|
i generally like their poo poo across the board but i've never worked with their radios. are you using a soc or module or?
|
# ? Sep 11, 2015 20:37 |
|
|
# ? May 16, 2024 18:58 |
|
I'M not using any Atmel stuff unfortunately. I was working with a contractor on an 802.15.4 device, we were forced to use MRF24J40 but said contractor (who is a lot more experienced) said he'd have much rather used the Atmel radios. The MCU+Radio chips are ok i guess but they're literally two dies in one package iirc, the radio's just hooked up using SPI internally. Actually the reason why we're using the MRF is because it comes as a complete module that you can just plop onto your PCB, and the Atmels do not. Which means you have to get FCC certification for Atmel based designs whereas you can just piggyback on the FCC cert for the Microchip module. We think/hope.
|
# ? Sep 11, 2015 20:49 |
|
modules for rev1, but it's sorta irrelevant from the FW side. im at the dev kit bringup stage, can't speak to real bringup just yet TI had much better integrated options with the radio module in the same package as app processor, but that kinda locks you onto their stuff
|
# ? Sep 11, 2015 21:00 |
|
why is every online verilog resource either strictly about simulation/modeling or aggressively terrible am i the only person in the world synthesizing poo poo???? gently caress i long for death
|
# ? Sep 16, 2015 19:32 |
|
Bloody posted:why is every online verilog resource either strictly about simulation/modeling or aggressively terrible even my formal education in verilog was treated as a gauche lingua franca to implement all our awesome computer architecture ideas
|
# ? Sep 16, 2015 19:37 |
|
http://www.asic-world.com/verilog/first1.html#Hello_World_Program yup
|
# ? Sep 16, 2015 19:39 |
|
there should be a verilog to synthesize a giant 3d middle finger sticking out of the silicon imo
|
# ? Sep 17, 2015 00:05 |
|
i'm pretty sure that is the motivation behind MIMS. the worlds tiniest bird-flipper.
|
# ? Sep 17, 2015 00:27 |
|
A buddy visited some assembly shop, he said they had no ESD control whatsoever that he could see, no grounding straps or wires or anything at any of the stations. He commented on this. guy giving the tour: "oh we're all wireless here"
|
# ? Sep 18, 2015 14:00 |
|
i have inherited this monstrosity from a co-worker who is leaving: and the powers that be have decided they want 12 more made, i was thinking i could probably get a pcb printed and solder all these boards to it instead of wiring them all individually, someone give me a sanity check
|
# ? Sep 18, 2015 14:19 |
|
Yeah that'll work. As long as your pcb routing is actually correct and you don't mind the time and/or expense of waiting for boards from somewhere. The result is smaller, cleaner, and less prone to errors if making anything more than ones & twos. If time allows then I have done exactly that in the past. Otherwise just suck it up and do it all by hand but 10+ of anything really makes you notice the inefficiencies in any kind of assembly work.
|
# ? Sep 18, 2015 14:29 |
|
we can do small run pcb printing in-house so ill run it by those in charge see what they say
|
# ? Sep 18, 2015 14:30 |
|
we can too but its usually faster/cheaper to just go out of house
|
# ? Sep 18, 2015 14:32 |
|
Using a custom made pcb to simply glue together some breakout boards might seem weird in a way but in a situation like what you show it's really just making the wiring job smaller, cleaner, faster to assemble, and less prone to errors in assembly.
|
# ? Sep 18, 2015 14:34 |
|
Bloody posted:we can too but its usually faster/cheaper to just go out of house quite true Mister Sinewave posted:Using a custom made pcb to simply glue together some breakout boards might seem weird in a way but in a situation like what you show it's really just making the wiring job smaller, cleaner, faster to assemble, and less prone to errors in assembly. this was my thought entirely, i can't be arsed to sit down for a week wiring up/debugging a million of these, so anything to offload that work is fine by me
|
# ? Sep 18, 2015 14:36 |
|
do you have lab techs?
|
# ? Sep 18, 2015 14:36 |
|
yep, thats me
|
# ? Sep 18, 2015 14:37 |
|
at least, im the odd job guy who it falls to to manufacture this sort of stuff a post doc will go "Im conducting a study and i need something that does this, go find or make one for me"
|
# ? Sep 18, 2015 14:38 |
|
oh
|
# ? Sep 18, 2015 14:45 |
|
sorry by proxy for all of the horrendous garbage ive dumped on people like you!
