Hi. I am in the process of completing a home-brew central heating controller which will be able to measure the temperature outside of the house as well as in 3 rooms inside.
Our house faces north/south and I plan to mount the outside sensor just under the eaves but do I mount it on the northerly side or the southerly?
Obviously, the northerly side never sees the sun but I am concerned that the sensor may be affected by stored heat, or should that be cold :) so, has anyone got any advice please?
In article , Phil writes
You don't want it to experience any solar gain so put it on the North side.
Thermistors are low cost sensors and pretty standard in commercial gear,
10k at 25degC is a typical example that would be fine on a long line using twisted pair (phone or cat5e). I didn't bother using shielded cable.A simple IP68 outdoor junction box will be fine for a housing and sticking the sensor to the lid would give it a low thermal mass with a response as least as fast as the weather.
On Thursday 19 December 2013 16:21 Phil wrote in uk.d-i-y:
On the north wall definately assuming you are in the northern hemisphere!
The mine for my weather compensator recommended the north side, and out of direct sun.
Keep in mind that even if you place the sensor on a north-facing wall, between the Spring and Autumn equinoxes the sun rises and sets north of east and west respectively, and between those times the sensor might give falsely-warm readings through being in the sun. Probably not an issue in full summer, but you can get really cold nights/days in March and April, and perhaps early May.
One of my garden fences faces north, and I had to place a temperature sensor on it where the morning sun can't shine on it, by hiding it behind a fence post. Evening sun isn't an issue as it sets here in summer at about 4:30pm due to 70' high trees on the ridge-line at the back of the house.
Both. Position two sensors such that one of them is always in the shade an d use the lower reading. Keep them spaced as far from the wall as you reas onably can - 0.5m perhaps om a stalk of some kind. That way you avoid the boundary layer of warmer air that rises up the walls in still winter condit ions.
John
You are one a hiding to nothing. These devices have long been available and took decades to perfect, some don't work well even now. So anything you make is unlikely to work well.
Go out and buy one.
On Dec 19, 2013, Phil wrote (in article ):
Temperature always varies a lot around heated buildings. If you really need to measure the true ambient outside temperature accurately you need to construct (or buy) a proper Stevenson Screen and mount it some distance away from your house. The official recommendation is a distance of twice the height of the building (or so I believe).
Meteorologists use a Stevenson Screen for housing measuring equipment. It's basically a louvered box, painted white, that stands on its own, getting full sun when it shines. The louvres keep the interior shaded but well ventilated. See
Put your sensor on a north wall, in a white plastic box with plenty of holes drilled in the upper and lower edges for ventilation.
On Thursday 19 December 2013 19:42 harryagain wrote in uk.d-i-y:
By that theory, toasters should actually be able to evenly and repeatibly toast bread by now!
Since they are basically a small plastic box with a thermister inside.
I have reasonable confidence that anything one makes will work equally as well as a commercial one, and that the performance will be more than adequate for most applications.
yeah right
instructions that came with
On Thu, 19 Dec 2013, "fred" writ:
I already have TMP36 remote solid state sensors using STP cable and they are pretty accurate. Certainly within 0.5C from -20C to +50C on a 5 metre cable. I have had them tested at work. You really must have STP as any noise messes up the readings as I found out when I tried them with UTP cable :o(
Thanks for the suggestion, I hadn't thought of IP68 boxes.
In message , Percy writes
I was about to reply to fred but you got in first. I too, have TMP36 sensors and I'm using the circuit given by the manufacturer which seems to be wrong. What remote V to I converter are you using?
I also agree that the IP68 box idea is good.
In message , harryagain writes
Are you for real? I would put my home-brew against any similar commercially unit any day. I use high stability components and the software does what I want, how I want it. It hasn't dropped off of the end of a Chinese production line.
Here's the basic spec of my controller:
Clock accurate to 500mS per day with battery backup
16 Analogue to Digital sensing ports 16 controllable outputs with relays, if required, to control pumps, motorised valves, outside lights, internal lights etc. Supports up to 8 remote keypads with LCD displays Fully programmable ON/OFF time periods and temperaturesTotal component cost including sensors but not the relays, GBP15.00.
The software writing is the really fun part.
So, where can I buy an equivalent controller and for how much?
Sounds like a good idea. I'll give it some thought.
In article , Percy writes
Useful device, thanks.
I figured the coarse and robust measurements of thermistors would suit this app if you were rolling your own system hardware and software.
Being 2 terminal devices they lend themselves well to balanced 2 wire measurement, noise in one limb is equally induced in the other, resulting in cancelling of noise if measured in balanced mode.
In a single ended voltage output device though (like TMP36) you get no benefit from twisted pairs although using a pair for the output and 0V should give the least susceptible result.
In either case, the way to avoid noise it filter it out or integrate the analogue output as response time is not important.
There are direct digital output devices available from Dallas that would do the job too. Another poster here has used them in his home automation system, search for Andrew Gabriel as poster and Dallas and you should find the devices he used. I think they're under a fiver each and have good application notes for interface software too. They have a large fanbase that has resulted in quite a few apps being written for them. It's recommended to range check your input, ignoring any results that are dramatically out from the norm as this weeds out received data that have been corrupted by (much heavier) noise.
Good for corrosion protection (condensation) and to protect from insect ingress. Some view it as inevitable that condensation will occur in outside enclosures (the boxes aren't vapur tight) so they drill a tiny (1mm) hole at the very bottom to allow drainage, the small hole minimising the risk of insects nesting. Another trick is to seal the sensor in adhesive lined heatshrink or to dip the wired assembly in thick varnish. Marginal increase in thermal lag but protects from corrosion.
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.