tomfleet.bsky.social
Hardware Developer @inMusicBrands
Technical Writer @Hacksterio
Mainly making things that make music.
209 posts
1,254 followers
193 following
Regular Contributor
Active Commenter
comment in response to
post
Insanity!
Now and then?!
Real ones know you use them as kindling, and chuck at least two more on each and every fire.
comment in response to
post
I guess it's more "tap" and "swipe" than it is "click" these days?
comment in response to
post
Don't forget the I2C pull-ups .. unless you are set on the internal ones in the nRF!
comment in response to
post
Nice. Seems like given your uses for it, it might end up as a good board for use with the SeeSaw libraries?
learn.adafruit.com/adafruit-att...
comment in response to
post
But yeah. the actual measurement offset here is absolutely down to the fact that these two sensors believe they are measuring drastically different air temps, so the are calibrating against incorrect data internally.
comment in response to
post
5 degrees*
My point is that these parts have complex dynamics. They can even have error introduced via the measurement/power mode chosen.
Hammering away in high perf mode gives you chunks of fairly significant current that needs to be dissipated eventually Thermals are everything with these.
comment in response to
post
Ok then, the other board reads nearly 22 - some 6 degrees out.
The idea is that figure should show (and match) the *gas* temp.
My point is, the two sensor algorithms are being fed different inputs, that directly feed into the calculation of the equivalent CO2 output value.
comment in response to
post
Speaking of thermal coupling. I guess there is AC on, and a vent blowing at the EPD board, probably with a reasonable groundplane acting as heatsink? If the sensor thinks it's sensing a *gas* temp that has that temp difference, well.. the calculated equivalent CO2 output is gas temp dependent, yeah?
comment in response to
post
These sensors, like the MOX chemristive iAQ sensors are a popular design topic recently... Very very few people properly implement these parts correctly, or, in such a way that will give the figures quoted for accuracy, etc.
comment in response to
post
It needs calibration, it needs *very much* decoupling from the board thermal contributions - ie,the ESP, the LDO sourcing that back-light... also from memory, keep it away from sunlight (strong wideband IR - as it's IR photo-acoustic)
comment in response to
post
Well that is some perfect silk screen commentary, if I don't say so.
comment in response to
post
I thought you were going to just say "I can get behind that".
Just like Sisyphus got behind that rock.
comment in response to
post
They are the same thing though. The screens in Garmin are MIP.
Source; my fenix that I need to put back together!
www.sharpsde.com/products/dis...
comment in response to
post
Both? As in, to my understanding, these MIP fall into that category of panel?
But yeah. Sharp have a couple of RGB MIP options for this tech.
comment in response to
post
And they also offer a couple of smaller colour panels, which are on my to do / interest list...
comment in response to
post
Oh, and the code is a fork of the Adafruit fork of this project - that explains how the "video" data is generated and packed.
Adafruit forked and scaled for memory LCD, all I did was source the data from the ESP32 FLASH SPIFFS, instead of an external SPI Flash device.
github.com/hackffm/ESP3...
comment in response to
post
For those interested in the board, it's this one here, from @fasani.bsky.social!
ESP32-s3 based.
www.tindie.com/products/fas...
comment in response to
post
A memory in pixel / memory LCD from Sharp.
Very fun things indeed.
www.sharpsde.com/fileadmin/pr...
comment in response to
post
Hell forgot the alt text on the photos. They are just two photos of a very broken e-paper display, showing garbled image data that gets a bit less garbled with a gentle touch!
comment in response to
post
Aw shoot. So looks like this display is borked. It got shipped with the FPC folded around the padding foam, rather than flat. I can get a picture if I tap the die bonded drivers *just right* - but it's definitely pretty dodgy!
But - wasn't the code at least!
comment in response to
post
Hm. Certainly not a Banksy itself, but in the style of. The tag is a good indicator, along with the stencil bleed.
Lot of this stencil around Hamburg.
comment in response to
post
I think i'll have another look - along with this, which supposedly works, and seems to have the bisected display approach, which this thing needs.
github.com/wavesharetea...
comment in response to
post
Well that patterning is my result from playing with bb_fastepd, having defined FUTURE (well. Removing the #ifdef...), but I'm missing something, clearly...
Partial updates gives noise on 1/4 the panel, and panel test gives what your can see currently...
Need to spend more than 5 mins on tinkering!
comment in response to
post
It's 5.79"?
792 x 292px, this one is B/W, but there are b/w/r and b/w/r/y versions also.
comment in response to
post
haha, it's not broken :)
comment in response to
post
I... I don't have a screen fault - the donor is fully functional :)
What you might be seeing is the garbage printed to the e-paper display, that is still visible from playing with it the other night!
comment in response to
post
I thought that had come to be less reliable?
Besides. I can more easily remember how to swear :)
comment in response to
post
I know a guy who will sell you some RP2350 with a few nRF ground up and mixed in?
comment in response to
post
Massgrave.dev
comment in response to
post
Ask it about massgrave
comment in response to
post
And don't forget that the easiest brute force password recovery method is often upon the owner...
comment in response to
post
Had the same idea a little while back - it does seem to help a huge amount!
comment in response to
post
Did this really need AI images?
What's the sense in talking about a real product, only to dump that rubbish for an article image.
comment in response to
post
www.aliexpress.com/item/1005005...
I swear this is real.
comment in response to
post
"Exquisite details"
comment in response to
post
Spot on, SysEx.
comment in response to
post
Oh I actually do that.
Along with the ones that respond follow the "statement, mitigation, moral overview, and 'what are my thoughts on this'?"
There's a very recognizable"arc" to a gpt answer.
comment in response to
post
Worse that when it's wrong (often) - these people will now start to argue with your corrections with newfound arrogance...
comment in response to
post
We had a Tesla retail showroom in the city, usually flipped it the bird on the way past.
It's disappeared now, which is just lovely.