Graham Sutherland / Polynomial

he\him

heavily ADHD.

into electronics, windows internals, cryptography, security, compute hardware, physics, colourimetry, lasers, stage lighting, D&B, DJing, demoscene, lepidoptera, socialism.

I am mothman. LÄMP 💡

nullsector/laser team @ EMF Camp

lasers & lighting orga @ NOVA Demoparty

I sell funny warning stickers at Unsafe Warnings: https://unsafewarnings.etsy.com

all posts encrypted with ROT256-ECB.

header photo by James (npub17qh…xe2r)

Web: https://poly.nomial.co.uk
Blog: https://blog.poly.nomial.co.uk/
Warning Stickers ⚠: https://unsafewarnings.etsy.com
Streaming: https://twitch.tv/gsuberland

Public Key

npub1szz2a5n67z8nfx2s5z87xqk5znexu3kfdmprq5cqva8g0gdd8snq07t4fx

NIP-05 Address

[email protected]

Profile Code

nprofile1qqsgpp9w6fa0pre5n9g2prlrqt2pfunwgmykas3s2vqxwn585xkncfspzpmhxue69uhkummnw3ezuamfdejsz9rhwden5te0wfjkccte9ejxzmt4wvhxjmcpp4mhxue69uhkummn9ekx7mq6r5s9z

Publishing to

nostr.wine

relay.damus.io

nos.lol

Author Public Key

npub1szz2a5n67z8nfx2s5z87xqk5znexu3kfdmprq5cqva8g0gdd8snq07t4fx

Show more details

Last Notes

2026-04-24 12:37:10 CEST

attempting to offset misery by buying comforting objects. it's not really working but at least we have ridiculously soft towels now.

2026-04-16 04:04:28 CEST

- reply

I already bought the 16i16 anyway, but yeah, more info can't hurt.
I'm actually not that bothered about the specs, 16-bit 48kHz is plenty. the main thing I care about is no hiss with no audio out and no standby nonsense.

2026-04-13 20:21:07 CEST

OH: "the pope takes double damage from nuclear weapons"

2026-04-12 07:58:12 CEST

- reply

last I checked it doesn't support casting to any of my TVs.

2026-04-12 07:24:38 CEST

plex seems to have adopted the "ask again later" model when it comes to consent popups. I must've said I do not agree a dozen times by now.

2026-04-06 17:49:17 CEST

do you know what "instagib" means, and what the "gib" part refers to, without looking it up?

2026-04-02 01:51:28 CEST

- reply

sounds like the space piss drama has been averted

2026-04-02 00:38:34 CEST

- reply

throttle down to 85%, SRBs separated

2025-12-26 17:44:13 CET

dropping a fully functioning heartbleed-style exploit against mongodb on xmas day is a fucknugget level dick move for defenders on call. I hope a horse kicks the guy in the crotch.

2025-12-16 16:48:50 CET

the temptation to put a DECT number in my display name despite not being at Congress is real

2025-12-12 21:22:55 CET

I hate default line length limits in linters. it's 2026, I'm not coding on a fucking VT100.

2025-12-12 00:57:20 CET

- reply

carrot-meters have now been added to the resistance units list for @npub1m8q…wweh

2025-12-12 00:57:03 CET

just remembered the carrot resistance measurement saga
https://chaos.social/@gsuberland/115152538274090027
#note1nas…ewpn

2025-12-11 09:37:20 CET

- reply

knew this before styro but yes that's what reminded me

2025-12-10 23:23:03 CET

did a prion write this ad
https://assets.chaos.social/media_attachments/files/115/697/592/368/911/580/original/d99af033178db9a4.png

2025-12-05 02:23:50 CET

- reply

I have never designed a fetch execute cycle in my life, so best I can say is that from an electrical perspective it seems sound.

2025-12-05 02:14:59 CET

- reply

same tbh, I had used the flip-flop loop trick before but had never tried making a no-overlap version. I guessed that xor'ing sequential outputs would work and it did haha, sometimes you just get a free win

2025-12-05 01:58:13 CET

- reply

oh no problem at all. at 5V Nexperia's 74HC74 can run at 20MHz at the bare minimum. so feed that circuit 8MHz and you'll get 8 phases at 1MHz.

