| brolin_empey | Apparently Oksana is not here but the video at the beginning of this video made me think of her comment about electronic waste: | 01:02 |
|---|---|---|
| brolin_empey | https://www.youtube.com/watch?v=6locxQ0jzdc | 01:02 |
| DocScrutinizer05 | >><KotCzarny> i can confirm that 'no sim card/connection error' is caused by old battery<< age old wisdom, anyway nice you can confirm it :-) o7 | 05:35 |
| DocScrutinizer05 | the nasty/annoying detail inweak-bat-noSIM syndrome in N900: the modem *should* and maybe even *does* reprt a "LOW VOLTAGE!" error before going into malfunction shutdowm. Alas nothing in maemo userland seems to check for and report this | 05:39 |
| DocScrutinizer05 | at _very least_ the idiotic modem could set a flag in own local filesystem like "abnormal forced shutdown, usually by unexpected power down" and delete/replace this flag on regular solicited shutdown. Then on next power-up / boot of modem it reports value of the flag to server system aka maemo | 05:43 |
| DocScrutinizer05 | though I'm absolutely sure the modem could run long enough from buffer caps to report the very condition on power removal even like popping battery out of device | 05:46 |
| brolin_empey | It could be worse, such as how Windows Desktop falsely says that the disk is not formatted and asks the user if they want to format the disk if the user tries to access a flexible disc and maybe rigid disc AKA hard disc too that is formatted with a file system currently not supported by the Windows installation even though Windows has Installable File System capability for a user outside of Microsoft to add support for additional file systems. Whilst I was | 06:11 |
| brolin_empey | in secondary school, my classmate formatted the flexible disc he brought from his Mac at home because the disc was formatted with HFS instead of FAT and the Windows 95 computer at my school did not have an HFS driver installed. | 06:11 |
| brolin_empey | Also, if I recall correctly, when I tried to use an active Telus Mobility SIM in my N900, the software behaved the same as if no SIM was present instead of telling the user that the current SIM is active but is for a cellular network that uses bands not supported by the cellular modem of the N900. To be fair, though, I do not know if any platform tells the user about this situation because I do not know if the computer is able to distinguish between the | 06:19 |
| brolin_empey | lack of a SIM and the presence of a SIM that the cellular modem cannot use? | 06:19 |
| brolin_empey | I wonder if the cellular modem of the N950 supports the (UMTS) bands used by Telus Mobility and Bell Mobility. It no longer matters by now but it is difficult to recommend the N900 for use in Canada when the cellular modem of the N900 does not support the (UMTS) bands used by two (Telus Mobility, Bell Mobility) of the three (Telus Mobility, Bell Mobility, Rogers Wireless) largest cellular network operators in Canada. | 06:26 |
| brolin_empey | The human population of California is slightly more than all of Canada, though, so Canada is often lower priority than the USA. The original Apple iPhone from 2007 had the same problem as the N900 in Canada but Apple released the iPhone 3G the following year and released only the iPhone 3G and later in Canada because of the same limitation of the cellular modem as the N900 has in Canada, which may be why Nokia never released the N900 in Canada. | 06:32 |
| brolin_empey | Also, Nokia released some Series 40 “phones”, such as the 6103b from 2005, that refuse to let the user user them if a SIM is not detected when the OS boots even though the computer is still useful without cellular connectivity. Nokia also released some Symbian smartphones, such as the C6-00 from 2010, that have a microUSB connector but can only charge from a separate barrel connector, not from the microUSB connector. | 06:41 |
| brolin_empey | s/user user/user use/ | 06:41 |
