ESP8266 WiFi gateway port for MySensors

[Mark Swift]

@Yveaux I believe so, the nNF24L01 is connected to a NodeMCU (I've tried feeding the radio off both the onboard 3.3v and via VIN and a 1117 3.3v regulator (I have caps across the radio GND / VCC).

No changes were made to the sketch - I'm using the latest development build (also on the nodes). I've tried with all of the ESP Arduino board versions, doesn't seem to change anything?

[Yveaux]

@Mark-Swift said:

kept getting st=fail at various times

Now I read over your post again I get the impression that it sometimes works, and sometimes you get st=fail. Is this correct?

[Mark Swift]

Example - I'm sat here right now with one of my nodes and the gateway connected to my Macbook via USB.

When using the serial gateway I get =ok after each item no problem; in fact I can't make it fail.

When using the ESP8266 MQTT gateway I get:

Starting repeater (RNNRA-, 2.0.0-beta)
Radio init successful.
send: 3-3-0-0 s=255,c=3,t=11,pt=0,l=13,sg=0,st=ok:R+M+L+T+D+Rpt
send: 3-3-0-0 s=255,c=3,t=12,pt=0,l=3,sg=0,st=ok:1.1
send: 3-3-0-0 s=1,c=0,t=1,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=2,c=0,t=1,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=3,c=0,t=16,pt=0,l=0,sg=0,st=fail:
send: 3-3-0-0 s=4,c=0,t=15,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=10,c=0,t=1,pt=0,l=0,sg=0,st=fail:
send: 3-3-0-0 s=11,c=0,t=1,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=12,c=0,t=23,pt=0,l=0,sg=0,st=fail:
send: 3-3-0-0 s=255,c=3,t=15,pt=0,l=2,sg=0,st=ok:

send: 3-3-0-0 s=255,c=0,t=18,pt=0,l=10,sg=0,st=fail:2.0.0-beta
send: 3-3-0-0 s=255,c=3,t=6,pt=1,l=1,sg=0,st=ok:0
Init complete, id=3, parent=0, distance=1
Moisture Sensor: 0
send: 3-3-0-0 s=1,c=1,t=16,pt=2,l=2,sg=0,st=ok:0
Rain Sensor: 0
send: 3-3-0-0 s=2,c=1,t=16,pt=2,l=2,sg=0,st=ok:0
LUX: 54612
send: 3-3-0-0 s=3,c=1,t=37,pt=3,l=2,sg=0,st=ok:54612
Distance: 0 cm
send: 3-3-0-0 s=4,c=1,t=13,pt=2,l=2,sg=0,st=ok:0
No rain or moisture detected, landroid is not waiting: Status(0)
Landroid is free to go and is now waiting on the schedule: Status(0)
Landroid is home charging!
Sending landroidHome (1) status to gateway: send: 3-3-0-0 s=10,c=1,t=16,pt=1,l=1,sg=0,st=ok:1
Sending landroidWaitingTriggered (0) status to gateway: send: 3-3-0-0 s=11,c=1,t=16,pt=1,l=1,sg=0,st=ok:0
Sending timeElapsed (0) status to gateway: send: 3-3-0-0 s=12,c=1,t=48,pt=2,l=2,sg=0,st=fail:0

Using the serial gateway on a nano (Not connected to anything, just the nano and nRF24L01) I get:

Starting repeater (RNNRA-, 2.0.0-beta)
Radio init successful.
send: 3-3-0-0 s=255,c=3,t=11,pt=0,l=13,sg=0,st=ok:R+M+L+T+D+Rpt
send: 3-3-0-0 s=255,c=3,t=12,pt=0,l=3,sg=0,st=ok:1.1
send: 3-3-0-0 s=1,c=0,t=1,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=2,c=0,t=1,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=3,c=0,t=16,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=4,c=0,t=15,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=10,c=0,t=1,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=11,c=0,t=1,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=12,c=0,t=23,pt=0,l=0,sg=0,st=ok:
send: 3-3-0-0 s=255,c=3,t=15,pt=0,l=2,sg=0,st=ok:

send: 3-3-0-0 s=255,c=0,t=18,pt=0,l=10,sg=0,st=ok:2.0.0-beta
send: 3-3-0-0 s=255,c=3,t=6,pt=1,l=1,sg=0,st=ok:0
Init complete, id=3, parent=0, distance=1
Moisture Sensor: 0
send: 3-3-0-0 s=1,c=1,t=16,pt=2,l=2,sg=0,st=ok:0
Rain Sensor: 0
send: 3-3-0-0 s=2,c=1,t=16,pt=2,l=2,sg=0,st=ok:0
LUX: 54612
send: 3-3-0-0 s=3,c=1,t=37,pt=3,l=2,sg=0,st=ok:54612
Distance: 69 cm
send: 3-3-0-0 s=4,c=1,t=13,pt=2,l=2,sg=0,st=ok:69
No rain or moisture detected, landroid is not waiting: Status(0)
Landroid is free to go and is now waiting on the schedule: Status(0)
Landroid is out cutting the grass!
Sending landroidHome (0) status to gateway: send: 3-3-0-0 s=10,c=1,t=16,pt=1,l=1,sg=0,st=ok:0
Sending landroidWaitingTriggered (0) status to gateway: send: 3-3-0-0 s=11,c=1,t=16,pt=1,l=1,sg=0,st=ok:0
Sending timeElapsed (0) status to gateway: send: 3-3-0-0 s=12,c=1,t=48,pt=2,l=2,sg=0,st=ok:0

[Yveaux]

@Mark-Swift It does work sometimes, so that rules out a lot of possible problems!
How do you power the NodeMCU & rest of the hardware?
Both NodeMCU & nRF are very sensitive to low/noisy supplies.

