-->
Page 7 of 34

Re: node.dsleep() reliability - zombie mode

PostPosted: Tue Feb 03, 2015 2:37 pm
by GeoNomad
gwizz wrote:Been asked via EEVBlog forum - what voltage is the GPIO0 oscillating to during zombie boot mode?

Thanks!



OK. It ran 2 hours and 25 minutes on a one minute cycle using node.dsleep() after sending the temperature.

Then it went into zombie mode.

GPIO is now oscillating from 0.8 to 3.15 volts at 25 MHz. Looks more like sine wave than square, but it's a 25 year old old scope, so not trustworthy at that frequency.

Current draw is about 20 mA (including LED).

Only connections are power and CH_PD.

ESP-01 board.

Hope that helps.

Look forward to hearing some feedback.

Peter

Re: node.dsleep() reliability - zombie mode

PostPosted: Tue Feb 03, 2015 2:46 pm
by gwizz
I'm trying to reproduce the problem myself, not happened yet!

To help keep us all on the same page, can we use a pre-built firmware? I've suggested that the problem occurs with nodemcu_512k_20141222.bin and nodemcu_20150126.bin - is it reproduced in the latest one also?

EEVblog forum member LabSpokane has suggested it would help to capture the scope trace and a photo of the setup? Can anyone with a scope oblige?

Thanks!!

Re: node.dsleep() reliability - zombie mode

PostPosted: Tue Feb 03, 2015 2:58 pm
by manawyrm
@gwizz: Can confirm.
I've linked you 2 osci. photos down there. The frequency is exactly 26 MHz. (the crystals freq.).

@GeoNomad: Yep, you're right.

Please be aware: These measurements have been done without any termination. (the real voltage is probably lower)

Re: node.dsleep() reliability - zombie mode

PostPosted: Tue Feb 03, 2015 3:00 pm
by GeoNomad
gwizz wrote: it would help to capture the scope trace and a photo of the setup?


Nothing worth photographing on the scope. Just a simple oscillation.

Image