PI PICO W minimum voltage..

The WiFi needs more....

Yes. but allegedly not *that* much...On test it was the ultrasonic sensor that gave out before the wifi. At around 3.8V

I don't know the Pico's power specs, but maybe there's resistance on one of the electrical connections, such that the voltage drops more under load than where and when you measured it?

I measure at the far end of the chain

It has since failed again after only a few hours.

Back to the drawing board

yes

or on batteries itself?

yes, later

nope..

Hello

TP> This is a bit of a puzzle. My oil sensor stopped transmitting last TP> night. I replaced the batteries just now and its working again. The TP> voltage from is 3 x AA supply had dropped to 4.2V. This is supposed to TP> be enough to work a PICO W off? And IIRC on test it worked down to TP> around 3.8V or so.

How did you check voltage? On working device, or on batteries itself? One of batteries may be depleted more than others and whole pack may give less current than usual.

Best regards, Nikita.

How? Multimeter, USB power measuring dongle, oscilloscope.

The first two will average the voltage readings over a short period. You really need a scope to spot transient drops, such as when the Pico is transmitting over WiFi.

---druck

Sigh

The Pi PICO W has three ADCs on board and one is connected directly to VSYS by a voltage divider. I transmit that data every time the device successfully activates.

This is the relevant code frangment while (TRUE) { temperature = read_temperature(); printf("getting distance\n"); distance=get_distance(); voltage=get_battery_voltage(); // check wifi is up and connect/reconnect as necessary..if cant connect in 10 goes, give up. for(i=0;i<10 ;i++) { printf("Attempting to connect to wifi %s\n",ssid); cyw43_arch_wifi_connect_timeout_ms(ssid, password, CYW43_AUTH_WPA2_AES_PSK, 20000); sleep_ms(1000); // allow DHCP handshake to take place if ((cyw43_tcpip_link_status(&cyw43_state,CYW43_ITF_STA)) == CYW43_LINK_UP) break; } if(i==10) { suicide(); sleep_ms(1000); }

and so on.

The battery voltage is read after wifi initialisation but before a connection is established. The log file on the wifi hotpoint reveals no successful connections at all in the last 30 hours

Not even one that aborts later from low/bad signal.

This is the last transmission

OIL-SENSOR OIL-TANK

-79dBm

Pico should never gulp more than 200mA and that is well within a dry cell . There is a 220uF cap as well to handle spikes.

From raspberry Pi forum

"Re: Pico W power consumption

"Sun Aug 07, 2022 1:32 pm Using a battery pack with 3 AA batteries, reading 4.6V, I measured 52 mA using wifi and the LED. While starting up it did spike up to ~150 mA briefly. "

Anyway it's all irrelevant as the new batteries stopped working almost immediately, Last reading was 4.5V

I suspect it us the Triffid like growth of the Russian vine over the oil tank plus a stiff breeze making the leaves flutter.

I will remove the batteries and check them with a multimeter. I reloaded with the cheapest possible amazon 'value' cells and it is entirely possible that they are in fact utter s*it

I will actually order energizer lithium AAs from Amazon and try those. They have a far higher OCV (up to 1.75v) and netter sustained discharge voltages as well

When I tested the unit indoors for a month with an admittedly better wifi connection it operated down to under 4V measured by the ADC. Limiting factor was the ultrasonic sensor. Not the wifi.

Ah well, These things are sent to try us.

Let me see what is going on with the unit now...

Further investigation shows that a brand new set of AAs has discharged more overnight than the last set had in six months...(4,3V)

That is not possible if the nano timer is operating correctly: I suspect there is a constant battery drain from the unit somewhere.

Fortunately it's a socketed unit and I have spares...

I need to 3D print a cover for the oil tank before I can remove it however..it looks like rain

I'm not sure I trust something to measure it's own voltage, when it might be experiencing low voltage.

---druck

That is probably because you have not understood the first thing about how a PI PICO W is constructed.

It has its own onboard buck/boost regulators and the ADC is driven from a stabilised 3V line

The rest of the chip runs on a less well regulated 3.3V.

The chip runs down to 1.8V

The ADC is not set up to 'measure its own voltage' but that of the incoming voltage TO the chip.

FWIW my rpi4 is able to give low voltage warnings, seemingly reliably.

Yes. I have the thing back on the bench and have tested with an ammeter and a scope as well as voltmeter. After several startups it blew a 200mA fuse in the ammeter . So short term peaks are a little over that. Well a good excuse to buy a much better digital meter then. I was already sick of the previous one, Ali express are giving the previous type away for free when part of another order...so that is what THEY were worth...

I did catch it in two fault modes, one was an oscillating low energy switch, which results in the PICO never being switched on, and the other was some kind of network issue that resulted in it never switching off.

Either might cause loss of connection and high battery drain.

Both have not been possible to reproduce reliably.

I've modified the code to hopefully eliminate the 'I am stuck trying to connect to a TCP socket' condition, which should not happen anyway.

The oscillation might be caused by a bad electrolytic, and cold

Its not reoccurred since testing on the bench