 Plus touchscreen controller issues are not new. ifixit.com/Answers/View/298409/…

Journeyman Geek

context?

bwDraco

Start here: chat.stackexchange.com/transcript/118?m=31963006#31963006

Bob

@ThatBrazilianHeadlessHorse Think of it like kickstarter - to an extent you're funding development of the game.

And/or they're trying to get you to spend money without seeing how bad the game is.

Journeyman Geek

@Bob and in the case of phyical copies helping the retailer work out how many to order

bwDraco

An iDevice repair shop isolated the problem to the touchscreen controllers literally coming off the main PCB due to the earlier "Bendgate" issues.

Bob

00:15

Meh. Premium hardware my arse.

Rule of thumb, every manufacturer has issues, expect failures.

Dog

I wonder what a BOB is

bwDraco

Apple didn't use underfill to reinforce the chips' bond to the board. This wasn't an issue with iPhone 5s (and there are valid reasons for not using underfill), but this, combined with the bending issues iPhone 6 had, is leading to a high failure rate.

If you remember the NVIDIA G84/G86 failures, that was caused by bad underfill that was not rated for the high operating temperatures of the GPUs.

Dog

Hey @Bob, do you know what a BoB is?

Bob

@Dog ?

Dog

I'm seeing BoB's mentioned in these LTE site upgrade schematics

Bob

00:19

...not a clue :S

Dog

But they're named after you!

Part number seems to be a custom one for this operator. But the closest thing is

> Fiber and Power Junction Box for 1 Hybrid Trunk / 4 Hybrid Jumper Cables with Internal Grounding

Bob

BreakOut Board? Box?

Dog

@Bob Hmm, breakout box makes sense

There's mention of a Commscope fibre+copper hybrid cable, so a breakout box does fit

And you still deny it when I say you're smart!

Hah!

Bob

@Dog Could also be a Big Orange Balloon

Dog

@Bob It's described elsewhere as grey, plastic, so it can't be a big orange balloon :-*(

Bob

00:26

lol

Anyway, home.icx.net/~ashburn/id1.html, cdn.sparkfun.com/datasheets/Sensors/LightImaging/…, context is fun

Dog

I'm pretty sure it's a breakout box. You were right. Amazing.

Bob

Amazing what my tea-deprived mind comes up with, sometimes...

Now if you'll excuse me, brb more tea

Dog

I wonder if you can figure out every other acronym on that schematic (BOB is the only one I didn't know, so if you can figure that out... :-o)

Bob

@Dog Probably not... was a lucky guess; I really have no idea about any of this.

Dog

@Bob It's just one of our operators bunging an LTE 2600 layer on top of an existing cell site.

Bob

00:30

Google to the rescue!

RRU => Remote Radio Unit

Journeyman Geek

lifehacker.com/…

Dog

@Bob Correct!

@JourneymanGeek o_0

I have an Anker USB-C cable :-o

Never used it though, since I have no USB-C devices

Bob

OBF => Optical Beamformer (god that was hard to find)

Journeyman Geek

@Dog I have one

Bob

RRH => Remote Radio Head

Dog

00:32

Also, a USB-C cable remembering voltage? o_0

Must be some sort of active cable I guess?

Bob

Huh. I thought BBU would be battery backup unit. baseband unit... ok then...

oh well it's straight after lte

not a clue what EE is

Dog

@Bob Could be either tbh. The Huawei BTS' have both in the same box

"E-marker" lol

@Bob The stupidest name for a mobile operator ever ("Everything Everywhere")

So a cable that's function and safety certified and has an "unmatched" e-marker chip, fries stuff?

Bob

MHA => Mast Head Amplifier

UCU => Universal Combiner Unit?

bwDraco

Anker makes good products. That's all I care about.

Dog

@Bob: Congrats, you now have a degree in RF engineering.

Bob

00:36

o.o

Journeyman Geek

@bwDraco I know for a fact that a good chunk of us tend to have more modern devices and are cable snobs

Seemed the sort of thing to pop in here ;p

Dog

Why does a basic, straight copper cable need the ability to "remember voltage" anyway?! wtf.

Journeyman Geek

@allquicatic is ~~a sith lord~~ iphone user so he should be fine since they use ~~sith~~ lightning cables anyway.

