The Ultimate Thermometer

[Dirk]

Well dont despair too much, that group you mentioned did end up buying Insteon and their servers are back to life. You do have to pay them a fee to use the servers, but it is fairly nominal (something like $35 a year). They also now have stock of the controller so if yours fails it can be replaced. I believe they are trying to get more stock of all the other stuff. Anyway at worst it’s breathed a few more years of life into my Insteon.

Ah, good news. Best of luck to them! I never used the Insteon servers for anything, so I was unaffected by the "dark days." I do hope they can bring some of the modules back.

Buying stock is what's going to work for me. I expect that'll eventually fail, but might prolong my system for a good while. I don't think any of you should do that, though. I think you all should forget about this conversation and just go away, and leave all the stock to me!

 

[Katodude]

They are currently out of stock on the hub, but you can buy more dimmer switches.

Hub 2 Central Controller

The Hub 2 gives you control over a huge family of connected devices, all from your smartphone and tablet Create customized scenes for instant recalling of your favorite lighting settings, & create schedules for individual devices or scenes Get instant alerts from sensors for things like motion...

shop.insteon.com

 

[Dirk]

Was it the hub that needed their servers? I don't use the hub, so that must be why I was never affected by their server dealio. I see the eight button keypad! I have a need for that. That's very encouraging. I most need the little plug-in appliance link. And the in-the-wall outlet link. If they could bring those back, I'd be all set for a while. I think I have a spare PLM, but I'd buy a third one if that showed up. Without the PLM, it's over.

 

[Dirk]

Update!

We finally got some warm weather!! I was monitoring my pool temp today with the temp probe described previously, but couldn't figure out why the pool temp was not rising in the afternoon as it should. It kept reading 76°.

Then I remembered I had adjusted my auto-fill level this morning, to bring the water level up an inch. Oops.

Because the temperature probe is sitting half-in-half-out of the equalizer tube through which my auto-fill valve feeds the pool, the probe was being subjected to a flow of fill water all afternoon, which is colder than the pool.

Soon after I turned off the auto-fill, the pool temp sensor started reading correctly, about 80°.

So that's a flaw in my pool temperature sensing setup. Not a serious one, as the normal auto-fill action is a few "squirts" periodically, not enough to influence the water temp surrounding the probe much at all. It was only because I had a constant flow for a few hours that the sensor cooled off (I hope.)

As I mentioned previously, I considered using the light niche for the temperature probe, or hanging it down from the light, just below the niche. I was concerned the light's temperature might affect the reading. Guess I got stung anyway. The temperature probe needs a dedicated niche.

Anywho, I left myself the means to relocate the sensor to the light niche if I want to. The cable feeding the probe already has a bunch of extra feet, coiled up underground, for this very purpose. So moving it won't be all that much if I decide to. We'll see how the summer goes, and if I continue to see random inaccurate pool temps.

Other than that, the system has been working flawlessly, inside and outside, since I fired it up.

 

[Dirk]

As mentioned in post #44 above, my pool temperature sensor was hanging half in and half out of the equalizer tube connecting my auto-leveler system to the pool. This had three problems:
- the sensor was vulnerable to kids or my pool vac messing with it, and
- the flow of the fill water was affecting the temperature reading, and
- the sensor and a bunch of its wire all coiled up in the equalizer pipe was affecting the fill-water flow.

@kimkats had given me the idea to en-cage it somehow, and I finally got around to it.

I fabricated a cage out of 3/4" grey PVC. Then I shaved off a side so that there would be a decent glue surface.



Sanding the shaved side with a belt sander left a couple of handy PVC flaps that made the glue surface a little wider.



I got in the pool with a mask and snorkel and plotted the position I wanted it with a level and a pencil. Then I tried out my idea of using my new ProTuff pole as some sort of clamping mechanism. I had several ideas to try, to make that work, but it turned out the simplest was all I needed. The pole was virtually the exact length I needed (jammed between the two sides of my pool), and my brush did a perfect job of holding the cage in place. I practiced assembling the pole and cage against the wall, so that I wouldn't be trying to figure that part out with the epoxy applied.



Then I gloved up and mixed and applied the underwater epoxy to the cage. That did not go well at all. The epoxy was quite unruly and the gloves made it worse. I did the best I could, then scrambled back into the water to get the cage in place before the epoxy set. It ended up being a bit sloppier than I was hoping for, but it's the kind of thing that the more you mess with it, the worse it gets, so I left well enough alone.



I pushed on the cage as best I could, to try and embed the epoxy. I had no idea if that was working or not. And trying to push anything against a pool wall doesn't work well at all. Like I imagine working in outer space would be.

Then I had to disengage the vac, so it wouldn't upset the pole during its mid-night run. I decided not to mess with the programming, and I couldn't just remove it. I didn't want to turn off the breakers either, so I just did this instead:



This morning I removed the pole, half expecting the cage to just fall off. It didn't. So I pushed on it a bit. Then a lot. The epoxy seems to be holding very strong. We'll see if it lasts. I then fitted the sensor into the cage and shortened and re-soldered its wire so that the extra slack I had previously stored coiled up in the equalizer pipe was reduced to a single strand. Much cleaner and now the wire won't impede the flow of water through the tube.

It's uglier than I had hoped. The grey of the PVC is better than white, but not as invisible as I had envisioned. But now the sensor is protected and out of the flow of fill water. It's not any uglier than one of those tied-up, dangling pool thermometers, I suppose. But this one talks to my home automation system and I use its reading many times each day, even in the off season. So it's a trade off.

It's possible cooler, freshwater fill water coming out of the equalizer tube could drop downward and affect the temperature of the sensor. I tested for that as best I could and didn't see that effect. If it does, it'll be in the 0.1° range, which I can live with. I might find a way to plug the top of the cage to minimize that, but that's not a concern for now.