2025-12-05 01:48:17 CET

- reply

ah, right, makes sense.
this would work, but you would need the source clock to be 8x and the D flip-flops would need to be fast enough to support that high speed clock.
https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0mRBmaBWAnKhCwA5sOQGwBMeqBIykFIALAhQKYC0YYAUGMsiE0YT6yICA7ORZghVSBy4CJcoUwSLBUKGwBKPIkRrgS2-hOzgQQsFSIoz1qShndlC8PPGSKD8KOd8xqqshsAO4CBFRuzhHSISx+8k5RwfomEWDeiTG+ImJZiUSQwtq6yQKo5gZ6AMIA0mz5hbz8TNjhJCkttCA1dQWRHSw0es1UVbX1IAkd4R2jPYVpYh1ggzwzXbUhbTw0JkTCijvqMWBh2yknVCdHoWc3V9ElTIfG4OQPW0+7uK-q47GXflOhz0AEUAOTsP5XK6xWgmUFgohzErQirwkDghDIrZXLbAjFgmjI-6RC4CFagyG9WERGnLPSgpHjF6pAwDBkYrHjPYqeQ88kckFEzh6CyQc4rMUpQ6WaiBcboFQrRXZTogqnzSDFdngLVDBZqxHIwThekufKqhFE5niiWi22WzkyUUWfWS123WXcQIASV14tuUp+dm9bD9rEgQ2eHvuUGsvv0cED+QBahgoZFidNkpTnrMcuNODdLuw+u8oPlvQkMLN1b85YJBGNWCVopbjvB2GNJGzop7HeEzv0ZAFw7EMvzoZtAehcHh6IACgAJDUuSM-CMMxdLplV3OopP45dcvcj1Ejo9L629Gg1+S3vwUjGVwp0czkN9ww2BEKfz60Jx7iSP9gRhGVkWUaZLG+YZOm6TZ9mcSDnGkThHAtCIaCTKIPHGQhTXkfDR1BJs8O8K5kHLdFwSbEJKPHEx6ODJImP-Ij-2kcYaAIIYZkmEZ1mAu8hG4-V5FQ2RROcZAVhwl8QGEdcIkU4s1UHcZ9lTTSvwRQcQm0-9tKA-SMMOFTPWRMJ2gCdI1ngihTgiKyUM8ZynJg-xcN6UtznkHziIxLtxmwHifhCrdDS7EJwtuGLjJAGL-38jjkWEZBFA6Ay7I2BSlNUNLeUkTwCucPBCrUBMJBHf8qvIWN0w8aLQqS5qDQeRKDUaes3jYAB7WhzDULDUEYuMIFlGghCxfq6FMEZIBG6gYHG-NZum2hRSGhbRuWtQRL0daaG4Fbhp26AVpExw2CAA

2025-12-05 01:21:11 CET

- reply

putting a buffer or inverter in where you did will certainly make the behaviour better, at least.

2025-12-05 01:15:43 CET

- reply

yup. the simulator can resolve the state of the clock signal being passed over to the AND inputs instantly without the NOT gate in place, whereas the state of the two flip-flops requires an additional timestep to compute, meaning the clock edge arrives exactly one timestep earlier than the updated signals from the flip-flops, resulting in the spike.
you can also place a Schmitt trigger (from Active Building Blocks) in place of the NOT gate and the blip will go away.

2025-11-26 23:58:59 CET

- reply

interesting choice to only have xnor. presumably that simplified the ALU.

2025-11-12 02:33:27 CET

- reply

ah, ok, so opamps can only handle input and output voltages up to their supply rails (or some drop away from the rails if it's not a rail-to-rail opamp). so if you have some input signal that goes both positive and negative (e.g. a sine wave centered around 0V) then you need to supply the opamps with both a positive and negative supply rail, otherwise it'll clip.
e.g. if you supply a rail-to-rail input & output (RRIO) opamp with 0-5V, it can only do 0-5V in and 0-5V out.

2025-09-25 01:12:15 CEST

insulin mutual aid scam posts are kicking up again. similar pretext, this time usually saying they're a family from Gaza. easily identified by the fact that they're brand new accounts and send dozens of identical messages to complete strangers, and often use the same pretext across multiple accounts.
as usual the appropriate response is to report them, let admins use those reports to set new spam filters, and donate/boost twice as hard for the real people who need mutual aid.

