ChanServ changed the topic of ##yamahasynths to: Channel dedicated to questions and discussion of Yamaha FM Synthesizer internals and corresponding REing. Discussion of synthesis methods similar to the Yamaha line of chips, Sound Blasters + clones, PCM chips like RF5C68, and CD theory of operation are also on-topic. Channel logs: https://freenode.irclog.whitequark.org/~h~yamahasynths
<cr1901_modern>
ValleyBell: Indeed, I got it down to like 680 lines after about 20 mins of pruning
<Lord_Nightmare>
i'm about 80% convinced cell E is a full adder with sum out, carry out (which is pulsed?), and A in, B in and Carry in
emily has quit [Ping timeout: 248 seconds]
emily has joined ##yamahasynths
Xyz39808 has joined ##yamahasynths
Xyz_39808 has joined ##yamahasynths
Xyz_39809 has quit [Ping timeout: 248 seconds]
Xyz_39809 has joined ##yamahasynths
Xyz39808 has quit [Ping timeout: 260 seconds]
Xyz39808 has joined ##yamahasynths
Xyz_39808 has quit [Ping timeout: 260 seconds]
Xyz_39809 has quit [Ping timeout: 260 seconds]
ZirconiumX has quit [Ping timeout: 260 seconds]
ZirconiumX has joined ##yamahasynths
<cr1901_modern>
ValleyBell: This will do for now. I'll call this "Version 0.1.0". It works, and it's completely usable for my purposes.
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
<fseidel>
It's always cool too see composers comment. Jeroen Tel comments on a bunch of covers of his stuff, and Toshiaki Sakoda used to comment on Musha covers until he deleted his old account and they all disappeared
<cr1901_modern>
That's the good wholesome shit I want to see
<cr1901_modern>
I got Takeshi Abo to idle here for about an hour last year, but I was the only one awake :(
<cr1901_modern>
Oh... he's the composer of Steins;Gate ._.
<cr1901_modern>
I... never knew that. I recognized him from his FM work lol
nukeykt has joined ##yamahasynths
<cr1901_modern>
Something changed in my setup, and now Inkscape doesn't scroll as smoothly as it should when I pan the die... hrm. I wonder if it has to do w/ integrated gfx or Windows 10...
<Sarayan>
It could be number of polygons
<Sarayan>
starting to hit the limits of cpu caches
<cr1901_modern>
I guess that's possible. Inkscape also required me to upgrade the svg from 90 to 96DPI
<cr1901_modern>
Sarayan: Btw, I have on my todoist: "Fix dietools atomic move breakage". Due Jan 31st.
<cr1901_modern>
Let's see how much I overshoot that deadline
<Lord_Nightmare>
E: 5-t cmos full adder with carry in and out
<Lord_Nightmare>
as suspected
<Lord_Nightmare>
i'm wondering if B is just a minor variant on E
<Lord_Nightmare>
or even the same gate, and i misidentified them as different
<Lord_Nightmare>
answer found:
<Lord_Nightmare>
B: 5-t cmos full adder with carry in and out, with high drive power for the SUM output
<Lord_Nightmare>
that was easy
<Lord_Nightmare>
the only gate cell left is A now, of which there are 12 of them forming what are presumably the bits of the integrator filter accumulator/dac driver
<Lord_Nightmare>
the datasheet only shows 10 bits to the dac, so i'm not entirely sure what's going on. its possible the low 2 dac bits aren't hooked up, or possibly the HIGH 2 dac bits aren't hooked up and allow the dac/integrator to 'saturate' while still retaining state, since the datasheet shows exactly this happening
<Sarayan>
what are you talking about?
<Lord_Nightmare>
hc55516
<Lord_Nightmare>
i'm working on getting a netlist/bit accurate implementation done because unlike the motorola cvsd chips, the harris hc555xx chips are almost completely digital, the only part not is the dac and the voltage comparator/input
<Sarayan>
from a die shot?
