1
\$\begingroup\$

I've got 8 ds18b20 (knock-off I assume) temperature probes monitoring the temperature in a freezer. I'm outputting the value of those probes each minute to an RRDTool graph using a Raspberry Pi.

However, occasionally I get some very strange readings and was wondering if anyone here would be able to help me figure out why.

Here's a graph for one of the eight probes (they're all pretty much the same):

RRDTool Tempearture logs

AS you can see there are 8, somewhat evenly spaced spikes in the graph.

The interesting thing is that I have 2 other probes also connected to the same 1-wire network that are just outside the freezer. They do not have any spikes at all:

enter image description here

The probes are hanging in free-air inside the freezer, not touching any surface.

Initially I thought that maybe there was some sort of electrical interference, maybe caused by the freezer's compressor. And this may still be the case, I'm not sure, but the fact that the 2 ambient temperature probes don't show any affect indicates to me that it's probably not interference. The external probes are pretty close to the freezer as well.

Also there's the fact that immediately after the spike, the temperature doesn't return to the value that it had before the spike, but seems to be cooling down and slowly leveling out as it reaches the freezer's set temperature.

I'd like to rule out electrical interference before continuing my investigation. If anyone knows whether this is possible, I'd really like to hear from you.

Some more information that may help:

It's a chest-freezer, It's not in optimal state as there's a small "leak"(?) there's water around it's base, I use old cardboard to absorb the water and it usually evaporates from the cardboard quite quickly.

EDIT:

So it looks like it's a daily thing. Here's some more graphs for the same sensor:

enter image description here enter image description here enter image description here

I'm guessing that this is just a 'feature' of the freezer, and even though it looks highly inefficient, I'll just have to live with it.

\$\endgroup\$
11
  • 1
    \$\begingroup\$ And what makes you think that these are not the real temperatures? Do you have known good probes that show different data? \$\endgroup\$
    – PlasmaHH
    Jan 16, 2017 at 11:06
  • \$\begingroup\$ Hey @PlasmaHH, thanks for the question, The only reasons is that it seems highly unlikely. I do not unfortunately have other probes, only the 10 ebay-bought ones. Would a regular (nothing-special-about-it) freezer purge cold air in this fashion? I've never heard about this sort of thing happening, surely that's a very inefficient way of doing things? \$\endgroup\$
    – Jim
    Jan 16, 2017 at 11:14
  • 1
    \$\begingroup\$ Well, you already said that it somehow leaks and generates water, so "regular" certainly does not apply here. You could try correlating things to the compressor activity, but I would recommend first checking if the probes work fine by adding an independent means of measurement. Don't you have e.g. k-type probes for your multimeter? \$\endgroup\$
    – PlasmaHH
    Jan 16, 2017 at 11:20
  • \$\begingroup\$ No I don't, just regular red-and-black positive and negative cables. But you're right, some sort of other 'independent' measuring is required. I'll have to figure out whether this is a 10PM-to-8AM regular occurrence and if so, camp out with the freezer for a while during that time and see if I can see/hear what's happening. \$\endgroup\$
    – Jim
    Jan 16, 2017 at 11:24
  • 1
    \$\begingroup\$ You need an independent measurement as per @PlasmaHH comment. \$\endgroup\$
    – Andy aka
    Jan 16, 2017 at 11:59

3 Answers 3

Reset to default
2
\$\begingroup\$

The drop in temperature during Sun AM establishes that the compressor works without producing pickup in your sensor, particularly if (as you indicate) the wires were not moved during the test, although the little glitch near the beginning of the trace may well be such pickup.

This pretty clearly indicates that the temperature swings you're seeing are real. The regularity of the swings, especially in amplitude, and to some degree in timing, also argue that the swings are real, and the freezer controller is doing something odd. The suggestion that this is a defrost cycle seems reasonable, except that I'd expect that more often, like once a day or something similar.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for the answer. You're probably right about the temperature drop on Sunday morning. Good spot. I'm not sure about the glitch near the beginning. It's interesting that it is a glitch towards 0 on both graphs. Unfortunately the server running apache and hosting those graphs has gone down, so I can't see what's been happening at all right now. But the Pi should still be up and running and recording all the temperature goodness. I should have some more data tomorrow. \$\endgroup\$
    – Jim
    Jan 17, 2017 at 14:06
1
\$\begingroup\$

AS you can see there are 8, somewhat evenly spaced spikes in the graph.

could be someone opening the door, or defrosting, etc.

put a few sensors there and if they all behave the same, you can rule out a bad sensor.

\$\endgroup\$
3
  • \$\begingroup\$ There are 10 sensors on the same 1-wire network, 8 of which are in the freezer, all 8 had the same kind of spikes at the same time. \$\endgroup\$
    – Jim
    Jan 16, 2017 at 12:25
  • \$\begingroup\$ would the same 8 sensors behave comparably if they were used to measure something else? like room temperature? it seems to me that the chance of all 8 of them going bad isn't that great. \$\endgroup\$
    – dannyf
    Jan 16, 2017 at 12:26
  • \$\begingroup\$ I agree, I just hope that there's no interference on them from something like the compressor. I'll try moving the sensors within the freezer to a new location within the freezer itself a bit later and see what the graphs look like tomorrow morning. \$\endgroup\$
    – Jim
    Jan 16, 2017 at 12:30
1
\$\begingroup\$
  1. The temperature is being controlled to better than 4°C. You haven't given any specs on what the temperature should be, so this is obviously within spec.
  2. Perhaps someone opened the freezer to cause the 8 events. You haven't said anything about how controlled the conditions are for this test.
  3. The 8 events could be some auto-defrost cycle.

In short, you've given no reason that indicates this isn't part of normal operation, or that the result is out of spec. There is no problem here.

\$\endgroup\$
1
  • \$\begingroup\$ (1) The temperature should be around the -21°C mark. (2) freezer remained closed. (3) I'm thinking this, however, I would like to leave this here to see if anyone else has experience anything like this (as with the comment by @Vladimir Cravero on the question above) with more solid info. \$\endgroup\$
    – Jim
    Jan 16, 2017 at 12:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.