#avr Logs

Oct 17 2020

02:18 AM rue_mohr: https://twitter.com/RueNahcMohr/status/1317341385993768960
02:25 AM rue_mohr: 6502, no ram, running lcd
02:25 AM rue_mohr: I only have 41 eproms left
02:32 AM vmt: using 6502s in 2020 is like gas guzzling f150
02:32 AM rue_mohr: yes
02:32 AM rue_mohr: its totally stupid
02:38 AM rue_mohr: but it does have 128k of program memory
02:38 AM rue_mohr: and will have 256k of ram
02:39 AM vmt: but it's a gas guzzler!
02:39 AM vmt: how about replace your avrs and 6502s with arms
02:39 AM vmt: because they don't guzzle gas!
02:39 AM rue_mohr: this 6502 thing is just a silly distraction
02:40 AM rue_mohr: even if it were a z80 it would have SOME Merritt
02:40 AM vmt: well i know you're not really into making things just tinkering with some random bs
02:41 AM rue_mohr: I need to be working on my power supply
02:41 AM vmt: that's what leads me to believing you might have an onset of schizophrenia
02:41 AM rue_mohr: I have an 8031 board mocked up on the bench too :)
02:41 AM rue_mohr: its from a while ago
02:42 AM rue_mohr: somehow, the simplicity of it all is attractive
02:43 AM vmt: well if it was up to you everything would be reduced to discrete components
02:43 AM rue_mohr: something about the idea of entering opcodes with a 20 button keyboard and having it run it, then dumping the memory to find your results
02:43 AM vmt: so, perhaps you should take some legacy hell chip from 70s/80s and design it out of discretes and sell that as a kit
02:43 AM rue_mohr: no
02:43 AM rue_mohr: I'm not good at selling things
02:43 AM vmt: you can do the same with any chip
02:44 AM vmt: if you really want to
02:44 AM rue_mohr: I think it I was gonna do it I'd need a sales person
02:45 AM vmt: you're already doing useless tinkering 90% of the time evidenced by your projects taking years and never reaching completion because you hop from one thing to another in no time flat
02:45 AM vmt: so why not try to design something useful?
02:45 AM vmt: well, useful in the sense of possible revenue
02:45 AM rue_mohr: I'v learned to not push myself to HAVE to finish one thing before I start another
02:45 AM rue_mohr: as you do things, you learn stuff, which can come in handy as you work on other things
02:46 AM rue_mohr: its like a horrid interlocked web
02:46 AM vmt: and never finish those either!
02:46 AM rue_mohr: and the only way thru is to just tug on the strings that will give till its all untangled
02:46 AM vmt: you have learned to be a slob, that is
02:46 AM vmt: no, you just spin more web all the time
02:46 AM rue_mohr: I'v had a lot of things I'v learned from the distractions come in handy on projects
02:47 AM vmt: i can't argue with that, but you just use that as a justification never to complete anything
02:47 AM rue_mohr: it can really be quite amazing how things come togethor sometimes
02:47 AM vmt: which is wholly fine by me, it just raises questions about your mental stability
02:47 AM vmt: that, and your whispering, your monologues, just to name a few
02:47 AM rue_mohr: sometimes finishing something isn't whats important about it
02:48 AM vmt: why do you even start then?
02:48 AM rue_mohr: sometimes, your learn more trying to do something than the worth of finishing it
02:48 AM vmt: that's wasting resources isn't it? i thought you're all against that
02:48 AM rue_mohr: some of the optimization stuff has lead to awesome paths
02:48 AM vmt: gas guzzling you know
02:48 AM rue_mohr: no, you learn things
02:48 AM vmt: you can't optimize something that isn't finished
02:49 AM rue_mohr: like I said, horrid intertangled web
02:49 AM rue_mohr: I finish lots of things
02:49 AM vmt: perhaps, akin to that of your mental state?
02:49 AM rue_mohr: but regardless of finishing, I learn
02:49 AM rue_mohr: brain thats evolving at one hell of a rate
02:49 AM rue_mohr: not sure if its fast enough tho
02:50 AM vmt: yeeea
02:50 AM rue_mohr: the 6502 turned out to be a great solution to a 16 bit counter as a 1 chip solution
02:50 AM rue_mohr: worked better than 6809
02:50 AM rue_mohr: which was a surprise
02:51 AM vmt: using a cpu as a 16 bit counter??