Bob

@Dog guy who found it commented here => m.reddit.com/r/Android/comments/5023o5/…

Dog

@Bob Well, I know RF engineers with degrees who don't know all those terms

bwDraco

00:37

It's just a way to tell consumers who don't understand that a correct resistor is required (or even know what a resistor is!) that the cable is standards-compliant.

Bob

@Dog never said I knew what any of em actually are. but my google-fu seems to be strong this morning.

Must be the tea.

bwDraco

Hmm...

Dog

> This particular combination of failures was unique to this C-to-C cable. Any C-to-C cable that claims "USB3.1 Gen1/2 support" has a little IC chip inside the plug.

Anker put an "extra resistor" in there that basically short-circuited the device detection safety. Any cable with this chip has the possibility for this screwup.

Oh come on

!!extraheaddesk

ChatBot John Cavil

@Dog That didn't make much sense. Use the !!/help command to learn more.

Bob

Anyway, as far as "remembering", that's just gawker being gawker, i.e. pants-on-head-retarded

bwDraco

00:42

Manufacturing error? Something seems to have failed with this chip.

Why is complying with the USB Type-C and Power Delivery standards so difficult?

Bob

Because complexity.

Dog

Well Android Police are using the same word

bwDraco

Anker's done a lot of things right. The occasional failure can happen, and the fact that they withdrew the product so quickly means they care about making good products.

Bob

s/making good products/their brand image/

bwDraco

Less honest companies will often just cover it up and continue selling bad product.

Bob

00:45

s/their brand image/not getting sued/

Dog

> [Attack of the Cables!] 11th #USB #TypeC analysis: Anker PowerLine 3.1Gen2 "100W" USB-IF Certified eMarked C-C cable. [Model A8185011.]
tl;dr: BAD. Spec-violating. eMarked 3A, not 5A. High IR drop. Lies about device, makes charger Vbus HOT. Somehow slipped by USB-IF Certification.

Journeyman Geek

shrug If you have one, and use it to charge a computer and a phone...

Bob

@Dog idk, then *shrug*

Dog

How is this much stupid even possible

Journeyman Geek

@Dog because cable makers arn't as snobby as you lot? ;p

Bob

00:46

I'm still waiting for my Tronsmart cables

Dog

@JourneymanGeek Not that implausible. I for example have 4-5 micro-USB cables dangling off my 6-port USB quick charger.

I just grab one, plug it into my phone. Next day might plug it into a tablet, or a watch, or a mouse, etc.

arda

I'm not feeling that bad for buying a regular micro usb type b phone after seeing all those terrible cables.

Dog

If it were USB-C I'd probably have a laptop connected at some point too

The cables themselves never get unplugged from the charger. So it's a totally plausible scenario.

arda

It is impossible to find high quality cables in turkey

Bob

s/manufacturing error/design fault/

Dog

00:47

Heck, how many people actually unplug the charger side of any laptop charger rather than just disconnecting it from the laptop?

bwDraco

@Bob Would using the actual !! s command be disruptive?

arda

@Dog same ones who remove drives safely

Dog

@arda Not me.

Journeyman Geek

@Dog that's only really an issue cause of USB C

The traditional chargers were hardwired to bricks

Dog

@JourneymanGeek Most USB-A/Micro-B chargers are not hardwired to bricks. Lenovo proprietary 20v USB cables aren't either.

Bob

00:48

@bwDraco Is not using it being disruptive?

CausingUnderflowsEverywhere

sorry @DavidPostill I was not online to read your message.

Bob

9 hours ago, by Bob

@allquicatic I do it with bot if I actually want to correct; without bot if I'm trying to make a joke

9 hours ago, by Bob

don't really like the bot doing an @Bob on the jokes

CausingUnderflowsEverywhere

Hello Mr. Arda ^^

@varfirstName , I am part roman..... mantic.

Journeyman Geek

Phones, maaaybe

@Dog most A/B cables arn't used at higher currents

arda

@CausingUnderflowsEverywhere hi

Dog

00:50

Considering the charger I have those USB cables dangling off is a QC 3.0 charger, then it's possible (and actually, I've done it deliberately) for it to fry a device if it accidentally "remembered" outputting 12v.

CausingUnderflowsEverywhere

Hey journeyman can I have a hug?

