💬 ESP-LINK ESP8266 WeMos D1 Mini Adapter Board

[NeverDie]

@Nca78
You're ahead of me. How much is supported?

[kalle]

@NeverDie
Do you noticed also the PSF-A85 and PSF-A85 boards on the itead website?

https://www.itead.cc/wiki/PSF-A85
https://www.itead.cc/wiki/PSF-B85

These boards contain the new ESP8285. I have seen this ESP8285 (ESP8285 internal Flash is 1MByte = 8Mbit) in the Sonoff Wall-Touch-switch and it was easy to write a Arduino sketch to it with an FTDI programmer (FTDI prg. is only for the first time needed, after that you can update the sketch OTA).

The Board has only ~14mmx14mm - so it is possible to shrink "Wemos D1 mini" based projects.
Here are a picture to see the Wemos size vs. PSF-A85 (red square):

Kalle

[Nca78]

@NeverDie said:

@Nca78
You're ahead of me. How much is supported?

Sorry I didn't go much further than seeing there's some code to modify to cheat in the flash read/write methods to access it from the ic when running, and some other modifications in arduino IDE to be able to write to more than 4Mb. As I don't have a need for the extra storage at the moment I decided to pass :)

@kalle thank you for the link I didn't notice those before. Very interesting indeed as it's an ESP8266 with extra memory. With a small ceramic antenna similar to the one on the D1 mini pro it can make a really tiny
I just discover the wall switches too, they are nice but you need neutral wire (not surprising to be able to power esp8266) and it's not available for the switch type I have here US(AU) which is a bit smaller than standard US type :(

[kalle]

@Nca78 I have tested both types of switches (EU/US) and they have both the same components inside - so maybe the EU wall switch will fit because it is even smaller. The price for such a switch is amazing (15$), but it is not glass on the top it is acrylic plastic.

[NeverDie]

@kalle
Thanks for the links. Itead describes them as "ultra alow power," but I'm wondering: what's ultra low power about them? Aside from turning it off (500na), everything else doesn't seem ultra low power, at least compared to the usual atemega328p + nrf/rfm radio module that we're all used to on this forum.

That said, I think the esp8266/85 actually could be made much lower power if it were possible (and it probably is, but I don't know how) to use its radio to simply broadcast a message without first linking to a router. Then, even if the power drain is high while active, it's active period could be kept so short that the energy drain would be quite manageable. Do you happen to know if there's a way to do that?

[kalle]

@NeverDie Itead said this chip is also for creating wearaible devices - so I think it is possible to use it with batteries. The ESP has a deepsleep function but I have no experience with it and don't know how it works - sorry. I have only "hacked" the Sonoff touch wall switch to use it without the terrible cloud app - I don't use it with the mysensors stuff, but maybe it is possible to use such a switch as ESPgateway node (the chip is powered with 3.3V) and we can try to upload the ESPgateway sketch on the switch.

[NeverDie]

@kalle

The deep sleep is good enough. What ruins the whole thing, though, is the fairly long re-acquaintance with the wi-fi router once it wakes up. That's where most of the power drain happens. If we could just drive the radio directly, then that problem goes away.

[NeverDie]

@NeverDie

OK, I just noticed that it supports wi-fi direct (https://www.itead.cc/psf-a85.html), so perhaps that could accomplish it, since that doesn't require the lengthy wi-fi router dance.

The other strange thing is that Itead says its module supports 2x1 MIMO, and yet I see only one antenna port on it.

[kalle]

@NeverDie To connect to the router take only 1-2 seconds for my wemos devices (static IP), but ok - maybe this is to long.
I think for a switch, it can be run all the time - or do I misunderstand something

[NeverDie]

@kalle
Yeah, 1-2 seconds is correct, but there's a consensus that for battery operation, because of the high current drain, that's actually 2 to 3 orders of magnitude too long. Think of a typical temp-rh sensor that updates every 5 minutes, for example. It adds up quickly.

[Nca78]

@NeverDie said:

@NeverDie

OK, I just noticed that it supports wi-fi direct (https://www.itead.cc/psf-a85.html), so perhaps that could accomplish it, since that doesn't require the lengthy wi-fi router dance.

No it doesn't. That's a claim that has been made by Espressif in the ESP8266 specs. But it has never been implemented...

Wearable ESP8266 (or 8285) is also a bold claim by iTead. The lowest power mode for radio transmit draws around 120mA, it's already over 50mA in receive mode. And if you use PWM, i2C (and I fail to see the use of wearable without them) or anything like that even when disabling the radio it will consume 15mA.

I have better hopes for the ESP32 as it's supposed to have some kind of very low power (for real) coprocessor to handle small tasks while the main cores will sleep. But it's still in development on the SDK side, and as seen with the "wifi direct" claim on the ESP8266 we should not take every promised feature from Espressif at face value ;)

[kalle]

@Nca78 Thanks for the explanation - and yes, I'm also still waiting for the ESP32

[scalz]

Hmm.. at Mysensors Team we already have our ESP32 LORA board :stuck_out_tongue_winking_eye:

This is a LIPO charger, LORA board i did for ESP32 evaluation. With a nice LDO, I added a neopixel, and prefer bigger tactile switch :) same width as Nano32 but 1cm longer, so one row available on each side for breadboarding. ESP32 is a nice little beast :)

I have other stuff coming for esp32. i'll release ;)

[NeverDie]

@scalz
Will the ESP32 be capable of "ultra low power" in a real sense? Have you run any measurements?

[scalz]

imho, ultra low power is sub uA. Above, it's very low power :)

So, i'm not sure for ultra low power (depends also on strategy..), but very low power yes. I've not run any power consumption tests, focused on other parts. And the SDK is still not complete including the low power aspects. Some stuff is missing in ESP-IDF, but ready in Arduino, and vice versa. Docs are not complete yet, but looks good.
That said, ESP32 team is doing a great work! And Espressif is working on a revision of their sillicon die (there is an errata)...

[NeverDie]

@scalz
In that case, what's your definition of "very low power"?

[scalz]

lol, my very personal view, for battery operated devices:

• ultra low power : <= 1uA
• very low power : < 10uA
• low power : 100uA, not sure for this one :)

The same for the efficiency ;)

[NeverDie]

Will the ESP32 be able to remember its state (including all variable values) just prior to sleeping, like an atmega328p can, or like the ESP8266 do you have to explicitly store the state in flash before sleeping and then restore it upon waking up?

[scalz]

for more infos, easier to have a look at their docs but it's a lot smarter than esp8266.

[NeverDie]

@scalz
Fair enough. It's just that it's hard to know, as Nca78 points out, what parts of their documentation can be believed, and what parts are pure fantasy.