Basenode woes

I have had problems in the past with my basenode but all has been running well until this recent cold snap came in. First I had an issue with boiler control lua script. It was crashing and taking down plugboard, obviously no longer controlling the heating. As all the rommnodes no longer reported after the crash I suspected the base node.

Before replacing I hooked up to a serial port and got the following at start of comms....

[RF12nocfg.1] A i1 g212 @ 868 MHz

I transferred the AtMega to a spare node and then I got...

[RF12nocfg.1] A i1 g212 @ 868 MHz

Member for

14 years 5 months

mark_baldwin

13 years 7 months ago

I fitted a new RFM12 to the basenode. Now I keep losing one of the nodes

Very strange...

Member for

17 years

derek

13 years 7 months ago

I'd first wondered if the issue that you are seeing was down to something temperature related. However, the outside monitor seems quite happy with the unseasonally cold readings. 

The baseNode iteslf looks good - in that the outside reports are steady. I'm sure that you'll have already checked the antenna length/position on the Living Room roomNode.

One thing that might be worth considering is reviewing the roomNode code to see if the AVR watchdog monitor is in use. I've not been there yet, but if there is an intermittent hangup on the micro, this would perhaps bring it back to life in a very short space of time.

Keep us posted on any progress.

Derek.

Member for

14 years 5 months

mark_baldwin

13 years 7 months ago

I've swapped over the living room and conservatory nodes. Now the one that was in the conservatory and is now in the living room has started dropping off. So maybe I have a range issue as the living room is farthest away.

Member for

17 years

derek

13 years 7 months ago

Try doubling the antenna length on that roomNode. I've also noticed that, despite being battery powered, the roomNodes do like to be away from mains powered kit. Definately away from TVs and PCs.

RF. Gotta love it.

Derek.

Member for

14 years 5 months

mark_baldwin

13 years 7 months ago

I'm gonna double check everything on the basenode first. That's the only thing that has changed. It has a magmount antenna so that shouldn't be an issue and all was fine until the RFM12 failed.

Member for

14 years 5 months

mark_baldwin

13 years 7 months ago

Checked the coax to the basenode antenna. Got a strange reading of ~250kΩ between screen and core. Cut the cable back an inch and remade, then the reading was ∞. All seems fine now, allbeit with the roomnodes swapped over but that doesn't really matter as they are identical except for the position of the temperature sensor. I'll have to extend the wiring to the DS to bring it outside the case. It's a little too unresponsive for heating control when insulated inside the node case.

Member for

14 years 6 months

magill

13 years 4 months ago

see next post

Member for

14 years 6 months

magill

13 years 5 months ago

Hi
I have started getting plugboard errors like this:

lua: /usr/share/lua/5.1/xap/init.lua:395: /usr/share/lua/5.1/xap/jeenode.lua:51: invalid value (nil) at index 2 in table for 'concat'
xAP Message being processed.
xap-header
{
uid=FF00D500
source=dbzoo.livebox.Serial
hop=1
class=Serial.Comms
v=12
}
serial.received
{
data=OK 34 215 1OK 34 21OK 2 215 0 200 0
port=/dev/ttyUSB0
}

stack traceback:
[C]: in function 'error'
/usr/share/lua/5.1/xap/init.lua:395: in function 'dispatch'
/usr/share/lua/5.1/xap/init.lua:97: in function 'fun'
/usr/share/lua/5.1/pl/List.lua:444: in function 'foreach'
/usr/share/lua/5.1/xap/init.lua:96: in function 'callback'
/usr/share/lua/5.1/xap/init.lua:203: in function 'dispatch'
/usr/share/lua/5.1/xap/init.lua:117: in function 'process'
/etc_ro_fs/plugboard/plugboard.lua:77: in main chunk
[C]: ?
Loading /etc/plugboard/JMaliasApplet.lua [ Alias with Heating Interpreter ]
Loading /etc/plugboard/JMjeenodeApplet.lua [ JeeNode ]
Loading /etc/plugboard/JMxapCacheWebserverApplet.lua[ xAP Caching web servlet ]
Running...