Dog

(Course someone would have to implement a USB cable that had a logic chip in it that lied to the charger, pretending there's a QC device connected, but there's nothing inherent about USB-A vs. USB-C that means it can't be done with A)

CausingUnderflowsEverywhere

oh dang, Sathya's name ends with an a. :)

oh yes you're on my hitlist

bwDraco

Most likely, Anker specified the wrong USB PD chip in the cable or made a mistake while designing the circuits around it.

Dog

> Furthermore, the cable appears to have a Rp inside of it that fakes the existence of a device. This breaks chargers and makes them Vbus HOT/dangerous. It will also break a ton of devices you plug it into -- since the Rp pulldown will be "doubled up". 5V Vconn voltages are also passed through to the CC pin (3.3 or 5v), which is BAD. 5v bleeding into a 3.3v circuit is BAD NEWS BEARS.

Fucks sake. Seriously. This level of stupid is something I'd expect from some $1.99 Ebay "fast charge" cable, not Anker.

CausingUnderflowsEverywhere

00:52

LOL NSA JOKE I LOVE IT @Hennes !

Journeyman Geek

@CausingUnderflowsEverywhere No. duuuuude.

Dog

On another note, the fact that a simple resistor can trick a charger/host into thinking there's a device connected? That sounds like a serious flaw in the USB-C spec itself.

bwDraco

It's also possible that the problem is the USB PD controller chip itself (at which point the designer of the controller is to blame).

CausingUnderflowsEverywhere

well isn't that what a battery does to charge? it resists the current?

are you saying you'd want all devices to talk over usb c to establish theres something there

Journeyman Geek

@Dog most of the issues with USB C cables are resistor related and using specific resistors is a simple/cheap/low pin count way to set things

Dog

00:55

@JourneymanGeek Which should not be possible.

Journeyman Geek

why not?

detecting resistance seems to be relatively simple

Dog

You can't simply "use a resistor to set things" to make a QC charger output 12v for example, it needs an actual device to explicitly request 12v via the data pins.

Journeyman Geek

unless you wanted active cables

Bob

@JourneymanGeek The biggest issue so far was cause Google was too cheap/stupid to properly protect their ports from overload.

Journeyman Geek

@Dog the issues with USB C cables and quick charging was with wrong resistor values mis-negotiating no?

Dog

00:57

@JourneymanGeek All USB-C PD cables must have an active chip anyway

bwDraco

Pretty sure this is an honest mistake. I will continue to buy Anker products. However, any product (regardless of manufacturer) that implements USB Type-C or Power Delivery will be subject to extra scrutiny because even the best manufacturers can make design mistakes.

Bob

They made the assumption that everything connecting to it would be perfectly to spec. They got burned. And now they're raising a massive fuss to protect themselves.

Journeyman Geek

also, having devices explicitly request probably needs some degree more intellegence.

Dog

@JourneymanGeek Again, shouldn't be possible

Simply having the "wrong" resistance on a cable should not be able to cause any connection to go into "fry device" mode.

Journeyman Geek

difference between comparing voltages/resistance/something vs "Hi, I'm an actual USB PD device. SHOW ME THE POWER!"

Bob

00:57

@JourneymanGeek Not even.

Ok, quick summary.

Journeyman Geek

@Dog are we talking the realm of possibility, or that strange lovely utopia of engineers who actually worry more about perfection than cost?

bwDraco

@JourneymanGeek lol @ "SHOW ME THE POWER!"

Bob

The whole resistor thing was meant to indicate that this is a USB-C-capable port and should be able to output 3 A.

Dog

@JourneymanGeek So you think it's OK to output USB-PD only voltages to a non-PD device just based on resistance? Because it's cheaper than implementing an actual PD negotiation chip?

Bob

That the devie was allowed to draw 3 A from the port.

Journeyman Geek

00:59

@Dog as someone who belives in proper engineering, hell no

Bob

Turns out some A-to-C manufacturers includes the resistor in the cable.

bwDraco

So is anyone here likely to stop buying Anker products because of this?

Bob

So the C device detected the A port as C and drew 3 A.

Journeyman Geek

Historically tho, smart cables have never been sucessful.

Dog

@JourneymanGeek USB-PD spec requires the cable to be smart.

Bob

00:59

Turns out certain Chromebooks didn't have adequade overload protection on their USB ports and burnt out.

