Did WinSCP just fry my Onion?



  • I was preparing my 2nd 2+ for the 1-Wire fix by copying from the Omega to my laptop via WinSCP. Almost finished the transfers and suddenly the connection failed.
    From time to time I see this as we have approx 30 devices connected to the router and I didn't think much of it.
    Noticed the Omega LED was off and now it fails to reboot.

    LED comes on, flashes for 20 seconds and then goes off.

    The file I was copying was Nano which I wouldn't normally bother with but I thought I would back that up too. I don't use dock$ so my power feed is not perfect but it has been stable for the last week or so.

    Switched my AMS1117 in case that had blown but no joy.

    Could WinSCP fry my Onion?

    As an aside, for firmware upgrades are most files independent of each other with this "router" device i.e. would simply restoring nano after the upgrade be fine and the same for the opkg stuff? Or does everything need to be installed via opkg etc?



  • @Costas-Costas First of all, a device or component having been fried means literally that: physical, hardware damage from too high voltage/current. WinSCP can't do that. Secondly, no, just copying binaries around won't work, you need to also copy any libraries that got installed with them and which they depend on, too, and if they have anything in /usr or /etc.



  • No humour in the @WereCatf household. Thanks for providing your definition of frying.

    After 15 to 20 failed reboots and boosting the power to "12V" the #2, 2+ is back up and running.
    I normally feed 9V into a step down regulator to 5.41V and feed this into a AMS1117. Increasing the initial feed to 12V worked first time.

    Is nano a standalone executable or does it have associated libraries?



  • @Costas-Costas Yes, it does.



  • @Costas-Costas said in Did WinSCP just fry my Onion?:

    After 15 to 20 failed reboots and boosting the power to "12V" the #2, 2+ is back up and running.
    I normally feed 9V into a step down regulator to 5.41V and feed this into a AMS1117. Increasing the initial feed to 12V worked first time.

    This suggests your upstream supply is too weak, current-wise. By boosting the voltage you have your linear regulator wasting more heat on a regular basis (do check on that!), but you do increase the effectiveness of the capacitors between the upsteam supply and the linear regulator input - by starting at a higher voltage, a pulse load can draw more energy out of them before they fall to the point where the regulator's output drops. A better upstream supply and or more capacitance between that and the regulator would be better fixes, though it's understandable why turning an available knob is tempting compared to going out and finding something new.

    As for damage, consider the hierarchy of

    • Electrical or mechanical damage, components literally need to be replaced or resoldered
    • Corrupted flash state of U-Boot, no recourse but JTAG or directly altering the flash via SPI pins
    • Corrupted Linux kernel or root filesystem, but U-Boot provides software recovery via USB, wired Ethernet or (painfully slow) serial Kermit
    • Corrupted overlay filesystem, but main Linux is okay and can be recovered either via failsafe mode or by erasing the overlay from U-Boot.
    • Small error/mistake of settings in overlay filesystem that can be corrected while the system is running with that mounted


  • @Chris-Stratton said in Did WinSCP just fry my Onion?:

    This suggests your upstream supply is too weak, current-wise.

    He's just got shoddy wiring. He tried to argue with me and others in another thread that wiring-length and -thickness don't matter and, well, it looks like he didn't learn anything from the thread.



  • I learnt plenty WC.

    IMHO Omega's are not user friendly and for most of the world there are better alternatives.

    For a tiny breed of EE's they must be great fun.



  • It is indeed true that they are complicated systems which require care to apply well. If that is done, they do present a lot of possibilities.

    But threads are making it clear that these are indeed not simply drop-in Arduino replacements, and are not ready for all the of the uses they've been imagined for, without additional engineering that is yet to be done by the user, supplier, or both.

