LoPy with 3.3V direct powering fails to start.

[jmarcelino]

Well I guess I've been lucky.

As @robert-hh says it is inherently safer to power it from the 5V/Vin pin with 3.3v, though there is a small voltage drop and presumable higher current consumption.

In the meantime I'll reach out to TI themselves and hopefully get a more conclusive answer.

[jmarcelino]

Official answer from TI's Simple Switcher group: It is a bad idea to power from the 3.3V pin (if directly on the regulator output as it seems to be)

"Yes it is a bad idea to power any regulator from it's output. There is a parasitic diode from input to the switch pin that will be turned on in this configuration and can cause damage.
The easiest way to power from two supplies would be to use a simple diode-OR connection.
However, this will cause a voltage drop in your supply rail due too the diode drop. You might also look at some of TI's "ORing" controllers such as the LM5050 or similar."

So always use the Vin/5V pin to power the board (even with 3.3V)

[robert-hh]

@jmarcelino Thanks for asking. @daniel, do you listen?

[cocoa]

@jmarcelino @ssmith @robert-hh

Thanks to everyone of you , now is more clear what's happened.

What indeed isn't so clear is why Pycom's guy @Fred stated , 5 month ago, to use pin 3V3 to power LoPy and mantain a deep sleep current around 2.5 mA.

See https://forum.pycom.io/topic/1022/root-causes-of-high-deep-sleep-current Root Cause 1 : Workaround.

Now there is this situation :
if I want to have a low current in deep sleep state I've to supply LoPy trough pin 3V3, but using this pin, as shown today, is not a good practice and could fry LoPy module.

Any idea ?

[ssmith]

@cocoa Not a great answer but if you have a hot air gun you can remove the power supply chip from the LoPy. How much power are you wasting by putting 3.3V into the 5V input? Seems like that shouldn't add much.

[jmarcelino]

@ssmith
The LM3281 data sheet suggests this will be 16 μA.

"7.3.3 Low IQ
An ECOnomy (ECO) mode of operation draws 16 μA (typ.) quiescent current, permitting the LM3281 to be used in “always-on” applications. This low IQ is achieved over the entire input supply range of 5.5 V to 2.6 V, irrespective of whether LM3281 is operating in regulation (ECO Mode or Analog Bypass mode) or in full dropout (full bypass)."

Thus substantially less - insignificant really - compared to the minimum current boards operate at (given the issues around deepsleep)

[jmarcelino]

@robert-hh
I e-mailed the Pycom team this morning with the findings and they've already changed the usage notes to reflect this new information

https://docs.pycom.io/chapter/datasheets/notes/

When powering via VIN:
The input voltage must be between 3.4V and 5.5V.
Please DO NOT power the board via the 3.3V pin as this may damage the device. ONLY use the VIN pin for powering Pycom Devices.

It's official. Thanks everyone for helping out.

[robert-hh]

@jmarcelino To reduce confusion, they should change the range for Vin to 3.3V to 5V. I just verified that the LoPy indeed starts with Vin even as low as 3V. Maybe less. I Did not go below that.

[jmarcelino]

@robert-hh
Yes I noticed that when I pasted the text here, already let them know :-) Thanks.

[cocoa]

@jmarcelino @robert-hh @ssmith

Thanks for help to solve this issue.

[jcaron]

@jmarcelino @robert-hh

Isn't this the opposite of what what stated here:
https://forum.pycom.io/topic/1022/root-causes-of-high-deep-sleep-current

Workaround:
Power the board via the 3V3 pin using a stable 3V3 supply that is able to deliver at least 400mA. This lowers the deep-sleep current to around 2.5mA.

?

Also, it's unclear to me what the setup on the Pysense/Pytrack actually is? Is the Pysense/Pytrack sending power to the module via 5V, and then receiving 3V3 back from it (which I suppose only goes to the 3V3_Pymodule pin)? Someone I had thought the module was powered by the Pysense/Pytrack via the 3V3 pin, but that's probably incorrect.

[jmarcelino]

@jcaron

Isn't this the opposite of what what stated here:

Not completely the opposite, it's still more efficient to use a regulated 3.3V supply (low voltage forces the onboard regulator into bypass ECO mode)

But it needs to be updated to reflect that 3.3V should be applied to the 5V/VIN pin.

Is the Pysense/Pytrack sending power to the module via 5V

Yes, all the various shields implementing deep sleep control the *py board's 5V rail