| infobot | brolin_empey meant: Also, Nokia released some Series 40 “phones”, such as the 6103b from 2005, that refuse to let the user use them if a SIM is not detected when the OS boots even though the computer is still useful without cellular connectivity. Nokia also released som... | 06:41 |
| brolin_empey | I had a C6-00 when it was current and it took me a while to realise that I had to use the Nokia power supply with the barrel connector for power because microUSB was already the charging standard outside of Apple land by then. | 06:45 |
| brolin_empey | Also, I can no longer use my Geeksphone Revolution because the only battery I had for it bulged/expanded and was recycled and the lame hardware/firmware (?) design always powers the computer from the battery terminals instead of from the USB connector used to charge the battery. | 06:51 |
| brolin_empey | Is there a mini PCI-E card with or to which normal SATA connectors can be connected? | 09:57 |
| brolin_empey | My actual question is: on a computer like this: | 09:59 |
| brolin_empey | https://www.aliexpress.com/item/low-cost-intel-celeron-J1900-processor-itx-industrial-motherboard-for-vending-machine/32869595642.html | 09:59 |
| brolin_empey | can the mSATA/mini PCI-E sockets can converted to normal SATA connectors so I can use those ports for 2.5-inch or larger drives instead of only for mSATA drives or mini PCI-E cards? | 09:59 |
| brolin_empey | s/can converted/be converted/ | 10:03 |
| infobot | brolin_empey meant: can the mSATA/mini PCI-E sockets be converted to normal SATA connectors so I can use those ports for 2.5-inch or larger drives instead of only for mSATA drives or mini PCI-E cards? | 10:03 |
| Maxdamantus | brolin_empey: apparently: https://www.pbtech.co.nz/product/ADPSTT690384/StarTech-mSATA-to-SATA-HDDSSD-Adapter-Converter | 10:24 |
| Maxdamantus | Seems hard to search for though, since most results seem to be the other way round. | 10:27 |
| Maxdamantus | Here's one on ebay: https://www.ebay.com/p/Syba-Si-ada40113-mSATA-to-SATA-Converter/1778623054?iid=382390398201 | 10:29 |
| Maxdamantus | seems like a more reasonable price | 10:29 |
| Maxdamantus | eh, maybe not a reasonable price .. "+ NZD27.02 addl. costs" | 10:30 |
| KotCzarny | the only worthwhile adapters are marvel based | 10:31 |
| KotCzarny | and even then, heatsink is a MUST | 10:32 |
| Maxdamantus | That's not an adapter. That's a controller. | 10:32 |
| KotCzarny | adapter, controller, just naming | 10:32 |
| Maxdamantus | He doesn't need any controller. All he needs is a bunch of conductors for his signal. | 10:32 |
| Maxdamantus | It's not. | 10:32 |
| Maxdamantus | There is no chip involved in the things I linked, so there is no heatsink needed. | 10:32 |
| Maxdamantus | It's essentially just cabling. | 10:32 |
| KotCzarny | mSATA/mini PCI-E | 10:33 |
| KotCzarny | he doesnt know which one he has | 10:33 |
| KotCzarny | if the post is msata, yes | 10:33 |
| KotCzarny | if its mpcie, he needs marvel based solution | 10:33 |
| Maxdamantus | Presumably he has both. | 10:33 |
| Maxdamantus | They have combinations of these things on the same fittings. | 10:33 |
| Maxdamantus | so you can put either a PCI-E card in, or an mSATA cardp. | 10:33 |
| KotCzarny | yes, some socs have programmable functions | 10:33 |
| KotCzarny | but i dont know about intels | 10:34 |
| Maxdamantus | It's not programming. It's just reusing the same physical fitting. | 10:34 |
| KotCzarny | yes | 10:34 |
| Maxdamantus | The actual pins are exclusive. | 10:34 |
| KotCzarny | but you need different controller for both functions | 10:34 |
| Maxdamantus | (except maybe power pins; not sure exactly how they work) | 10:34 |
| Maxdamantus | If you have an "mSATA/mini PCI-E" slot, you already have controllers for both of those things. | 10:34 |
| KotCzarny | and for marvel socs you SET which function particular port has | 10:34 |
| Maxdamantus | Hm. I thought they used exclusive sets of conductors. | 10:35 |
| KotCzarny | https://en.wikipedia.org/wiki/PCI_Express#Mini-SATA_(mSATA)_variant | 10:36 |