    It's really easy to understand the frustration from an end-user perspective; many engineers wear that hat too. However from an industry perspective, everything has issues - Arduinos are absurdly memory limited and in usual form have only a single UART, Pi's are weakend by the SD card dependency, irreproducibility and concealed internal details, while embedded WiFi stacks for micro-controllers take forever to wrestle to stability and require major rework for trivial changes, etc. Apart from the serious issue of missing source code, the main issue here is perhaps not so much the under-delivery, but the overly optimistic expectations that have been encouraged. Embedded is hard - one can admit that upfront, or discover it later.



  • @Chris-Stratton As usual, the marketing-department exaggerated things out of proportion and try to present the Omega2(+) as much more user-friendly than it really is. I think everyone would be better-served if they just straight-out admitted that it's not a very suitable product for beginners due to the many quirks and limitations of it.

    For a beginner/less-experienced user I, personally, would recommend either the C.H.I.P. or just a regular Arduino, depending on what they're looking for: C.H.I.P. runs a "real" Linux-distro and has 4GB storage, so it's a lot friendlier in many ways, and if one wants to really get to understand electronics and microchips on a lower-level, Arduinos have a shitton of information and tutorials everywhere and a billion gajillion different ready-made sensors and modules available. Hell, I'd recommend getting both and learning how they differ from one-another -- it's a good learning-experience.



  • Don't forget Pi's can use USB's for their filesystems like the Omega.

    ESP's are my weapon of choice. Just plug and play. Their relationship to Arduinos means there is a vast amount of info available too.



  • @Costas-Costas said in Did WinSCP just fry my Onion?:

    Don't forget Pi's can use USB's for their filesystems like the Omega.

    A USB filesystem is not an improvement for traditional embedded tasks, nor can most pi's boot from one anyway.

    An SPI NOR flash like used in a typical MT7688 system or the budget routers for which this chip is intended, is a far better boot source. It's basically the same approach an ESP8266 uses, only often with a larger NOR chip.

    For a compact embedded system that benefits from Linux complexity but needs to keep doing its thing with as little that can go wrong as possible, it can be really useful to run from a ramdisk and only save persistent variables in a very carefully chosen and controlled way, much as a microcontroller or your ESP8266 would.

    An MT7688 can be wrestled into such a constrained role where it might for some purpose directly compete with an ESP8266 (the half of smart outlets that don't use the ESP8266 typically use the AR9331 that was in the Omega1) but as a result of unfortunate decisions made at the factory an ordinary pi (especially the zero, which otherwise might be tempting) cannot.



  • Pi's and ESP's are doing just fine in the "mainstream" iOT sector.
    11 million Pi sales is something Onion must dream of.



  • @Costas-Costas said in Did WinSCP just fry my Onion?:

    Pi's and ESP's are doing just fine in the "mainstream" iOT sector.

    Really? How many actual soundly engineered products are built around a pi? There have been a few attempts, but it's not very suitable for real use, which becomes obvious when you seriously try.

    11 million Pi sales is something Onion must dream of.

    The primary market of the MT7688 is not hobbyists, it's gadgets, so forget about onions and start counting routers, cameras, switches, audio streamers, etc. And if you are going to count all varieties of pi, then you also have to reach back to when the design team that made the MT7688 was at Ralink and count those routers and ip cameras, too.

    Anyway, go ahead and experiment with a pi if it suits your needs; it's utterly unworkable beyond the demo stage for many others.

    ESP8266's on the other hand do find application in gadgets; of the side-by-side smart switch solutions on the market last year, the ESP8266 designs make more inherent sense than the competing ones based on the MT7688 preceding-competitor AR9331. But what makes theoretical sense and what you can get cheaply that will do the job on the day you have to commit the design can be distinct subjects as well. Unlike a pi, the MT7688 or AR9331 before could be reasonably shoehorned with a NOR flash into a smart plug where they are overkill, and be a good fit for tasks that are just enough more complicated to expose the limitations of an ESP8266.

    The roles you seem to want to push a pi for are instead being filled with things like the Freescale/NXP IMX, TI Sitara, and tentatively various far east tablet chips - take actual IoT products apart and that is the silicon you will find in the jobs at the level above what "router chips" can do.