Journeyman Geek

@bwDraco nope. But if someone owns that cable, they better do the needful.

Bob

So, dodgy cable and shitty laptop design.

Dog

If the cable isn't "smart", it is not USB-PD.

Bob

Google went on a crusade to highlight cable issues (while not addressing their laptop design issues...)

Dog

Also what @Bob said.

Journeyman Geek

01:00

@Dog and smart has variable degrees of smartness.

one could 'easily' design a brilliant cable that would prevent anything bad from ever happening. Probably at twice the size, and quadriple the cost.

bwDraco

My Griffin Technology A-C cables have not had issues with this. The device doesn't attempt to draw excessive power from any of my chargers. It falls back on the old approach to charging: sensing how much power it can safely draw and only drawing that much.

CausingUnderflowsEverywhere

smart monitoring ability right in time

Dog

I don't see your point

Bob

@Dog iirc the resistor thing wasn't used to change voltage. it was to identify the charger to the device, not the device to the charger. I could be misremembering though... wasn't too interested in the PD part of it cause I have no PD devices yet

Journeyman Geek

@Dog engineering is all about compromises. At this point, I think it tends to lean towards cost over pure reliability

Bob

01:01

Basically the resistor thing works the same way as Apple's thing. Or USB BC.

Dog

@Bob I know, but @JourneymanGeek seems to disagree with the fact that resistance alone shouldn't be able to trick a USB-PD negotiation

Journeyman Geek

@Dog I'm saying its a (poor) engineering compromise

Bob

@Dog I dunno how the resistor thing got involved in this Anker issue :\

Seems like two different issues.

bwDraco

Reddit's giving me 503s, dammit.

Dog

@Bob Apparently, the resistor thing is the whole Anker issue

(The eMarker chip indicating 3A instead of 5A is a completely separate issue)

Journeyman Geek

01:03

we can agree this even happening is a bit shitty?

Bob

The "USB-C resistor thing" @JourneymanGeek is talking about (see: Chromebook) was for A-to-C.

Dog

AIUI, the resistor in the Anker cable tricks the USB-PD host into thinking a device is connected when there isn't.

Bob

This is C-to-C and different...

Journeyman Geek

@Bob cause @Dog brought it up

Dog

Hence, you connect a 20v PD device, it requests 20v. You disconnect the 20v PD device, the host is unable to detect it being disconnected because resistor, and continues supplying 20v.

Journeyman Geek

01:04

and I was saying its a cheap way of setting a setting ;p

Bob

I'm confused.

@JourneymanGeek You're talking about A-to-C, yea? The cables that do not have an active IC in them at all?

cause I think @Dog is talking about the Anker C-to-C with the IC in there.

Journeyman Geek

@Bob I was talking the general use of resistors in electronics as a way of setting things

as opposed to active signalling

or just jumpering shit

Dog

@JourneymanGeek But the whole point is active signalling is required by the spec for any of this to exist in the first place.

Journeyman Geek

@Dog ah

I think I missed that

Dog

USB-PD does not exist without active signalling. It requires an active cable, and active negotiation by the device.

Bob

01:06

@JourneymanGeek The resistance thing is part of detecting if the charger is capable of C current limits (and abused by A-to-C). Not whether the device supports USB-PD - that's active.

Dog

Indeed, the A-C resistor shenanigans was basically where the resistor was an extension of USB Battery Charging spec.

bwDraco

Any USB Type-C or Power Delivery device needs scrutiny. If even the big guys are messing up, that means the standard is hard to correctly implement and extra care is needed to ensure that you're getting a compliant product.

If even Anker and Google (!) are geting it wrong, no single manufacturer can be blamed for compliance issues.

At least my Griffin A-to-C cables work as expected and do not cause anomalous behavior with my Nexus 5X or chargers.

Journeyman Geek

@bwDraco I don't really think we're blaming anyone ;p

Dog

USB-PD required FSK signalling over Vbus with v1.0 and BMC signalling over data pins on 2.0+ type-C

Without active negotiation with USB-PD, a passive port should only ever provide 5v, at up to 2A or 3A depending on resistance.

Bob

@Dog Oh wait I missed the "fakes the existence of a device" line. O_O

Dog

01:10

@Bob Lol

That was sorta key to the whole issue :-P

Bob

brb downloading USB-PD specs. again.