This locks up the plugboard and requires a reboot.
I have a roomnode and a roomnodetwin attached. It looks like a data collision but I don't know how to sort it.
thanks
John

Member for

15 years 7 months

garrydwilms

13 years 5 months ago

Looks like your getting combined jeenode signals. Strange. 

What versions are you using. There have been quite a few changes recently although they work ok for me. 

That said, I haven't used the recent HAHcentral node yet!

Member for

14 years 6 months

magill

13 years 4 months ago

Garry

I've upgraded to the 309 betas and now back to 309 all giving the same errors.

I have also flashed the new HAHcentral to no effect.

John

Member for

15 years 7 months

garrydwilms

13 years 5 months ago

How odd. This just started happening?  No obvious changes to initiate it?

what do you get if you monitor the base node output (via putty or arduino serial monitor, etc)

Is this a node issue or xap-serial?

 

Garry

edit: just seen this old post. Is this the same issue?

/content/hahcentraljeenodeapplet-issues

Member for

16 years 10 months

brett

13 years 5 months ago

Well its not happy as this data string is clearly broken as the data repeats three times - see the OK marker ?

data=OK 34 215 1OK 34 21OK 2 215 0 200 0

The data is different which would imply you might have either

Two nodes with an ID of 2 there is some data overrun happening?  Perhaps xap-serial gets two data feeds really fast and drops the \n and combines them into one.   I'll look if this could happen.

Brett

Member for

14 years 6 months

magill

13 years 4 months ago

Thanks Garry for the link.
Kevin seems to have got his setup sorted, but begs the question why is no-one else having issues?
I had been experimenting with current cost and had left it enabled on the GUI. When I disabled it all came good.
No more corrupt data like this:

data=OK 34 215 1OK 34 21OK 2 215 0 200 0
data=OK 4 1 0 246 168? 33 70 255 8 186 45 66 15 214 51 45 190 45
data=? 139 1476 227 31 117 217 23 233 215 61OK 4  ? 3  196 124 128 138 153 237 192   169 238 155 214 119 244 79 24 ? 59 60 121 246 6
data=OK 4 1 0 23OK 5 217 1 0 162 48 50 38 146 12
data=OK 5 218 1 0 161  ? 141 161 149 151 196 14 37 254 146 94 ? 185 25 2 18 23 43 77 183 177 68 120 6OK 5 217 1 0 16O ? 182 0 130229
data=OK 4  ? 10 114 241 150 145 246 234 116 0 187 255 167 252 251 51 77  93 174 102 40 158

Anyway same question, has no-one else encountered this?
John

Member for

16 years 10 months

brett

13 years 4 months ago

You had TWO processes bothing trying to access the same serial port /dev/ttyUSB0
Those being xap-currentcost and xap-serial.  Strange that the system exhibited this corruption as a symptom I would have thought the second request to open the port would have been flat out denied as the port was already open.  Having said that I've seen this being allowed before.

It would certainly explain why nobody has run into this as its a configuration error not a software one.  Glad you figured it out.

Brett

Member for

16 years 10 months

brett

13 years 4 months ago

I was looking at xap-serial code and I could see how this could happen as each line is processed and sent as a xAP packet.
It would have to mean that the corrupt string you are pointing at would have had to be emitted from the baseNode itself, which I also find stange as it too can't do this.  Which only leads me back to the conclusion that two process esopen the same port and both where consuming data.

That is not to say there isn't a bug, anything is possible, I can't see how it can be happening.

UPDATE: I tested this and I was able to open the same serial port from multiple processes.  I've made code changes so that a serial port is opened exclusively to a single process. 

With this code in place for John's misconfiguration xap-currentcost would have grab it first, as it starts before plugboard, and  xap-serial would deny plugboards request instead push an xAP serial.error packet to say it can't open the port something already has it open, and all your nodes would have stopped working.  From there hopefully it would have been a little easier to diagnose this configuration issue.

Brett

Member for

14 years 6 months

magill

13 years 4 months ago

Thanks Brett. Silly oversight by me. Sometimes you can't see the wood for the trees!