Search

Few weeks late, sorry. Anyone doing this on a newer part should look at the C3, deep sleep is about 5uA and the same board rules apply. Devkit numbers are meaningless, as above.

Pro Micro, 12 keys, QMK. - matrix is 3x4, diodes all the right way round after the second attempt - qmk compile with my own keyboard folder builds, layout works. Key 7 doubles sometimes, might be a bad joint or could need more debounce - wa…

About 90ms awake total. Wifi and mqtt was like 2 seconds. Massive difference!!

Good numbers. Told you to come back with them! What do you see for wake time with ESP-NOW?

Kept going after /progrider/t/0ve0JwhGsLJ0 - bare C3 module now, 6uA asleep, I measured properly this time! - 18650, 1W panel, CN3791 charge board - wakes every 10 min, sends over ESP-NOW to a base station. Way less airtime than wifi+mqtt -…

Mine did this until a BIOS update. Check the Lenovo release notes for your model, there was a thermal fix for several of them. Not saying it's your problem, but it's a cheap thing to try first.

Had this on a T470. Reloading the thinkpad_acpi module after resume fixed it for me, put it in a systemd sleep hook. Never found the root cause.

T480, Debian. After resuming from suspend the fan runs flat out until reboot. Sensors show normal temperatures. Anyone seen this? Happens maybe one resume in five.

Yep, RMT's good too. PIO just feels like cheating.

ESP32 has RMT for the same thing!! Works great. But yes never bit bang, ever

Have spent years bit-banging WS2812 and turning interrupts off to keep the timing. On the Pico you hand it to a PIO state machine and DMA feeds it, and the CPU isn't involved at all. There's a ws2812 example in pico-examples that's about 10…

Different cable, plugged straight into the laptop. Been stable for two days. Resolving.

minicom if you want menus, screen /dev/ttyUSB0 115200 if you don't. picocom is what I'd pick too.

Try without the USB hub if you have one. Mine was a power issue; changing terminal settings didn't help.

The adapter connects fine at 115200 but eventually stops receiving. Unplugging and plugging it back in fixes it for a while. dmesg says "usb disconnect" when it happens. Same result with two different terminal programs. Going to try another…

picocom is fine. For logging, --logfile session.txt exists in newer versions. Older ones don't have it, which is probably why people wrap it in script.

Have a spare ThinkPad with a cracked corner and a working screen. Thinking of leaving it on the workbench for serial consoles. Any reason to use something other than picocom? Mostly routers and an Arduino. Would like to save the session to …

30cm is fine at 100kHz. If it gets flaky drop the pull-ups to 2k2 before you change the cable.

Box is a Pi Zero in a plastic junction box, on the shelf above the door. - BME280 on i2c, address 0x76, cable is 30cm which is about the limit before it gets flaky - writes to local sqlite, porch box pulls over ssh every hour - condensation…

Nice. Post the numbers after a month on the battery, I'd like to see how close it gets to the sum.

GOT IT. 14uA with the 1M divider and DHT on a GPIO!!! kris you are a legend. Marking this solved. Burden voltage was also real, it kept resetting until I shorted the meter during wake.

Also how are you measuring? A multimeter on the uA range has a big burden voltage. Can brown out at wake-up when it pulls 200mA through that range. If it reboots in a loop on the meter and not on the bench supply, that's why.

There you go. A 20k divider on a 4V cell is 200uA all by itself. Go to 1M + 1M and put 100nF across the bottom resistor so the ADC has something to sample. And power the DHT from a GPIO, or a P-channel switch, so it's off during sleep.

DHT22 and a divider, 10k + 10k for battery voltage. Oh no.

What else is on the rail? Sensor? Voltage divider for the battery?