|
# ? Sep 18, 2015 14:46 |
|
the cool thing about grad school is that as soon as you'd need to support something you're gone!
|
# ? Sep 18, 2015 14:48 |
|
Bloody posted:sorry by proxy for all of the horrendous garbage ive dumped on people like you! Ive got to make some cool stuff actually! gimmie a sec to rustle up a picture of something
|
# ? Sep 18, 2015 14:51 |
|
this is an electronic punching bag i helped some high school interns make, those silver things are buttons we made and it uses the location of a punch in tandem with an accelerometer to calculate force, it also has a led array on it as a kind of rudimentary screen
|
# ? Sep 18, 2015 14:59 |
|
thats actually really cool
|
# ? Sep 18, 2015 15:00 |
|
in that plastic box is 2 arduino megas that handle all the input and output, everything is calculated on the computer, it has a high score function too
|
# ? Sep 18, 2015 15:04 |
|
I like to imagine that sign refers to the punching bag itself
|
# ? Sep 18, 2015 15:50 |
|
makign and troubleshooting a pcb seems like a lot of hassle for 12 units
|
# ? Sep 18, 2015 16:54 |
|
Mido posted:makign and troubleshooting a pcb seems like a lot of hassle for 12 units they had previously told me they might have wanted up to 50, but they seem to have relented Mister Sinewave posted:I like to imagine that sign refers to the punching bag itself the punching bag is tired
|
# ? Sep 21, 2015 11:57 |
|
does anyone have Strong Opinions about message pack or protocol buffers? i have a lot of very regular data that i need to punt up to the cloud with a bunch of goofy embedded constraints (like samples being 3b instead of something reasonable) and i'm not going to package things up in JSON
|
# ? Sep 24, 2015 17:17 |
|
protocol buffers require a schema, but will be smaller if you want to use message pack but need to optimize for size you'll end up giving fields one-letter names
|
# ? Sep 24, 2015 20:51 |
|
thanks! I'm really sensitive to size, and one end of this is in C which seems to rule out protobuf. i know the format of the 31bytes I want to send. at this point i know enough to spin up a few dummy test cases and work them out. some of the message pack examples looked like they were outside of k:v pairs in JSON format though debating asking the gray forums
|
# ? Sep 25, 2015 23:29 |
|
JawnV6 posted:thanks! what is the thing running C? if it's an OS you can still use google protobuff and just link it cleverly contriving my own stupid schema is never a problem, the most difficult thing is keep tracking of MSB/LSB madness depending on the platform/what is serializing/deserializing in the chain of things
|
# ? Sep 25, 2015 23:39 |
|
Mido posted:what is the thing running C? if it's an OS you can still use google protobuff and just link it cleverly im in enough fights with my tools that I don't want to add another language though. and there's actually another programmer around this time, so I don't have the luxury/curse of writing both sides
|
# ? Sep 25, 2015 23:44 |
|
JawnV6 posted:m4 hmm, well if i were in your situation i'd probably still use protobuff and just make a little script that generates a static library from the protobuff schema file that is linkable (there's probably reasons it's not this simple) failing that -- let me know what you end up going with/link your greyforums post here
|
# ? Sep 25, 2015 23:56 |
|
JawnV6 posted:m4 here's the wire format for protocol buffers, fwiw https://developers.google.com/protocol-buffers/docs/encoding it wouldn't be that hard to mash together valid messages ad-hoc if you're just sending if you're receiving it's sketchier but i'm pretty sure there exists c bindings, even if they're not official
|
# ? Sep 26, 2015 00:31 |
|
I'm pretty sure the C++ req is for the repeatable elements and leveraging the class structure to handle the goopy bits the C side just needs to emit, for downlink stuff it can be totally different formatting cool, protobuffs back on. idk if i mentioned but the 3b values will probably end up as 4b for Other Reasons
|
# ? Sep 26, 2015 00:58 |
|
what about capn proto?
|
# ? Sep 26, 2015 05:04 |
|
Bloody posted:what about capn proto? mentioned in my gray post, but I'm not wild about it looks like less support and I'm not crazy about moving all in-memory representation over to something goofy. i still like being able to navigate a memory dump on paper & pencil.
|
# ? Sep 28, 2015 18:25 |
|
i work out my dumps with a pencil too!
|
# ? Sep 30, 2015 15:43 |
|
|
# ? May 16, 2024 18:58 |
|
bump
|
# ? Oct 8, 2015 04:51 |