| KotCzarny | i think you thought about usb part of mpcie slot | 10:38 |
| KotCzarny | in which case yes, those have predefined pins for usb function | 10:38 |
| KotCzarny | but msata/pcie function is muxed if supported by soc/board | 10:38 |
| Maxdamantus | Hm. So how does it know what controller it should be connected to? | 10:40 |
| KotCzarny | i just told you, you set it in bios/uboot | 10:40 |
| Maxdamantus | Ah. | 10:40 |
| KotCzarny | sometimes with pins on board | 10:40 |
| Maxdamantus | so in that case you still shouldn't need any chip. | 10:40 |
| KotCzarny | if his board supports such functionality, yes | 10:40 |
| Maxdamantus | anyway, the board he linked seems to just have mSATA and mini PCI-E separate. | 10:42 |
| Maxdamantus | (it has both, one slot for each) | 10:42 |
| Maxdamantus | if you want to connect it through the mini PCI-E slot, you'd need a mini PCI-E SATA controller. | 10:43 |
| KotCzarny | yup | 10:44 |
| Maxdamantus | https://www.ebay.com/b/Disk-Controllers-RAID-Cards-for-Mini-PCI-Express/90715/bn_2785152 | 10:44 |
| KotCzarny | #onlymarvell | 10:44 |
| KotCzarny | unless you dont care about performance that much | 10:45 |
| KotCzarny | then asm is fine too | 10:45 |
| KotCzarny | there are details such as number of pcie lanes controller uses etc | 10:49 |
| KotCzarny | but for a first buy, sure, go with cheapest | 10:49 |
| brolin_empey | Thank you for the links, I will try to get around to opening them. My reason for asking is that I want a small, low-energy, fanless x86-64 computer that can run Ubuntu 10.04.x LTS (for x86-32 if it matters) so I can replace a full-size ATX Core 2 tower computer that heats and adds fan noise to my bedroom but I want to connect at least three 2.5-inch or larger SATA drives using normal SATA connectors, not mSATA drives. I have a 2.5-inch SATA SSD for the | 11:01 |
| brolin_empey | boot drive and two HDDs contaning off-site backups of servers in my office: one 2.5-inch HDD and one 3.5-inch HDD that can be cloned onto a 2.5-inch HDD and is 3.5-inch instead of 2.5-inch only because 2.5-inch HDDs had not yet reached 2 TB in 2011 but 3.5-inch HDDs had. | 11:01 |
| KotCzarny | then your only solution is 4 port mpcie adapter in mpcie slot (not msata), msata could be used for boot drive | 11:02 |
| brolin_empey | Or install the 2.5-inch SATA boot drive on the computer I linked and connect the two HDDs indirectly via USB; this computer does not need to boot from the HDDs, only from the SSD. | 11:06 |
| brolin_empey | Is there a small, fanless 80686 or x86-64 computer that can have at least three SATA ports using normal SATA connectors without additional boards? I looked a few months/years ago but most seemed to have only up to two integrated SATA ports. I even found a small VIA C7 computer but it had only at most two integrated SATA ports if I recall correctly. I know that the VIA C7 is only 80686, not x86-64, but it should still work for this application because the | 11:11 |
| brolin_empey | current OS is only x86-32. | 11:11 |
| brolin_empey | Actually, I could install the current 2.5-inch SATA boot drive in my HP 2133 Mini-Note VIA C7M notebook computer to see if the current drive seems to work on the VIA C7. | 11:15 |
| Maxdamantus | Isn't the SATA port availability mostly an issue of what motherboard you're using, which shouldn't have much to do with fanlessness? | 11:20 |
| Maxdamantus | ie, presumably if you have a particular low-power CPU or something in mind, you should be able to find a motherboard with 5 or 6 SATA ports | 11:21 |
| brolin_empey | The current microATX LGA775 motherboard (Asus P5E-VM DO) has six SATA ports but the desktop Core 2 Duo or Core 2 Quad CPU becomes burning hot in mere seconds without a heatsink, which has a fan. Also, the full-size ATX tower case has at least one fan (on the rear, below the power supply when the case is vertical) and the normal/full-size ATX12V power supply has at least one, maybe two, fans. Is there a socketed LGA775 CPU that can safely and reliably run | 11:28 |
