Background: A couple of times I've had my garage freezer defrost. I'm pretty sure it's not failed, just that the door isn't properly shut, so I want to find a simple way to monitor it. I already have some basic wireless energy monitors and I figure a reduction in average usage (eg it failed) or an increase (door open) would be a good place to start.
I've not done anything like this before and I'm thinking about building it around a Pi. But for the "if the energy drops below X or goes above Y for a period of Z then send me an email/text message/other alert" bit I figure I should use something standard.
IFTTT came to mind (never used it, been aware of it for years). Or, if I care about my data, a FOSS alternative (Huginn and Bip.io came up from a Google search).
Anyone here had any experience of anything like this to make a recommendation?
This is intended to be the start of an adventure where I start playing with all sorts of potential triggers and actions so to be honest I'm not so fussed about a specific solution as I am in finding a gateway into something to play with. IoT is something I'd like to experiment with in general, but due to lack of time I really need to be able to get something useful up and running pretty easily that I can then extend when time permits.
Mark
On 17/07, Mark Rogers wrote:
IFTTT came to mind (never used it, been aware of it for years). Or, if I care about my data, a FOSS alternative (Huginn and Bip.io came up from a Google search).
Anyone here had any experience of anything like this to make a recommendation?
I don't know of any alternatives but I'll be interested in hearing about anything you find :)
Steve
On Fri, 17 Jul 2015 15:28:40 +0100 Mark Rogers mark@more-solutions.co.uk wrote:
Background: A couple of times I've had my garage freezer defrost. I'm pretty sure it's not failed, just that the door isn't properly shut, so I want to find a simple way to monitor it. I already have some basic wireless energy monitors and I figure a reduction in average usage (eg it failed) or an increase (door open) would be a good place to start.
Surely you're monitoring the effect rather than the cause. If the cause is that the door is left open, monitor the door! It would also be much simpler to do.
On Fri, Jul 17, 2015 at 04:24:05PM +0100, Chris Walker wrote:
Mark Rogers mark@more-solutions.co.uk wrote:
Background: A couple of times I've had my garage freezer defrost. I'm pretty sure it's not failed, just that the door isn't properly shut,
Surely you're monitoring the effect rather than the cause. If the cause is that the door is left open, monitor the door! It would also be much simpler to do.
I'd suggest that the thing you *really* want to monitor is neither the door or the power. If there is some doubt about what the problem is you want to measure the temperature (and probably the door and power too).
I would suggest the really easy way to do this is get a computer, get a temp sensor, get some kind of energy monitor and something that can read open/shut for the door and graph it all with a tool like munin or nagios and then set some alerts on that if you really want.
Adam
On 17/07/15 18:13, Adam Bower wrote:
On Fri, Jul 17, 2015 at 04:24:05PM +0100, Chris Walker wrote:
Mark Rogers mark@more-solutions.co.uk wrote:
Background: A couple of times I've had my garage freezer defrost. I'm pretty sure it's not failed, just that the door isn't properly shut,
Surely you're monitoring the effect rather than the cause. If the cause is that the door is left open, monitor the door! It would also be much simpler to do.
I'd suggest that the thing you *really* want to monitor is neither the door or the power. If there is some doubt about what the problem is you want to measure the temperature (and probably the door and power too).
I would suggest the really easy way to do this is get a computer, get a temp sensor, get some kind of energy monitor and something that can read open/shut for the door and graph it all with a tool like munin or nagios and then set some alerts on that if you really want.
Adam
Not sure if it's the simplest way, but there's a Pi based project called Open Energy Monitor. This can monitor power usage, and it can also monitor temperature too. http://openenergymonitor.org/emon/
Any use?
Cheers Steve
On 17 July 2015 at 21:22, steve-ALUG@hst.me.uk wrote:
Not sure if it's the simplest way, but there's a Pi based project called Open Energy Monitor. This can monitor power usage, and it can also monitor temperature too. http://openenergymonitor.org/emon/
Thanks for this, it looks useful (although more so in general than for this specific project).
However it also looks quite expensive: £70 for a transmitter with CT, plus £20 for the Pi receiver module.
What I currently have is a CurrentCost monitor (~£50, with USB interface) which gets data from appliance monitors (~£25/ea). I think this might now be discontinued as a product though.
I've also looked at the new Mi|Home kit [1] which appears to be able to talk directly with IFTTT (which was one reason I started thinking along those lines) but I haven't looked yet to see how FOSS friendly the Mi|Home stuff is.
[1] http://www.ledhut.co.uk/products-of-the-future/smart-home-products/adapters/...
On 18 July 2015 at 10:00, Mark Rogers mark@more-solutions.co.uk wrote:
I've also looked at the new Mi|Home kit which appears to be able to talk directly with IFTTT (which was one reason I started thinking along those lines) but I haven't looked yet to see how FOSS friendly the Mi|Home stuff is.
I'd forgotten that Energenie also have Pi-Mote [1] which suggests a level of FOSS friendliness
[1] https://energenie4u.co.uk/index.phpcatalogue/product/ENER314
I think I might give this a try.
On 17 July 2015 at 18:13, Adam Bower adam@thebowery.co.uk wrote:
I'd suggest that the thing you *really* want to monitor is neither the door or the power. If there is some doubt about what the problem is you want to measure the temperature (and probably the door and power too).
Agree that temperature would be the ideal property to monitor. However I have looked at wireless monitors without finding anything that looked much good (especially if I wanted to get the data out into something else). A wired sensor would be OK but I'm not sure about the effect that might have on the door seal and whether it might exacerbate any issue I have with the door potentially not shutting properly. But in principle a temperature measurement would be preferable.
