#garfield Logs

Jul 08 2020

#garfield Calendar

12:00 AM rue_mohr: hahahah
12:05 AM rue_mohr: and a z80 cant go up to 8Mhz, so I cant use it as a counter (and its only 16 bits)
12:05 AM rue_mohr: HOWEVER
12:05 AM rue_mohr: a 386 could
12:05 AM rue_mohr: hahahaha
12:26 AM rue_mohr: polprog, oh
12:26 AM rue_mohr: well
12:27 AM rue_mohr: better you use it for a computer than I use it for a counter
12:28 AM polprog: heh
12:28 AM polprog: dont worry
12:29 AM polprog: ill find one someday :v
12:29 AM rue_mohr: I can send this to ya
12:29 AM rue_mohr: I'm sure there are a number of cpu's around I can lock into NOP cycles to be a counter for me
12:29 AM polprog: heh. clever
12:29 AM rue_mohr: :)
12:30 AM rue_mohr: I just need something with 17 or more address lines
12:30 AM rue_mohr: the 386 is looking good
12:30 AM rue_mohr: they dont mux the address lines on a 386
12:30 AM polprog: nop it and make it jump to zero on 2^17
12:30 AM rue_mohr: na let 'er roll
12:31 AM polprog: ive veen trying to get a 486 board to run yesterday
12:31 AM rue_mohr: k
12:31 AM polprog: because i finally cleaned the bench up
12:31 AM rue_mohr: whats the 29k for tho?
12:32 AM rue_mohr: I have no idea if its ok...
12:32 AM polprog: just pure curiosity. i read about it on wikipedia once
12:32 AM rue_mohr: hah
12:32 AM polprog: that they were used in printers for a while, and i'm looking for one just because I haven't seen anyone do a homebrew computer with it
12:33 AM rue_mohr: I need to know if the 386 is 5V signalling
12:33 AM rue_mohr: yea
12:33 AM rue_mohr: vcc is 5V
12:34 AM rue_mohr: the 386 will totally work
12:34 AM polprog: im pretty curious how will you hook it up
12:34 AM rue_mohr: this would be halarious
12:34 AM rue_mohr: a 386 and a 1M eprom
12:35 AM rue_mohr: you would think computer
12:35 AM rue_mohr: but no, counter and sine wave table
12:35 AM rue_mohr: hahaha
12:36 AM rue_mohr: I still want to dead-bug a 386 computer
12:36 AM rue_mohr: run it at like 4Mhz
12:36 AM polprog: haha
12:36 AM rue_mohr: ah cool ok
12:37 AM polprog: i think 386 is still possible
12:37 AM polprog: not surr about 486
12:37 AM rue_mohr: I need to find the more plausable 386 for this
12:37 AM rue_mohr: I dont know if I can use a 286
12:37 AM rue_mohr: maybe I can
12:37 AM polprog: "Several versions were introduced during the period from 1988 to 1995, b"
12:37 AM polprog: from 29k article on wikipedia
12:37 AM rue_mohr: 486 needs the cache and junk
12:37 AM polprog: i need to find a dead printer from that period..
12:37 AM rue_mohr: this one is "Am29030-25GC"
12:38 AM rue_mohr: (1991 / 1992)
12:38 AM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1160543.jpg
12:39 AM polprog: neat
12:40 AM rue_mohr: I'm not a retro guy tho! >:|
12:40 AM rue_mohr: :)
12:40 AM polprog: ill have to look for something to send you back :)
12:40 AM rue_mohr: I dont think the 286 is fast enough
12:40 AM rue_mohr: I dont know, even with the 386, how many cycles the addressing takes
12:41 AM rue_mohr: haha
12:41 AM rue_mohr: polprog, I bet that 29k IS from a printer
12:41 AM rue_mohr: a laser
12:41 AM rue_mohr: a compaq pagemarq
12:42 AM polprog: hmm..
12:42 AM polprog: well the laserjet in the basement had asics in it
12:42 AM polprog: not even a flash memory iirc
12:42 AM polprog: unless it's hidden somewhere deeper..