02:51 AM rue_mohr: yea
02:51 AM vmt: that's gas guzzling right there
02:51 AM rue_mohr: not really
02:51 AM rue_mohr: recycling maybe
02:51 AM vmt: no. gas guzzling
02:51 AM vmt: think of the inefficiency
02:51 AM vmt: it's like three F150s!!
02:52 AM rue_mohr: vmt one chip, 16 bit synchronous counter, 1Mhz, what would you use?
02:52 AM vmt: you need discrete logic!! and you need to build those out of discretes too!
02:52 AM rue_mohr: cmon
02:52 AM vmt: yes
02:52 AM rue_mohr: prove your supperiority on this
02:53 AM rue_mohr: ah ya cant do it can ya
02:53 AM vmt: what? on a 16b counter?
02:53 AM rue_mohr: yer just a shit-talker
02:53 AM rue_mohr: damn
02:53 AM rue_mohr: thought I might learn something new
02:54 AM vmt: you can find plenty of 16 b counters as ics??
02:54 AM rue_mohr: go on, tell
02:54 AM vmt: lmgtfy?
02:54 AM rue_mohr: cause there aren't
02:54 AM rue_mohr: lots of high-bitcount ripple counters
02:54 AM vmt: oh there aren't?
02:54 AM rue_mohr: no synchronous
02:55 AM rue_mohr: not as 1 chip
02:55 AM vmt: why is a single chip of any importance here?
02:56 AM rue_mohr: we should have progressed from building it from NAND gates
02:56 AM vmt: weeeeellll
02:57 AM vmt: it's roughly on par with using an arduino
02:58 AM rue_mohr: nope
02:58 AM vmt: probably is an order of magnitude more efficient as well
02:58 AM rue_mohr: synchronous
02:58 AM rue_mohr: you cant update 16 bits at once with an arduino
02:58 AM vmt: yep.
02:58 AM rue_mohr: and I dont think an stm32 can push 1Mhz to a 16 bit port, does it even have a full 16 bit port available?
02:59 AM vmt: you latch it to a buffer then
02:59 AM vmt: and i'm pretty sure they can
02:59 AM rue_mohr: heh
03:00 AM rue_mohr: cmon, your failing on a simple design requirement
03:00 AM vmt: just use a bunch of latches and there's your counter
03:00 AM vmt: less gas guzzling than a fucking 6502 lol
03:01 AM rue_mohr: so, you cant do it in 1 chip
03:01 AM vmt: of course i'm just fucking with you, but you're failing your own standards here
03:01 AM vmt: weellll given enough flip flops on the same die, yes you can!!
03:01 AM vmt: so, do it with an fpga!!
03:01 AM rue_mohr: mhm, so whats your 1 chip
03:01 AM rue_mohr: see
03:01 AM rue_mohr: I was waiting for that answer
03:01 AM rue_mohr: thats an acceptable answer!
03:01 AM vmt: but seriously, why is it important to be single chip?
03:02 AM rue_mohr: took ya long enough tho
03:02 AM vmt: you do realize it's not a measure of anything?
03:02 AM rue_mohr: I'd have also accepted PAL
03:03 AM rue_mohr: oooo
03:04 AM rue_mohr: I want to play with getting the fpga to operate a state machine thats stored in the LUT of a serial eeprom...
03:04 AM rue_mohr: maybe when the snow comes in feb
03:09 AM rue_mohr: Sir Antony Hoare, 1980
03:09 AM rue_mohr: Very simple ideas lie within the reach only of complex minds.
03:09 AM rue_mohr: http://kim.oyhus.no/ <-- that is an interseting page
03:09 AM vmt: i think a cortex-a soc is the best "single-chip" choice for a 16b counter
03:10 AM [1]MrMobius is now known as MrMobius
03:11 AM rue_mohr: hahah
03:11 AM rue_mohr: "But I've never seen the Icarus story as a lesson about the limitations of humans. I see it as a lesson about the limitations of wax as an adhesive."
03:11 AM rue_mohr: ^^^ hahaha
03:18 AM rue_mohr: I like the thinking-out-of-the-box solutions, the 555 is the lowest parts count RC oscillator, 3 components if you wire it right (555, resistor, and cap)
03:42 AM vmt: but an ic can be thought of as multiple components
03:42 AM vmt: so gas guzzling!!