2025-07-09 17:15:08 CEST

as someone who rather vocally does not like Linux at all and also does a lot of Windows stuff, there seems to be an expectation that I have some more-positive opinion on Windows 11 but like... no. absolutely not. if anything I'm *MORE* pissed off than the rest of you are. y'all didn't really seem to like the platform in the first place, I fuckin' loved it.
(please do not proselytize for alternatives in the replies, I am in no mood)

2025-03-13 05:01:14 CET

rate my buck converter
https://assets.chaos.social/media_attachments/files/114/153/101/520/184/433/original/19b7e2f4b678a016.png
https://assets.chaos.social/media_attachments/files/114/153/115/682/984/542/original/e2566d462cc78de7.png

2025-03-07 12:32:02 CET

- reply

looking at the 008004 caps again.
they're so small I can't even get a useful photo of them with my phone camera's dedicated macro lens.
I can't believe they squeezed 100nF into these, let alone with a normal DC bias falloff curve. the fact that they then went and did it with 006003 afterwards is just bonkers.

2025-03-07 12:13:52 CET

- reply

also some cool stuff I forgot I had, like a bunch of LM431-style 100mA 40V precision zener references designed for driving high power LEDs. with proper mounting they can handle about 2W of dissipation, which is impressive.

2025-03-07 12:11:08 CET

- reply

also going through my stash of parts and there's some weird stuff in here. why do I own so many video decoder/encoder/transcoder ICs? I vaguely recall thinking about an analogue video mashing / glitching project but I don't know why I bought so many different video ICs for it. there must be a dozen different P/Ns here.

2025-03-07 11:02:02 CET

hackspace donation inventory is happening
https://assets.chaos.social/media_attachments/files/114/120/567/693/174/561/original/06e09b1d5888a235.jpg

2025-03-04 12:47:56 CET

- reply

yeah I feel like people get the wrong idea a lot of the time when I say "I want a tool not a toy" but it is in exactly this sense
so many things are toys these days

2025-01-29 20:01:57 CET

- reply

twitch VODs mute 'cos copyright, but I locally record the stream in OBS *and* record the set via Rekordbox (better dynamics on the audio) so I've got double evidence :D

2025-01-29 19:58:55 CET

listening back to my mix from last night and I'm pleasantly surprised that it was not a mess. I was, uh, somewhat inebriated.

2025-01-11 17:46:03 CET

- reply

ngl I felt gross writing that one

2025-01-11 17:45:34 CET

- reply

subcritical and breedable

2024-12-27 08:10:20 CET

what is the weirdest commercial off-the-shelf IC that you're aware of? critically: must still be in production, or at least produced in the last few years, and still available for purchase from distributors.
#electronics

2024-12-21 13:00:32 CET

- reply

(at some point I will eventually do it anyway in spite of the voice drivel and then promptly forget literally all of it as I go build weird packet networks)

2024-11-23 05:28:05 CET

- reply

wouldn't need Linux at all. pretty simple firmware on a microprocessor.

2024-06-28 07:52:04 CEST

- reply

also good because for audio frequency you need way more capacitance for DC blocking than a C0G/NP0 MLCC can provide, and class 2 (X5R, X7R, etc.) MLCCs have zero-crossing distortion due to polarisation hysteresis effects.

2024-05-19 23:21:42 CEST

- reply

that's 10 pins on the top, then all the bottom row is just GND.

2024-05-19 23:20:45 CEST

- reply

I'd probably not bother with 1.2V on there because at that low a voltage you really want to regulate it onboard to avoid droop from ohmic losses. it's easy to regulate onboard from 3.3V too.
I'd go with something like TX, RX, GND, 3.3V, 3.3V, 5.0V, 5.0V, 12V, PRSNT#, PG
the device pulls PRSNT# to GND to indicate that something is plugged in. that's useful if you end up putting load switches on the supply to help manage inrush current.
PG is useful too for DC-DCs that have a PG out.

2024-05-17 08:46:18 CEST

how has nobody leaked gerbers for the pi boards, you'd think that would've happened several times by now.
I have no interest in using them but it would be very interesting to review their board layouts.

2022-11-25 20:16:34 CET

if you've ever messed up a dimension or a hole position on something you're building, don't be too hard on yourself.
at least you're not the Cisco design engineer who caused an entire product line recall by placing the mode button (which resets the switch if held) directly above an RJ45 port.
https://assets.chaos.social/media_attachments/files/109/406/013/793/351/730/original/da3803bec6cbff70.jpg
https://assets.chaos.social/media_attachments/files/109/406/031/039/700/085/original/f69224ef7ec8a545.webp