12:43 AM rue_mohr: I might have more of them
12:44 AM rue_mohr: looks like 4 cycles per address
12:44 AM rue_mohr: so I'd need 32Mhz
12:44 AM rue_mohr: and the 286 didn't come that fast
12:44 AM rue_mohr: furthermore, the 386 requiers 2x clock input
12:45 AM rue_mohr: so if its the same, I'd need a 64Mhz oscillator and like a 40Mhz 386
12:45 AM rue_mohr: oh, and this one IS 40Mhz
12:45 AM polprog: haha i just realized that this is for the inverter
12:45 AM polprog: hahahaha
12:46 AM polprog: put some more parts and you will have a remotely controleed one, via a web server ;)
12:46 AM rue_mohr: I KNOW EH!?
12:46 AM rue_mohr: I can use 3 74LS393 + 74LS74
12:46 AM rue_mohr: or ONE 386
12:46 AM rue_mohr: or ONE 386-40 ;)
12:47 AM polprog: what is the dac gonna be?
12:47 AM rue_mohr: its not!
12:47 AM rue_mohr: thats whats kicking my ass
12:47 AM rue_mohr: its pwm
12:47 AM polprog: oh
12:47 AM rue_mohr: the rom directly outputs two channels, one for the high side and one for the low side
12:48 AM polprog: i wanted to try and make a small r2r music player using an eeprom but found not enough counter ICs
12:48 AM rue_mohr: I'll keep trying to do it with the atmega88, but if the cpu clock really does have to be more than 4x the timer clock, it cant do it
12:48 AM rue_mohr: :) well there
12:49 AM polprog: breakfast time :D
12:55 AM rue_mohr: yea, I'm late for bed, I better eat supper
12:55 AM rue_mohr: this is epic
12:55 AM rue_mohr: I found my counter.
02:06 AM polprog: https://www.reddit.com/r/Skookum/comments/hmvhu5/why_i_had_to_walk_naked_across_a_factory_floor/
04:00 AM polprog: that's a great story there ^
09:39 AM rue_mohr: but...
09:39 AM rue_mohr: it sounds like its 3 phase
09:39 AM rue_mohr: the electricians hooking it up should have checked the rotation before using the machine...
09:48 AM rue_mohr: but there are SO many questions on that one...
09:48 AM rue_mohr: like why was he asked to wire that when he'd not done anything like that before
09:48 AM rue_mohr: and what lead to them building that for them?
09:59 AM polprog: it was the kid who hooked it up
08:46 PM rue_mohr: yea...
09:09 PM rue_shop1: YES
09:09 PM rue_shop1: the code works for the avr
09:10 PM Tom_L: cool
09:28 PM rue_mohr: so far
09:28 PM rue_mohr: not doing sine yet
09:38 PM Tom_L: how are avr data sheets now that pic took over?
09:40 PM rue_mohr: dunno, still using the old ones
09:40 PM rue_mohr: eep
09:41 PM rue_mohr: this might work...
09:44 PM Tom_L: i gave you that file with them all didn't i?
09:45 PM rue_mohr: the first file got me there
09:45 PM Tom_L: all the docs...
09:45 PM rue_mohr: but the independent clocks I need for the cpu and the timer are the same pin
09:45 PM rue_mohr: but it still might work
09:45 PM rue_mohr: every 256 cycles I need to update the pwm registers
09:46 PM rue_mohr: so, atmega88
09:46 PM rue_mohr: I'm using T2, in asynchronous mode, with external clocking
09:46 PM rue_mohr: it takes that external clock from EXTAL1
09:47 PM rue_mohr: then, I need to set an external clock for the processor
09:47 PM rue_mohr: because it takes 2 instrctions to set the pwm value, I need the clock for the processor to be /128
09:47 PM rue_mohr: the external clock for the processor is the EXTAL1 pin
09:47 PM rue_mohr: but I can use a cpu clock divider
09:48 PM rue_mohr: if it doesn't screw up the timer
09:48 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/avr/tech-lib/
09:48 PM Tom_L: tech lib
09:48 PM rue_mohr: I'm ok
09:48 PM rue_mohr: nobody to date has done this
09:48 PM Tom_L: it's there if you ever need it
09:49 PM Tom_L: it's got tech notes and pdf for parts
09:49 PM rue_mohr: ok
09:49 PM rue_mohr: the datasheet said that the asynchronous clock wouldn't work if it was faster than 4x the processor speed
09:49 PM rue_mohr: so far I proved that wrong
09:50 PM rue_mohr: I just need to make sure the processor divider doesn't screw with the timer
09:51 PM Tom_L: same with programming them. the spi clock must be 1/4 the clock or less
09:52 PM rue_mohr: well this worked, so I'm ok
09:52 PM rue_mohr: ok, next test is the cpu clock divider
09:54 PM rue_shop1: looks ok!