| brolin_empey | without a fan even if it still needs a heatsink? | 11:28 |
| brolin_empey | I have a nice, very efficient and quiet ATX12V power supply from SeaSonic but I had to stop using it after around a decade only because the fan developed a problem. | 11:30 |
| Maxdamantus | Maybe you need a bigger heatsink. | 11:31 |
| Maxdamantus | our old HP desktop (from around 2003 or so) had only a heatsink on the CPU, no fan, and that heatsink seems a lot bigger than the typical HSF that comes with Intel CPUs. | 11:33 |
| Maxdamantus | You're not just using a HSF assembly where you've taken the fan off, right? | 11:33 |
| Maxdamantus | heat directly corresponds to power consumption btw, so if you have a CPU that uses half the power, it will emit half the heat. | 11:34 |
| brolin_empey | It seems that usually the low-power (as in electric power, not capability) 80686 or x86-64 CPUs are BGA ICs or some other package meant to be soldered on an embedded motherboard, not a socketed package meant for use on a desktop motherboard. Note that “embedded motherboard” can include a notebook computer in this context. | 11:35 |
| brolin_empey | Which makes sense in practice because connectors often cause problems in the long term, which is bad if you need a computer to run continuously without servicing the hardware for at least a decade. | 11:38 |
| brolin_empey | And at my current age of 32 years, a decade seems significantly shorter than it did when I was younger. | 11:40 |
| brolin_empey | Is it possible to use a Core 2 ULV CPU on a desktop motherboard with a socket for the CPU? I used to have a Core 2 ULV notebook computer (a Dell Vostro V13) but I think the CPU was soldered, not socketed. | 11:45 |
| brolin_empey | I guess it could be possible if you could make a board to adapt the soldered package of the CPU, probably a BGA IC, to the socket of the desktop motherboard. | 11:47 |
| brolin_empey | Do any of the 80686 or x86-64 “mobile on desktop” motherboards use a socketed CPU? Are socketed CPUs still common in notebook computers now that the x86 CPU makers sell separate mobile/embedded and desktop/workstation/server CPUs? | 11:49 |
| brolin_empey | Maxdamantus: Which CPU did your 2003 desktop computer have? It may be pre-NetBurst, assuming it is an Intel CPU, probably meaning P6 microarchitecture, meaning Pentium Pro, Pentium II, or Pentium III or Xeon or Celeron using P6 microarchitecture, which may explain why the CPU could run without a fan if the airflow in the case was sufficient. | 11:55 |
| Maxdamantus | Some Pentium 4. | 11:56 |
| brolin_empey | NetBurst microarchitecture was used by the Pentium 4 and Pentium D. | 11:56 |
| brolin_empey | Intel used the Pentium 4 brand for around eight years, though, so that is vague. | 11:57 |
| Maxdamantus | I'm not sure how the microarchitecture is important, except in making the CPU more power-efficient, or more heat tolerant. | 11:58 |
| Maxdamantus | the case had a fan though, and there was a duct that directed that fan towards the CPU | 12:00 |
| Maxdamantus | (dunno if it was blowing in or out of the case) | 12:00 |
| KotCzarny | my i3-4130 board is very quiet, cool and power efficient | 12:00 |
| KotCzarny | using some crappy psu i get 20watt with 'off' and 40watt idling in win7 | 12:01 |
| Maxdamantus | seriously though: it's about energy consumption and heat tolerance | 12:03 |
| brolin_empey | Intel discontinued NetBurst partly because of the high power consumption and heat, same reason as why AMD stopped producing the Athlon Thunderbird (first Athlon for Socket A/Socket 462 after the original Athlon for Slot A) after the 1.4 GHz version. | 12:03 |
| Maxdamantus | anything involved in the CPU that affects how well it works with different cooling systems just leads to those two things. | 12:04 |
