Lessons being learned...
... and some 'crow' consumed per prior rant ... (trust me, does not taste like chicken, no matter how much tabasco...) Issues and Challenges Taken Head-On... and some Take-Aways - download and try to 'build' node(s) with prior or latest build- 'bad' supply, new low voltage errors - enjoy the ease of asl-menu- why doesn't it recognize my eth0 and RA-40? We often hear "bad card..." - despite indeed some "bad card" attempts, changing methods allowed 'bad' cards to work. Probably NOT a bad SD card or in my case USB sticks. So far we can't even tell what specs/speed, etc. absolutely will work, absolutely will NOT work, or might work. We don't even know what internal schemes card makers employ in their products. I'm checking out a couple of supposed flash memory device diagnostics to see if anything might be obvious. I'm thinking a full-write and verify utility like h2testw (https://www.heise.de/download/product/h2testw-50539) may help - perhaps as a card 'conditioner' before writing an image to stick. This will not be a short-delivery cycle diversion... - I've variously used Win32DiskImager or Etcher in a wonderful Win10 workstation-class main PC... may be time to re-think that (haven't looked at Rufus, or unetbootin, etc. yet) and either seems to yield various results... but it's not very clear/apparent that various other system disk and USB I/O operations or just how Windows handles all of this. - before firing up a 'proper' Linux system to do all this with, or just use Raspbian on the Pi itself, I dug out my MacBook Air and Applepi-baker... which has proven quite consistent, reliable and happiness through over a dozen 'stick'/card image writes and saving images to reproduce to save time making other new nodes. I know - who just happens to have a 'dusty' MacBook laying about? Point is - is the problem Windows? The box x-nix is running on, or what? Not quite sure. - Bottom-line - unless you have a platform and tool you KNOW works, don't totally shame Windows - not there yet. "It's Your Power Supply..." - no, no it's not, unless/until it IS... whether using a generic USB-wart of 1.5A, *the* accompanying 2.5A 'wart' with many Pi kits, or a PowerWerx USB Buddy 5v 3A DC device ... "low voltage error" still can/has appeared out of nowhere with clean 5.1v into the MicroUSB. I've had no less than 4.93 at the connected RA-40 and 4.96 at the Pi - clean, scoped, flat DC. The Pi should not 'react' until 4.65v. - if we determine, somehow, that there is something imposing noise on the 5v bus or the Pi's internal 3.3v... maybe we can preempt this with a suitably placed add-on filter or two. TBD. "asl-menu didn't work..." - AH-HA! Big thanks to Steve and Nate for sleuthing this one FAST! - First take-away - start fresh - do NOT import any config files from "some other builds" - Yet to be identified, parsed, handled is what the custom/undocumented (?) non-ASL configuration items are... (but certainly there are standard parameter lines we built and want... save those off and bring them in one at a time after you have a good basic working node...) - Even with a fresh start... well, see next point... - It didn't know what to configure for 'eth0' because in some x-nix flavors, and the Pi, 'eth0' does not exist, but 'enxab12cd34ef56' does - the adapter "enx+MACaddress" - I don't know how or if this can be resolved... 'ifconfig' is your friend to get the 'NIC' identity to use for manual IP configuration steps. "Why doesn't Asterisk/rpt see my radio interface device?" - I don't have an answer for this one... back to the "bad SD card" conundrum perhaps... ------------------------------- Once I worked around and through the various challenges, I have established one good working node as a reference platform to clone, reproduce, edit, and deploy as a consistent upgrade for 4+ other nodes in our system. I captured a working image with ApplePi-baker and made 6 additional core working sticks. Onward! The goal is to off-load a handful of $400+ equally-oddly-custom-programmed multi-port controllers to reduce rack space, variables, power consumption... most interfaced with RA-35 or RA-40, and testing a newish comprehensive "pi-hat" out of Canada. I'm still working through a couple of points - mostly "cheat sheet" items to try and contribute yet-another-AllStar-how-to (OMG!) I've got 6-8 others lying about and they are all 'challenging' in many points - too assuming of knowledge of x-nix, paths, logs, scripts, permissions, skipping steps, apt-get, etc. This is a $35 platform 10-year-olds get to work making robots and media servers... we can do this... yes? Back to the rest of my 'crow' and following through. Thanks for listening.
The goal is to off-load a handful of $400+ equally-oddly-custom-programmed multi-port controllers to >reduce rack space, variables, power consumption... most interfaced with RA-35 or RA-40, and testing a >newish comprehensive "pi-hat" out of Canada.
Being that your goal is somewhat the same as mine was some time ago... I might suggest not using the Pi for primary radio control. To 'cantankerous' so to speak. I said a long time ago when everyone went for the small micros that I would never put my 'time-on' these devices since you are stuffing 10lbs of stuff into a 5lb package. But never to I want to stand in the way of development for anyone to figure out a way to 'do it'. It has been done and it's great. But not reliable enough for me everyday with the things 'I DO', your mileage may vary ! Been using Intel atom D2700 boards with 2 cores 4gb ram using Parallel port for i/o on a 12v power supply for 5 years now for radio systems. Somewhat limitless for most of the stuff we do and experiment with. The cost factor for me is no different for the path I took vs Pi's and BBB's but much easier to work with and plenty of resources to do it with. I'm not looking to update them anytime soon either. Utill a Pi is out there with the same resources, I'll stick to what I have and just play with the Pi's for cute novel things. So they might use 50-70% more power than a Pi3b+ and 4 times the size, but I get a repayment in the form of time messing with it and available resources. If I have the need for something more powerful, I can now make use of Virtual ASL installs on VPS to 'get-r-done'. Just my opinion, but it works for me. ...mike/kb8jnm. On 07/16/2018 10:10 PM, Jim Aspinwall No1PC wrote:
The goal is to off-load a handful of $400+ equally-oddly-custom-programmed multi-port controllers to reduce rack space, variables, power consumption... most interfaced with RA-35 or RA-40, and testing a newish comprehensive "pi-hat" out of Canada.
participants (2)
-
Jim Aspinwall No1PC -
Mike