09:55 PM rue_shop1: hahah
09:55 PM rue_shop1: with the scaler preset, it cant program the chip
09:57 PM rue_shop1: how do I slow down the programmer bit rate
09:59 PM Tom_L: i don't remember. i think it's a spi fuse setting but i may be wrong
09:59 PM rue_shop1: -B 10kHz
09:59 PM Tom_L: default chip is as slow as it can go
10:01 PM rue_shop1: ooo
10:02 PM rue_shop1: bugger
10:02 PM Tom_L: internal osc at 1 or 4Mhz depending on the chip
10:02 PM Tom_L: so 1/4 that
10:02 PM rue_shop1: the cpu prescaler is before the takeoff for the timer
10:03 PM rue_shop1: I cant do the cpu/128 and the timer /1
10:04 PM rue_shop1: timer 2 is asynchronous, but uses the same damn external clock source pin as the external clock for the cpu
10:05 PM rue_shop1: the diagrams for this aren't entirely accurate
10:09 PM rue_shop1: mega8 is the same thing
10:09 PM rue_shop1: but
10:09 PM rue_shop1: I think mega32
10:10 PM rue_shop1: aha
10:11 PM rue_shop1: they seperate it on the atmega32
10:11 PM rue_shop1: into clock pins for the timer and clock pins for the cpu
10:12 PM rue_shop1: the 328 combines them
10:13 PM Tom_L: same family as the 88
10:13 PM Tom_L: 8 is it's predecessor which may be a bit odd
10:13 PM rue_shop1: heh
10:14 PM rue_shop1: the tiny26 does not have the edge detector that causes the issue
10:14 PM rue_shop1: this is funny
10:15 PM rue_shop1: the code I have for sine wave now uses a tiny26
10:15 PM rue_shop1: but just outputs 8 bit values
10:15 PM rue_shop1: it seems clear I cant do this with a mega8 / 88
10:17 PM rue_shop1: hah, the tiny26 does not have a cpu prescaler tho
10:17 PM rue_shop1: which means I need to go back to looping an asynchronous timer back into the clock pin
10:18 PM Tom_L: take a peek at the tiny10
10:18 PM Tom_L: it's got a 16 bit timer and maybe an 8 i forget
10:18 PM Tom_L: maybe not enough pins for you
10:18 PM rue_shop1: not enough memeory
10:19 PM rue_shop1: I dont even know if t26 does
10:19 PM rue_shop1: 2k? maybe
10:19 PM Tom_L: internal osc on the 10
10:19 PM Tom_L: i forget how fast
10:20 PM rue_shop1: no, I need the cpu to operate at 15kHz
10:22 PM rue_shop1: ok, well it looks like I have to play dual clocks with the t26, but it looks like it can do it
10:22 PM rue_shop1: wild ride with the m88
10:26 PM Tom_L: dead end
10:28 PM MoonyMoon: 989 2656
10:28 PM rue_mohr: I think the t26 can do it
10:28 PM rue_mohr: oh
10:28 PM rue_mohr: is that land line or cell?
11:22 PM rue_mohr: so...
11:22 PM rue_mohr: I'v not done pwm with the tiny26....
11:22 PM rue_mohr: heh
11:22 PM rue_mohr: oh wait...
11:23 PM rue_mohr: I kinda did
11:23 PM rue_mohr: for the mowerbot
11:24 PM rue_mohr: dual 4 bit parallel in, 2 pwm channels out