Author Topic: Wireless outdoor thermometer  (Read 2055 times)

0 Members and 1 Guest are viewing this topic.

Offline pfletch101

  • Senior Contributor
  • ****
  • Posts: 242
    • Personal Website
Re: Wireless outdoor thermometer
« Reply #25 on: March 07, 2018, 08:11:07 PM »
One more question:  any reason why you didn't just display the results on the Raspberry Pi?

Because I don't have a display on it (and didn't see a reason to get one) and I need them on my main machine, anyway. The inside temperature trends in a couple of places are among the things that I record and use to manage my HVAC system.
Vantage Pro 2+ with WeatherLink - data incorporated in domestic energy production (PV) and use monitoring system.

Offline Bushman

  • Forecaster
  • *****
  • Posts: 6927
    • Eagle Bay Weather
Re: Wireless outdoor thermometer
« Reply #26 on: March 07, 2018, 11:08:53 PM »
OK - you can run headless.

Offline pfletch101

  • Senior Contributor
  • ****
  • Posts: 242
    • Personal Website
Re: Wireless outdoor thermometer
« Reply #27 on: March 07, 2018, 11:19:19 PM »
OK - you can run headless.

I do, but it is a lot easier to have readings reported to a gadget on my Windows desktop than to have to look at the virtual display.
Vantage Pro 2+ with WeatherLink - data incorporated in domestic energy production (PV) and use monitoring system.

Offline pfletch101

  • Senior Contributor
  • ****
  • Posts: 242
    • Personal Website
Re: Wireless outdoor thermometer
« Reply #28 on: March 09, 2018, 09:44:20 AM »
Also, what kind of range does the little dongle/antenna setup give you?  Do you use an outside antenna with the little coax over to the dongle?  Your solution sounds interesting.

Sadly, it looks as if my initially positive earlier response to this question was premature, though probably more or less correct if you are only monitoring one sensor, or if you have more than one sensor but all of the sensors are located at similar distances/separations from your receiver. There are a couple of issues:
1) The time between transmissions for each sensor is approximately 48 seconds, but it is not precisely the same for each sensor. Consequently the temporal separation between transmissions changes over time, and two sensors can end up transmitting at very close to the same time for a substantial period. Under these circumstances, the data from the more 'distant' sensor is completely lost.
2) Even if the sensors' transmissions are reasonably well spaced in time, the presence of a closer sensor appears seriously to interfere with data reception from one that is further away. I am using the automatic gain setting for rtl_433, and this issue may reflect the auto-selected gain (perhaps the average of what would be the two optimal values) being inadequate for the more distant sensor. In testing a two-sensor configuration in which only occasional data from the 'far' sensor was being recorded, moving the 'near' sensor to a location close to that of the 'far' sensor led to the restoration of almost 100% reception from both. I will be exploring the use of manual gain settings, to see if this helps.

{Edited March 14} There is a workaround for the problem noted in 2) above, but it involves a bit of fiddling with the RTL_433 options. Further study has largely confirmed my impression that the hardware's automatic gain control does not respond fast enough to cope with two (or more) alternating burst transmissions with very different signal strengths on the same frequency. With a bit of experimentation, however, it may be possible to find a 'manual' gain setting that allows both signals to be properly handled - there were two such in my situation (with everything back in its original configuration). Parenthetically, it appeared that the issue in my case was not primarily that the more distant sensor's signal was too faint but that the nearer one's signal was so strong that it pushed the preamplifier into overload with anything but a quite low gain setting.

In the course of exploring these issues, another, unrelated but significant, issue became apparent: the ID of a given sensor is not (as I had assumed) hard-wired into it; it appears to be selected at random on power up (battery insertion), though it does then remain constant unless/until the batteries die or are removed. Changing the batteries for a sensor will, therefore, change its ID, which is certainly less than optimal for their long-term use in temperature/humidity monitoring applications.
« Last Edit: March 14, 2018, 05:21:15 PM by pfletch101 »
Vantage Pro 2+ with WeatherLink - data incorporated in domestic energy production (PV) and use monitoring system.