<Lord_Nightmare>
yep
<Sarayan>
nice
<Sarayan>
pr0n?
<Lord_Nightmare>
no, on sean's site. it probably SHOULD be on pr0n, but it isn't, yet.
<Lord_Nightmare>
E: 5-t cmos full adder with carry in and out
<Lord_Nightmare>
F: 4-t cmos 2:1 MUX
<Lord_Nightmare>
G: 5-t cmos 4-input NAND
<Lord_Nightmare>
H: 3-t cmos 2-input XOR
<Lord_Nightmare>
I: 2-t cmos inverter
<Lord_Nightmare>
J: 3-t cmos 2-input NAND
<Lord_Nightmare>
K: 2-t cmos inverter with high drive power
<Lord_Nightmare>
L: 3-t cmos 2-input NOR
<Lord_Nightmare>
M: 5-t (really 4-t) GUESSED: cmos d-latch with ? edge triggering, open collector inverted and non-inverted outputs
<Lord_Nightmare>
N: 5-t cmos 4-input OR
<Lord_Nightmare>
M is weird, and there may be a stitching error with that one gate, I need to ask sean for the unstitched source images
<Lord_Nightmare>
it might not be open collector, as i found out from the weird way some of the metal is hooked to vcc through some sort of buried contact layer
<Lord_Nightmare>
on the B and E cells
<Sarayan>
open collector is damn rare in cmos
<Sarayan>
it is rather rare in nmos too, but I've seen it happen
<Sarayan>
the wd1772 has a beautiful one
<Lord_Nightmare>
there's definitely some strange design decisions in the M gate, there's a bunch of poly traces which don't actually connect to a gate and just 'end'
<Lord_Nightmare>
the M cell, i mean
<Lord_Nightmare>
it looks like someone may have manually hacked it up a bit very late in design
<Lord_Nightmare>
there's only one M cell, and it connects to the digital output pin and its input (and /Q output) go to a MUX, allowing the mux to choose whether the digital output gets a signal, or if it inverts every other clock (the 'force zero' and 'alternate plain text' "quieting pattern" described on the datasheet)
<Sarayan>
are you sure about the poly traces?
cr1901_modern has quit [Quit: Leaving.]
<Sarayan>
I mean there seem to be horizontal poly under the big metal vcc/gnd
<Lord_Nightmare>
no. a delayer would definitely help clarify things
<Sarayan>
connecting a bunch of gates together
<Lord_Nightmare>
spotting the horizontal poly under the vcc and gnd traces is a lesson in eyeballing it
<Lord_Nightmare>
and comparing between multiple gate cells
<Lord_Nightmare>
since its very very hard to see
<Lord_Nightmare>
that's why i'm really not sure about the M cell, since there's only one
<Sarayan>
yeah, that doesn't help
<Lord_Nightmare>
the N cell also only has one cell, but that one its fairly obvious what it does
cr1901_modern has joined ##yamahasynths
<Ultrasauce>
theres something wonderfully perverse about 22 megabytes of base64-encoded jpeg sitting in an xml file
<Sarayan>
the original jpg is 16M, I think it's the original you have in the svg :-)
<Lord_Nightmare>
ah. it was the .xcf which was much larger, i guess gimp doesn't even bother compressing an imported image
<Lord_Nightmare>
i gave up on the .xcf since it took too much memory and was annoying to trace
<Lord_Nightmare>
one thing svg is missing which xcf has is layer groups/heirarchy. i'm wondering if that limitation is part of the .svg format or just something the version of inkscape i have hasn't implemented
<Lord_Nightmare>
i'm running 0.92.0, latest is 0.92.4
<Lord_Nightmare>
oh neat, 0.92.2 fixed the bug with the color selector freaking out and showing a glitched 5th slider which does nothing
<Lord_Nightmare>
i should probably upgrade
futarisIRCcloud has joined ##yamahasynths
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
nukeykt has quit [Remote host closed the connection]