10:49 AM nohit: some low power, small package cortex-m0 stm32 is the best and cheapest solution
10:49 AM nohit: it has 16 bit port
01:02 PM markasoftware: Cant you link together two 8-bit synchronous counters together?
01:03 PM cehteh_: you mean the hardware counters in AVR?
01:04 PM markasoftware: Nah there's a 4000 series chip for this
01:04 PM markasoftware: 8 bit, but you can connect multiple ones synchronously
01:05 PM cehteh_: read datasheet, i suppose that should work somehow
01:05 PM markasoftware: But I guess that fails the single chip requirement of rue_mohr
01:05 PM cehteh_: just need to have shared latch/start/enable whatever they use
01:27 PM vmt: the single chip requirement is really moronic to begin with
01:27 PM vmt: and i'm pretty sure he pushed it just for the argument's sake
01:28 PM vmt: as he firmly maintains everything designed in the 70s and 80s are vastly superior in every sense and everything else just isn't "low-level" enough for him
01:28 PM vmt: i think 6502s date back to the 70s... jesus h. fucking christ
01:38 PM rue_mohr: markasoftware, there are no 8 bit synchronous counters, only 4 bit
01:38 PM rue_mohr: markasoftware, those 4000 series are ripple counters
01:38 PM rue_mohr: but yea, the goal is 1 chip, I was able to use a 6502
01:39 PM rue_mohr: locked it into 0x49 on the data bus, which is LDA $49
01:39 PM rue_mohr: it counts the address bus twice for each instruction
01:39 PM rue_mohr: (unlike NOP, that has a dummy cycle)
01:40 PM rue_mohr: vmt, why is it moronic to limit a simple function like 16 bit counter, to one chip
01:40 PM rue_mohr: the fpga was a valid answer, it took you like almost an hour to think of it
01:48 PM vmt: there are sync 8 bit counters int he 74 family as far as i know
01:48 PM vmt: and you do realize how stupid it is to use a fucking cpu or an fpga just because you want a single chip solution???
01:49 PM vmt: it serves no purpose whatsoever to have it "single chip" instead of 2x cascaded chips
01:55 PM polprog: you can do single chip if you do PLD
01:55 PM polprog: but I havent taught him GALs yet :v
01:59 PM vmt: as rue_mohr lives in the 70s/80s, it's going to be nothing but through-hole for him, so you can get as "single chip" solution as a 6502, if you just stack 2x counters on top of eachother and wire them appropriately
02:00 PM vmt: you can fill the gap with epoxy and paint it black and there's your single package
02:05 PM rue_mohr: heh, 2020 and you need multiple chips to make a synchronous 16 bit counter
02:05 PM polprog: hehe
02:07 PM vmt: if you think there's a big demand, well then feel free to roll your own silicon and start selling it
02:08 PM polprog: ill make a verilog n-bit counter and sell you a licence for up to 16 bits
02:08 PM vmt: i can't believe it's 2020 and there's still no discrete solution for my latest-and-greatest ad-hoc rube goldberg involving N amount of 555s (everyone's favorite) and M # of discrete logic gates
02:08 PM polprog: i can even throw bidirectionality in for no extra charge
02:09 PM vmt: i also can't believe it's not butter
02:09 PM vmt: let M, N be a random integer between 9 and 10
02:10 PM vmt: polprog: that's pretty sick!
02:10 PM vmt: i'm already horny for that ip, i'm throwing money at the screen please take some of it
02:10 PM polprog: ;)
02:11 PM polprog: our CPLDs are made with 100% real fuses, slow blow kind
02:13 PM vmt: slow blow is at least an M20, i don't think your cpld stands a chance
02:14 PM polprog: my CPLD is only 5 racks
02:15 PM vmt: hmmmm
02:16 PM vmt: at any rate, i'm still entertaining a sick fantasy of interleaving quad uvc (or csi, i don't really care) buffered video streams with an fpga and feeding that to some reasonably fast micro or then straight to an sd card
02:17 PM vmt: but i might actually have to give up this perversion and just go for some arm sbc
02:48 PM polprog: hmm, i want to do some video stuff
02:49 PM vmt: you can do this, i'll pay you in discrete logic chips
11:50 PM day_ is now known as day