    What you won't find in those is Broadcom's set-top-box chip, and in the actual set-top-boxes you won't find it relying on an SD card in the way it does on a pi.



  • Each to their own but I didn't mention MediaTek, the reference was Onion.
    What I'm pushing for is mainstream and Onion seem to be doing a pretty good job of ensuring that will never happen.



  • @Costas-Costas said in Did WinSCP just fry my Onion?:

    Each to their own but I didn't mention MediaTek, the reference was Onion.

    Onion is just making Mediatek slightly more accessible in some ways than it would otherwise be. There are several other modules for the same chip on the market, some of which provide what Onion has so far failed to. You have a range of choices (and Mediatek themselves have competition in Atheros), so the general solution is valid even beyond its particular embodiment.

    But with a pi, you're stuck with the actual literal pi boards - unless you have many millions and want to enter into lengthy negotiations, you literally have no access to the technology apart from those boards, and so are stuck with their unworkable choices.

    What I'm pushing for is mainstream and Onion seem to be doing a pretty good job of ensuring that will never happen.

    You seem to be overlooking that the pi had it's own notorious issues with schedule slip not to mention power supply issues.

    But pushing for mainstream in the manner of the pi would be a serious mistake, if it means something that similarly lacks utility, and if it continues to mean that resources get devoted into chasing too many over-promises rather than solving core problems.

    Do one thing and do it well - let the community worry about the variety.

    One would think, with a Linux based router chip that would be reliably boot up and talk on a network. That's an entirely achievable goal, but it hasn't been being treated as a priority.



  • SSH access has disappeared now via WinSCP and PUTTY.
    Onion have stated there are some "issues" with the new firmware.

    0_1486251875920_WinSCP-Omega-bash.JPG



  • Well, here's something that Onion could productively learn from the Raspberry Pi:

    https://www.raspberrypi.org/downloads/raspbian/

    Raspbian is the Foundation’s official supported operating system.

    https://www.raspbian.org/

    Note: Raspbian is not affiliated with the Raspberry Pi Foundation. Raspbian was created by a small, dedicated team of developers that are fans of the Raspberry Pi hardware, the educational goals of the Raspberry Pi Foundation and, of course, the Debian Project.

    Fairly early on, the pi foundation abdicated the task of being the primary provider of the software stack to the community, and that has worked so well that as the above indicates, the official stack isn't even one they author.

    In contrast, despite starting with a quite similar mix of mostly GPL components with a few proprietary bits, due to incomplete source releases Onion is still parked in the critical path of trying to be their customer's only system software solution.

    As many threads here amply demonstrate, getting that right for the envisioned diversity of use cases (or even at this point even just consistency for boot up and talk on the net goals) is an overwhelming job for a small company; even after empowering the community to solve the software blockers, hardware and supply chain concerns could probably keep all hands busy.



  • @Chris-Stratton said in Did WinSCP just fry my Onion?:

    Really? How many actual soundly engineered products are built around a pi? There have been a few attempts, but it's not very suitable for real use, which becomes obvious when you seriously try

    Interesting tangent this thread has taken. The Raspberry Pi is aimed at the education market, not the commercial product market. I don't think they could be more clear about that. Certainly they have been adopted by hobbyists and have found a niche in some products, but, except for the "compute module" variant the latter isn't something they are targeting. Moreover the Pi is more of a small server or high end graphic user interface sort of device with some GPIO, not really in competition with node devices like the Omega or ESP8266.



  • The Pi is closer to being a "$5 Computer" than the Omega is.



  • @Costas-Costas said in Did WinSCP just fry my Onion?:

    The Pi is closer to being a "$5 Computer" than the Omega is.

    Only if your definition of a computer is "something that can do nothing but complain about the lack of boot media"


Log in to reply
 

Looks like your connection to Onion Community was lost, please wait while we try to reconnect.