I would suggest the really easy way to do this is get a computer, get a temp sensor, get some kind of energy monitor and something that can read open/shut for the door and graph it all with a tool like munin or nagios and then set some alerts on that if you really want.
Alerts are the most important point, as aside from the first few days after setting it up I'm not going to be interested enough to go and look at pretty graphs of my freezer temperature - especially given they really shouldn't vary very much if the thermostat is doing its job!
I don't have any experience with Munin, and the experience I had with Nagios wasn't positive (too complicated to set something simple up). I'm leaning towards the IoT approach where I could have lots of things "tweeting" their status and something else "following" them and reacting to changes. I *think* IFTTT would be a candidate for this but I could be wrong.
On Sat, Jul 18, 2015 at 09:23:14AM +0100, Mark Rogers wrote:
I'm leaning towards the IoT approach where I could have lots of things "tweeting" their status and something else "following" them and reacting to changes. I *think* IFTTT would be a candidate for this but I could be wrong.
To me, that sounds stupidly complicated when what you're doing is monitoring something!
Adam
On 19 July 2015 at 22:31, Adam Bower adam@thebowery.co.uk wrote:
To me, that sounds stupidly complicated when what you're doing is monitoring something!
Can I ask why you think that?
Setting aside the secondary objective (to learn something I can extend into different areas), the primary objective is to get notifications if the freezer power usage goes outside certain ranges.
Somehow, therefore, I need to monitor something and when certain criteria are met I need to get a message to (say) my phone.
I could script something myself (I have played a bit with the data that the CurrentCost devices give me over USB) but that's not really going to be easy to extend and means reinventing some wheels. I could use a typical monitoring package, such as Nagios (which pretty much has "stupidly complicated" as its tagline in my experience!)
What I'm currently suggesting is that I use something which is already capable of sending data into "the cloud" and use tools already there (IFTTT) to validate the data and send me notifications (email, text, or to a phone app (I believe)). There's a massive amount of complex infrastructure involved in doing that I agree, but at my level it should still be just a case of connecting a few building blocks together, and I then have the rest of that infrastructure to play with in the future.
The biggest downside is handing my data over to a third party I have no control over (free beer without freedom), which is why I was looking at FOSS IFTTT alternatives - although it's probably not a bad idea to try out the hosted IFTTT first before trying to set something up of my own (and I'd certainly need a bigger use case than checking my freezer before that was worthwhile).
On Mon, Jul 20, 2015 at 09:42:30AM +0100, Mark Rogers wrote:
On 19 July 2015 at 22:31, Adam Bower adam@thebowery.co.uk wrote:
To me, that sounds stupidly complicated when what you're doing is monitoring something!
Can I ask why you think that?
Because you want to pull some data from a device and do something with it.
If you control all those systems then this is easy, pushing it to the "cloud" (which is a load of bollocks in this scenario) you are adding further steps to something which has already existed for over 20 years and is well understood and a problem that is pretty much solved.
Adam
On 20 July 2015 at 22:50, Adam Bower adam@thebowery.co.uk wrote:
If you control all those systems then this is easy, pushing it to the "cloud" (which is a load of bollocks in this scenario) you are adding further steps to something which has already existed for over 20 years and is well understood and a problem that is pretty much solved.
Surely I can't avoid the "cloud" in the broadest sense, if I want to get notifications to what is essentially a cloud device anyway (my phone).
Having control of the key systems is preferable (hence looking for self-hosted alternatives to IFTTT, which increases complexity beyond a sensible level unless I'm doing it for other reasons as well, which I am in this case).
The age-old monitoring systems I have looked at (such as Nagios, or the sort of tools you might use for graphing your CPU temperature) have been around long enough but they're not trivial to use (unless you already have the skills, which I don't).
It feels like a parallel argument might be that Gmail is overkill for email when all I need is a local Postfix and Mutt installation (where of-course the latter is actually far more hassle than the former). I'm sure that's not the argument you're making, but I'm not quite seeing what I'm missing. (There are dozens of reasons why Postfix+Mutt might be preferable of-course, but "overkill" isn't one of them.)
I guess what I'm asking is: if the cloud is the wrong way, what tools *should* I be using to do it differently. You mentioned nagios/munin before but I'm struggling to see how setting that up is going to be easier in the absence of prior experience with either. I haven't tried Nagios recently but I have tried hard with it on and off over the years in the hope of getting a basic server monitoring setup working without ever really getting something that delivered what I couldn't do in a few minutes using free (beer) cloud tools elsewhere. (This isn't about the freedom aspect, it''s about which is the most "stupidly complicated". The "freedom" aspect I'm already sold on.)
Mark
On 17 July 2015 at 16:24, Chris Walker alug_cdw@the-walker-household.co.uk wrote:
Surely you're monitoring the effect rather than the cause. If the cause is that the door is left open, monitor the door! It would also be much simpler to do.
Actually power usage is a symptom rather than the effect I actually care about (temperature drop), but it's relatively easy to monitor with the right hardware, which I already have (and I'd be interested to know what the usage is like anyway).
I'm no sure that monitoring the door being open would actually be that easy. In this context "open" would mean "almost but not completely closed", and to be honest I'm only guessing at this point that the door not being properly shut is the cause so if that guess is wrong I'm going to miss the next time it happens.
-
Adam Bower -
Chris Walker -
Mark Rogers -
Steve Engledow -
steve-ALUG＠hst.me.uk