Dog

lol

CausingUnderflowsEverywhere

should be giving you 808's

bwDraco

Apparently, the USB-PD controller in the cable isn't detecting that the device was disconnected and therefore fails to signal that the charger needs to stop supplying power.

Dog

Funny enough that image contradicts the paragraph above it on the page it came from -_-

Bob

01:12

> Negotiation
•
Over V
BUS
only, no data line usage or reliance
Start
-
up
•
Legacy 5V
V
BUS
start, voltage / current adjust after
negotiation

bloody 60 MB specs

Dog

Heh, funny this article got published less than a month ago: extremetech.com/computing/…

"How USB charging works, or how to avoid blowing up your smartphone"

> If you take a phone which came with a 900mA wall charger, and plug it into a 2,100mA iPad charger, as an example, will it blow up?

In short, no: You can plug any USB device into any USB cable and into any USB port, and nothing will explode — and in fact, using a more powerful charger should speed up battery charging. We now do this all the time with our mobile devices here at ExtremeTech, and we’ve never had a problem.

Haha WRONG! xD

Bob

The whole idea of communicating with the cable plug just sounds... wrong.

Dog

@Bob Yeah...

Bob

And now they're introducing cable authentication.

Dog

That said, having a cable be able to identify what it's capable of is a good thing

Bob

01:17

Cable-DRM

@Dog This is one case where passive-ID makes sense.

Dog

But the charging voltage and current through any configuration should always =MIN(Host,Device,Cable)

Bob

IMO cables should be purely passive.

Which would've avoided this issue, too.

Dog

@Bob I agree, but

@Bob I'm not so sure about that

Bob

@Dog Hm. Still reading the spec, but based on the G+ post wasn't the fault that the Anker cable's chip failed to detect a device change?

Dog

If the USB-C spec's "detect when a device is disconnected" capability is defeated by a simple resistor, then IMO there's a bigger problem than just active cables.

@Bob Not as I understand it

Bob

01:20

A purely passive cable (resistance to ID what current it's capable of) and the voltage negotiation would be purely host and device like QC...

Dog

As far as I could tell it's the charger that could not detect a device change, because the cable somehow defeated the USB disconnect detection with a resistor. I could be wrong.

I don't see there being explicit mention of the cable pretending to be a device in and of itself

Bob

@Dog The info is really sparse, so I'm not sure...

Dog

@Bob Same

And I've not even read all of it

All I've seen with regard to the chip in the Anker cable is it identifies itself as 3A, rather than 5A it's advertised at.

bwDraco

This is the cable I normally use to charge my phone:

amazon.com/Griffin-USB-USB-C-Charge-Cable/dp/B00WF4WC0A

Dog

> Cord is illegal Vbus hot (!!!)

Has Rd pulldown inside the cable, for
“compatibility” with broken devices like Nokia
N1. Active circuitry complicates understanding
of failure mode.

Bob

01:23

@Dog USB-PC Rev 3.0 V1.0a § 2.6.1?

Dog

drive.google.com/file/d/0B2OJRSgNnm4GS05DQ3o1Y2hTNTA/view

bwDraco

I don't use the fast charger that came with my phone, and if it's non-compliant (such that the only safe cable to use with it is the supplied cable), I likely will not use it anyway.

Dog

According to that report, the cable identifies as "passive" and doesn't respond as a device

> Cable Termination Type
<0b00> Vconn not required (or Cable Plug only
supports Discover Identity)

Bob

@Dog Here:

Dog

@Bob Nice find

Bob

01:28

So... for the charger (source) to not reset, it needs to fail to detect detach (the resistor?) and also needs to keep receiving messages?

Hm. Failure to receive a response to a message. So if the source doesn't send messages..?

Is there a periodic message sent?

Wait. WAIT.

@Dog ^ (!!!)

So failure to receive a response leads to a soft reset... which doesn't reset voltages?!

Dog

@Bob Well apparently yes, but also apparently if the soft reset doesn't successfully renegotiate within 1-1.5s a hard reset is performed automatically which does reset voltages

Bob

@Dog Ah, must've skimmed over that. Me dumb.

