robbi5 has quit [Quit: Ping timeout (120 seconds)]
robbi52 is now known as robbi5
m4ssi has joined #glasgow
afkarl is now known as karlyeurl
ExeciN has quit [Quit: so long king Bowser]
setrofim has quit [Ping timeout: 268 seconds]
setrofim has joined #glasgow
hl has quit [Quit: Quit]
hl has joined #glasgow
hl has joined #glasgow
<whitequark> does anyone here have access to dssz.com?
<whitequark> oh, pudn also works
<hl> what are you looking for?
<whitequark> for very cursed reasons, naturally
<hl> interesting, is this leaked?
<whitequark> i would assume so
<whitequark> i actually don't want the core itself
<whitequark> i just want to know how its "OCI" works
<hl> downloading
<whitequark> thanks!
<hl> so apparently it's possible to procure datasheets from aliexpress, but I don't know the details
<whitequark> datasheets?
<whitequark> for OCI?
<hl> for chips, with NDA'd datasheets
<hl> also SDKs, apparently
<whitequark> right
<whitequark> unsurprising
<hl> but yeah, if anyone knows how to go about doing this, I'd be interested to know
<whitequark> thank you much!
<whitequark> let's see if that at least has the right IR size...
<whitequark> yes!!!
<hl> / Copyright (c) 2006 Actel Corporation. All rights reserved.
<whitequark> i think the nRF24LE1/LU1 have the same core
<whitequark> IT WORKS
<tnt> huh, what does ?
<whitequark> i can sample PC of the 8051
<whitequark> and i can single step it :D :D
<gruetzkopf> neat
<whitequark> gruetzkopf do you even have any LU1/LE1 devices.
Stormwind_mobile has quit [Ping timeout: 272 seconds]
Stormwind_mobile has joined #glasgow
<tnt> whitequark: oh so you just "randomly" found the source code of the 8051 core they use in the nRF to figure out its jtag debug interface ?
<whitequark> tnt: yes, i was not actually looking for verilog
<whitequark> i was just looking for docs
<whitequark> but... no docs... yes verilog... i'll take it
<gruetzkopf> whitequark: somewhere, yes
<gruetzkopf> a few LU1
<whitequark> gruetzkopf: ah. i made a bootloader for them and a bridge to the radio
<gruetzkopf> currently building a radio interface over usb thing for the cc2531 with the linux-wpan people to work with mac802154
<whitequark> coool!
<gruetzkopf> they built something like that for the atmel radios, plan is to have only one kernel driver for all of them
<gruetzkopf> linux is currently gaining 6lowpan route distribution support
<zeiris`> that's awesome
<whitequark> annnd i have code execution too!
m4ssi has quit [Remote host closed the connection]
<Stormwind_mobile> I'd like datasheet and SDKs for mediatek's MT61** chips, please.
<Stormwind_mobile> The SoC that they use in those 5$ crappy smartwatches. The OS and UI and features are terribly inconsistent, sometimes look like demo code, used without a second of thought if it actually works for the usecase, or just fake to begin with.
<hl> iirc some smartwatches use nordic RF stuff
<Stormwind_mobile> But if one could actually do open source programming on a unbeatably cheap platform, that would be sooo cool and potentially useful.
<Stormwind_mobile> But as it is, those watch-impostors are barely useful for telling the time - because you have to charge them every other day, and you need to turn the screen on manually
<Stormwind_mobile> One of my favourite UI fails: setting the alarm clock with a keypad with a size of 3 by 3 mm per key. With your fingertip.
<Stormwind_mobile> You need to hit this tiny area 4-6 times in a row with pinpoint accuracy. Or else the resulting incorrect input will be autocorrected to something completely different from what you intended, raising the number of keypresses required by another two or three. Which can also fail, of course.
<Stormwind_mobile> I really don't understand how so many people put so much work into building such a complex system, then produce millions of them, which are close to unusable in everyday life.
<hl> it's nuts isn't it
<hl> and even if it were usable, they'd expect you to dispose of it and get a new one in a few years
<Stormwind_mobile> Makes me think of one giant con job, where a series of hundreds of people managed to con the next one into buying the thing, by promising them what incredible value this is and how much profit they can expect to make...
<Stormwind_mobile> You can buy those for around 30-40 currency units on amazon, or for 5-15 on AliExpress, or get dozens of them as defective. I'm willing to bet that every single watch has been bought by some merchant who failed to make those promised profits and threw them out wholesale.
<Stormwind_mobile> Or they have been bought by a customer who realized within 10 minutes that this isn't even close to the Apple or Samsung
<Stormwind_mobile> device they copied from the outside appearance.
<Stormwind_mobile> And then get returned to Amazon, resold by some returned goods reseller, or dumped or whatever.
<Stormwind_mobile> I don't ask what consequences the energy and time and materials and fuel had on climate, but compared to the use humanity got out of the product, it's probably abysmal.
<awygle> i keep running into extremely useful chips, which have datasheets, but which don't exist on their manufacturer's website
<awygle> which makes me hesitant to use them
<furan-> looks handy
<awygle> like, the Realtek RTL8211EG-VB-CG is 1/4 the price of the KSZ9031, and is available on lcsc, but doesn't exist on realtek's website at all, and the docs are marked "confidential" lol
<gruetzkopf> definity seen that phy somewhere
<gruetzkopf> 90% of realtek chips seem to be in a state of semi-existance
<awygle> yep
Stormwind_mobile has quit [Ping timeout: 260 seconds]
<zkms> the power of the realtek crab~
Stormwind_mobile has joined #glasgow
<davidc__> awygle: Dunno if you're planning to use the KSZ part; but if so - check the Errata
<davidc__> I remember there being some ugly bugs in there
<awygle> davidc__: not anytime soon, but thanks for hte heads up
<davidc__> awygle: I used an older variant of the KSZ90x1 series ethernet PHYs on a board a while back. There were a number of bad bugs; including their reference design being so under-specced that a pass-transistor desoldered itself
Stormwind_mobile has quit [Ping timeout: 260 seconds]
<davidc__> And their timing skew adjust register not working, so I had to add a few ns of wire in a big loop to get the clk-to-data skew right for the FPGA