| brolin_empey | I thought the amount of switching in the CPU is important too because more switching means more electrical energy converted to waste heat; x86 computer designs are CISC, which I thought implies that they use more transistors than RISC designs such as the ARM, which implies that they switch more. | 12:08 |
| KotCzarny | evrything is cisc nowadays | 12:09 |
| KotCzarny | i mean risc | 12:09 |
| KotCzarny | with cisc layer built on top of that | 12:09 |
| KotCzarny | microcode etc | 12:09 |
| KotCzarny | intel learned the lesson when amd bought alpha and made athlons | 12:09 |
| KotCzarny | and a bit earlier, but production took a while | 12:10 |
| Maxdamantus | I'm not very well-versed in CISC vs. RISC, but my guess was always that RISC is able to be more efficient by reusing the same circuitry. | 12:13 |
| Maxdamantus | or rather, that CISC is less efficient because it uses more disparate circuitry; signals have to travel further. | 12:13 |
| KotCzarny | and that risc can be clocked higher | 12:17 |
| brolin_empey | KotCzarny: Yes but architectures such as the ARM have been designed to be RISC from the beginning in the 1980s, whereas x86 has been evolving from around 1978 (8086 but that evolved from earlier CISC designs in the 1970s) and still has to maintain the same CISC design used by the software even now over four decades later so x86 CPUs have even more transistors because of the overhead of converting the 1970s-era CISC interface used by the software to the | 12:17 |
| brolin_empey | internal RISC architecture of the modern x86 microarchitectures such as P5 (original Pentium from 1993) and later, which means converting from x86 instructions to micro-ops, then converting from micro-ops to field-programmable microcode, and so on. | 12:17 |
| KotCzarny | in short, dont use old tech | 12:18 |
| brolin_empey | If I recall correctly, the x86 evolution goes back at least as early as 1971 with the Intel 4004. I guess it actually goes back at least as far as 1968, when Intel was founded. If I recall correctly, Datapoint (a minicomputer architecture?) was the predecessor of x86 or what became x86. There was a man with surname Poor or Poore (?) who was associated with Datapoint who died a few years ago. | 12:24 |
| brolin_empey | Or the x86 ISA evolved from the Datapoint ISA? | 12:25 |
| KotCzarny | z80 ftw! | 12:26 |
| brolin_empey | IBM has a long history of maintaining backward compatibility in its computer designs because backward compatibility is important to businesses; I believe x86 is the oldest microprocessor architecture still commonly used for new designs. | 12:33 |
| brolin_empey | An x86-64 microprocessor from 2019 can still run object code from as early as 1978. | 12:36 |
| brolin_empey | Look how successful companies such as Commodore and Apple that keep breaking backward compability are with business users compared to companies such as IBM that have maintained backward compability as far back as the 1960s. | 12:40 |
| brolin_empey | Wikiwide: <brolin_empey> Apparently Oksana is not here but the video at the beginning of this video made me think of her comment about electronic waste: | 13:12 |
| brolin_empey | <brolin_empey> https://www.youtube.com/watch?v=6locxQ0jzdc | 13:12 |
| brolin_empey | Wikiwide: Also, it seems strange to me that the nVidia Web site categorises Australia as part of Asia instead of as Ocenia? | 13:13 |
| brolin_empey | Oceania. Maybe my font is too small after all. | 13:14 |
| brolin_empey | Here is the VIA C7 computer I mentioned: | 15:09 |
| brolin_empey | https://www.wdlsystems.com/VIA-Technologies-VB7009-16-Mini-ITX-1-6GHz-C7-D | 15:09 |
| brolin_empey | but I really should buy an x86-64 computer if I am serious about this project, which I am not because, honestly, it is a low priority for my life and I always have too much I want to do in too little time. | 15:09 |
| brolin_empey | Calling a purely electric car a “zero-emission vehicle” seems quite inaccurate to me because the vehicle still has a significant carbon footprint. | 15:27 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!