@Dog How does the mentioned Rp map to that? Oh wait DFP perspective.

bwDraco

A standard that is difficult to comply with is a bad standard.

Bob

Yay, another spec to download.

bwDraco

I likely will never understand the standard itself, but if everyone is having trouble getting it right, something is wrong with the standard.

Bob

01:38

bwDraco

I've looked at the standard before.

Bob

@Dog ^ Attach/detach detection definitely depends on that resistor.

So. Either the charger is not sending periodic messages (and not receiving a response), therefore no soft reset, therefore no hard reset.

Or the Anker IC is (incorrectly!) processing those messages and providing a valid response.

Cause detach detection would never happen if the Rd pulldown inside the cable is pretending there's a device there.

bwDraco

I think part of the problem is the fact that the Vbus connectors are no longer just floating (from the perspective of the supply) when there is nothing connected because of the active nature of the cables.

Therefore, you can't just detect infinite resistance and reset the power output.

Bob

@bwDraco No, the attach/detach detection is specifically, by standard, based around whether the CC line is pulled to ground.

In this case, the cable seems to be acting as a device itself.

bwDraco

Hmm.

Anker might be using the chip in an incorrect configuration.

Bob

01:44

The C-to-A issue (cc @JourneymanGeek) was the cables pretending to be a valid USB-C power source.

Here it's the C-to-C cable pretending to be a valid USB-C power sink.

I'm still confused over why the soft reset doesn't happen.

Maybe it takes time, and only blows up if they switch it quickly?

@Dog Wait. How does the cable identify as passive, but fake an attach?

This whole thing is utterly broken o.O

bwDraco

Before USB Type-C, chargers can simply detect that a cable is disconnected using the simple fact that the pins are floating when no device is connected. USB-PD active cables invalidate this assumption.

Anker already has fixed cables on the market.

Dog

@Bob Yeah

Bob

@bwDraco Before USB Type-C, chargers didn't give a damn whether anything was connected.

By, by spec anyway. Excluding QC.

You did have various "smart" chargers that tried to detect the device type, but that's all nonstandard.

Dog

Also, I think "UCU" doesn't mean universal combiner, because there's another doc that says "X contains 3G flexi modules, Mk1 UCU, and combiners"

bwDraco

Regardless, this means you can't check for the absence of a load and reset. Cables that pose as devices (due to incorrect implementation of an active cable) only cause more problems.

😑

Bob

01:53

@Dog -_-

CausingUnderflowsEverywhere

you guys and maybe gals are so cool :)

Bob

@Dog Upper Control Unit? Paired with an LCU? For satellites though...

Journeyman Geek

@bwDraco there's a very simple solution for not understanding the standard. Just find a reference implimentation and follow it to the letter? ;p

Bob

> Now in use across the UK, the Universal Combiner Unit (UCU) has dramatically cut the operating costs and energy consumption of mobile telecommunications networks.

The UCU is a radio frequency filter combiner that allows mobile telecommunications operators to share common antenna infrastructure in the mobile network.

idk

Journeyman Geek

@Bob they're talking about doing it here

Dog

01:57

@Bob Fine, you're smarter than me

Heh, gotta love how advanced RF filters/combiners are just a box of air with a very complex set of holes drilled into it.

Bob

@JourneymanGeek Make a new standard*

Journeyman Geek

@Dog and those traces

Bob

@Dog I really don't have a clue, but that's all I could find for UCU

Journeyman Geek

@Bob gee, why would anyone ever do that... looks at all the dead USB varient standards

Bob

@Dog how the hell do they even caulculate this shit o.O

Journeyman Geek

02:00

@Bob also note the identical wierdly shaped PCBs with blobs

bwDraco

@Bob Simulation software that costs hundreds of thousands a seat.

Journeyman Geek

@bwDraco actually, antenna design in many cases is simple

that thing is 'simply' 2 identical units in a box

bwDraco

Aug 11 '15 at 19:32, by DragonLord

EDA software is an extreme example of this. Licenses cost hundreds of thousands of dollars per seat and there are only a few thousand licensees each year.

Journeyman Geek

@bwDraco quite literally an antenna is a piece of wire of the right length

usually 1/4 the wavelength

EDA is for designing the chips and other complicated shit

bwDraco

Yeah but it's not just antennae.

Dog

02:05

@Bob Magic/HFSS

bwDraco

We're talking about RF simulation software.