[Mark Swift]

@Yveaux I power the NodeMCU from either my Macbook Retina, or from a decent USB wall plug, tbh, it seems to make little difference which I use. I presume everyone else is powering them via the USB also?

I've also tried a spare NodeMCU and a bunch of radios. In fact, I've been trying to get this working for 3 weeks, it's literally going to cause a divorce soon :-/

[Yveaux]

@Mark-Swift The log mentions Moisture Sensor, Rain Sensor, LUX, Distance.
Are these real sensor connected (and yes, how) or virtual?

[Mark Swift]

@Yveaux real sensors connected to my robot mower garage node...

[Mark Swift]

@Yveaux Node code is here for reference:

http://pastebin.com/mxkLq1Rp

[Yveaux]

@Mark-Swift To be frank: I don't know.
I'm quite sure it has nothing to do with the ESP port, or software in general.
You could try different orientation of the nRF, or move it closer to the sensor(s).

Are you using an amplified nRF24 btw?

[Mark Swift]

@Yveaux That's a shame, I'd really like to solve this. I've tried the radios orientated different ways but it makes little difference. Not using a PA+LNB, I was but gave up on it as I thought it was the issue (Now realise it wasn't).

Strange that I only see these issues with the ESP, and as soon as I switch to a Nano running the serial gateway they're gone.

[Yveaux]

@Mark-Swift I always use my sniffer in this case to see what's going on on air, but I don't know if your marriage will survive another sniffer build :innocent:

[Mark Swift]

@Yveaux I doubt it will, seems I'm destined to build a serial gateway, crap, I've invested so much time in this and tried everything!

[Mark Swift]

@Yveaux How about this, for a laugh I just tried rolling the ESP8266 gateway back to v1.5 (master).

Guess what? It works every time with no fails (communicating with my 2.0.0 beta node). So the issue lies with the latest development version somewhere, any ideas?
~~
Insane, I just took the gateway down the other end of the house and not one fail, wow! So okay, what was changed in version 2.0.0?

UPDATE: After more investigation, the reason it was working is because v1.5 only has a straight forward gateway version, and until now I've been using the MQTT client. It seems the gateway versions work fine, but the MQTT client version create the fails!

UPDATE 2: Okay, I can reproduce the fails with the gateway too, let me try and explain. When the gateway is not connected to anything (for example Domoticz) I get OK all of the time, no fails whatsoever. If I connect to Domoticz, I start to get fails against certain child sensors (it's always the same!). I presume it's sketch related, or even the amount of node updates I'm sending? As I believe the fails are the same I'm seeing with the MQTT client.

UPDATE 3: Nearly 3 weeks in and I've just figured out what has been causing me all this pain. I've always commented out the default 9600 serial baud define on the ESP8266 sketches. It seems this combined with the 200ms TCP delay has been causing me my issues when connecting to my Windows 7 machine. S$%&*!

[Yveaux]

@Mark-Swift Well, look on it from the bright side: your marriage is saved now :bowtie:
There have been more reports from issues with connections to external servers (e.g. MQTT, Domotiocz) that take too long.
Maybe we should pull @hek in as he wrote the MQTT client for ESP.

[hek]

What adjustments do you propose to the MQTT sketch/driver?

[Yveaux]

@hek I'm not familiar with the exact implementation at the moment. Just hoped you would have a hunch where these issues might arise from or how we can prevent them...

[hek]

Nope, I don't get why baud rate would affect radio/tcp traffic. Maybe @Mark-Swift could shed some light on it.

[robosensor]

ESP8266 gateway is quite unusable in setups where TCP ACKs can be delayed. This is because thread is blocked until TCP ACK received in ESP wifi code. So all wireless messages will be rejected if NRF24 receives more than 3 messages during this 200ms delay: http://forum.mysensors.org/topic/1870/esp8266-wifi-gateway-port-for-mysensors/235

[hek]

Hmm.. blocking code.. crap.

[robosensor]

The problem is slightly wider than it seems. Wireless messages should be fetched from internal NRF24 FIFO and stored into memory as soon as possible (using interrupts?). But lack of RAM makes the problem difficult to be solved.

The same thing happens not only with ESP8266 gateway and TCP ACKs, but also with all blocking calls during sensor reading (for example for old blocking call to read DS18B20 data with 750ms delay), with improper use of delay() instead of gw.wait() and so on.