Journeyman Geek

MOST of that box is ~~magic~~ air and antennae

Dog

Journeyman Geek

@Dog heatpipes?

Dog

@JourneymanGeek RF waveguides

Lol, doesn't look that intimidating from the outside

Journeyman Geek

02:09

"cavity duplexer" sounds like a SR weapon...

Dog

Funniest thing is the latest and greatest microwave filters/combiners/etc. are purely mechanical devices

Journeyman Geek

There's just something about machined components <3

tho I wouldn't be surprised if those were cast, then machined

Bob

@JourneymanGeek In the "tear [you] a new arsehole" sense? :P

Journeyman Geek

@Bob in plural

Dog

Journeyman Geek

02:13

@Dog passive mechanical things are 'cheap' can be made with machines and reliable

Dog

@JourneymanGeek I'm not sure building those extremely high precision cavity boxes is particularly cheap

Bob

@JourneymanGeek Machining that probably costs a lot more than etching a PCB

Though most of that would be the initial price of the machinery

Journeyman Geek

'cheap'

yup

Dog

@Bob I think it's more that that kind of task just can't be done with micro-electronics for some reason

I know the high power stuff is usually mechanical, I mean those boxes can handle kilowatts, if not more of RF

Journeyman Geek

@Dog and those bits will never fail, overload, overheat...

Dog

02:17

@JourneymanGeek Well... valve amps?

It's just funny how pretty much everything RF has been gradually replaced by smaller and smaller microelectronics

Journeyman Geek

@Dog valve amplifiers are pretty fragile

Dog

But just this one standout niche piece of kit that's still built from giant hunks of metal

Journeyman Geek

and produce a ton of heat, and noisy (sometimes in all the right ways...)

bwDraco

19 mins ago, by Bob

@Dog how the hell do they even caulculate this shit o.O

!!/xkcd 1

ChatBot John Cavil

Dog

02:50

!!Caaaaaat

ChatBot John Cavil

Dog

So cute

bwDraco

03:35

WAFFLES!

Journeyman Geek

03:50

If a waffle is in midair... is it a luftwaffle?

bwDraco

04:07

ibuypower.com/Store/MSI-GS63VR-Stealth-Pro-Gaming-Laptop

They're already shipping Intel's new SSD 600p series?

(IMFT 32-layer 3D TLC NAND, NVMe M.2.)

intel.com/content/www/us/en/solid-state-drives/…

newegg.com/Product/Product.aspx?Item=9SIA1UH4T13989

NVMe storage for the masses!

Ramhound

darn you journey

that dog is cute

bwDraco

So the prices of NVMe SSDs have finally come down to earth. The SATA barrier is no more 😇

Journeyman Geek

04:26

bwDraco

I'm not too happy with the endurance numbers. While the Samsung SSD 950 PRO is rated for upwards of 0.4 DWPD (400 TBW, or 224 GiB of writes per day over five years on the 512GB version), the Intel drive is rated for a mere 72 TBW, or 40 GB a day.

Then again, your typical consumer isn't going to write more than 20 GB per day onto their drives under normal use conditions...

Journeyman Geek

are you going to write 40gb a day and use a drive for 5 years without backups?

bwDraco

CrystalDiskInfo's records indicate that I average about 15-20 GB a day on the main drive, with about 7 GB a day on the other small SSD from File History (see here for an explanation; this is primarily due to File History configured to back up every 15 minutes).

I'll probably dial back on File History since I haven't been doing much that warrants particularly frequent backups. The default is once per hour, after all.

Bob

!! s/ write 40gb a day and(.*?) for 5 years/$1/

ChatBot John Cavil

@Bob are you going to use a drive without backups? (source)

bwDraco

04:39

I need to learn regex.

Journeyman Geek

04:56

or that

tbh, my drives are at 2 years as a mainline drive

If I upgrade again, I might end up using an older SSD as a scratch/game storage drive

Bob

05:23

@JourneymanGeek I've been doing that with a couple OCZ drives :P

Journeyman Geek

05:47

The company I am contracting with called and are saying there's a delay in processing security clearances ._.

so I'm starting almost 2 weeks later than planned

Bob

O.O

Journeyman Geek

and blah, singapore employment law is from date of joining

Transcript for