Like my water level ruler project, if I find I dont like this gizmo, or it doesn't work out, I can always try something else. But this project is almost done. Just have to hide the wire that crosses the deck. That's a project for another day.

 

[Brian PacNW]

Well done!

 

[plnewb]

Phenomenal project. I am getting addicted to pool chemistry and since launching rockets isn't stressful enough, I am going to be starting a few projects that will allow me to monitor my aunts pool, and then, eventually my own (at which point I will need to fire my pool guy, not because of financial reasons but because if I do things right, he wont have much to do)

Are there a few things you would have done differently?

Thank You for sharing!

 

[Dirk]

Are there a few things you would have done differently?

The little PVC cage I built for my underwater temp sensor is the kind of fabrication that just comes out better the second time you do it. So if that falls off or otherwise fails, the next one will be a bit better. And it's not as camouflaged as I had hoped, so I might have done that better, though short of painting it I'm not sure how. But other than that, I'm pretty pleased with everything.

What needed to be done differently is something I couldn't achieve. It's a limit of retrofitting. For example, if I ever built a house, I would wire the heck out of it. I'd have ethernet in every room, ceiling cams everywhere, a temperature sensor in every room, etc. So much easier during construction. I did some amazing wiring considering my obstacles, but I can't do all I still want to. Same for a pool. I've mention this elsewhere, and I don't imagine it'll ever catch on, but a pool should be built with some sort of sensor array (that has yet to be invented other than in my mind). Something you could install like a light niche, recessed and inconspicuous, and be able to access and modify after the pool is full (again, just like a light niche). This sensor niche could house pH and FC sensors, a water temp sensor, a salinity sensor, etc. And be upgradable over time as water monitoring technology improves. Instead, we have floating thermometers and have to scoop out water every few days to test it. Even high-end chemistry monitoring systems rely on drawing water at that pad, which can only be done while the pump is running. Or there are types that float in the pool but they require batteries and WiFi. No thanks. The sensors should be "hard wired" in the pool, and available 24/7. We're still in the Stone Age when it comes to pool telemetry. It's barbaric, I tell you!

The cables for all these sensors and gizmos would run back to the pad, buried deep inside conduit, just like the lights' power cables are. For my pool temp sensor, I had to use direct burial wire, a few inches deep, because I wasn't going to destroy my mature landscaping for this experiment. And it works just fine, but the next owner of this house will likely discover my handiwork by shoveling through it, so that could have been installed better I suppose.

So, sure, I wish I had more sensors, and a more robust (yet stealthier) installation, but it's the best I could do for a retrofit.

Keep us posted with what you come up with!

 

[plnewb]

but a pool should be built with some sort of sensor array

What are your thoughts about using the (area around) light fixture that most (every?) pool has to house this array?

I am loving what you propose - proprietary systems exist precisely to do what you propose - but proprietary systems don't exist in my universe.

 

[Dirk]

What are your thoughts about using the (area around) light fixture that most (every?) pool has to house this array?

I think it should work fine, most of the time. My original plan was to use the light niche. And as I mentioned, it was a backup plan if running the sensor through the auto-filler equalizer tube didn't work.

I decided against the light niche for three reasons:
- I had the equalizer tube available and that was just a simpler route
- I was concerned that the light would affect the temperature reading
- Running low voltage wiring through the same conduit as high-voltage wires is usual against some sort of code and generally not a good idea

I pictured the sensor wire traveling through the same conduit that feeds my light, then through the niche and out its bottom to the sensor, dangling some number of inches below the light's trim ring. My light gets quite warm, possibly enough to affect the sensor's reading. Originally, my alternate was just housing the sensor inside the equalizer tube, where it wouldn't be seen at all. I thought maybe the water from the pool would influence the water in the equalizer tube enough to get a reliable reading. I just hadn't considered how the temperature of the water coming out of the auto-fill valve would affect the sensor. Duh.

I just realized that's not what you asked. The ideal location of this fantasy sensor niche would be determined by the pool shape, which may or may not be where a light would go. Just like how we advise where to take your water testing sample, the niche should be mid pool, between shallow and deep end, 18-24" deep, and as far away from any returns and skimmers as possible. And the conduit that feeds it should be independent, without any other devices' wires running through it. It would probably need to be bonded, too.

If you're asking about retrofitting such a thing, you'd want to satisfy all of the above, plus figure out how to get wires to it. I suppose something like that could be addressed during a surface refinish, if you could access the back of the shell well enough to drill a hole through it. When my pool was refinished, the crew removed my main drains and added eyeballs to my returns, so a lot is possible during a refinish.

The other way to wire it would be to have, or convert to, a low voltage lighting system, and then it would be OK to use an existing light niche conduit for the sensor wires. Sharing that conduit with more than one cable might pose a challenge, as fishing even just the one light cable can be hard to do. But adding a second, small cable is probably possible. And then getting the wire from the light niche to some sort of sensor pod would look much like my setup does, which I suspect most people would not like much. And this pod would have to be far enough away from the light to avoid the warmth of the light affecting the temperature reading. I'm not sure what that distance would be.

Since a pool light is not on all that much, and only at night, when a temperature reading would arguably be of less value, the problem, if any, of the light affecting the temperature reading could be considered acceptable.

Some light niches, like mine, are full of pool water. The temp sensor could be housed inside a light niche. The heat from the light would definitely affect it there, but as I mentioned, maybe that would be acceptable. WIth the light off, the water temp inside the light niche should be pretty close to the water out in the pool, but it's bound to be affected somewhat. I didn't experiment with that.

 

[JamesW]

I would place temperature sensors at increments of 1 foot depth from the top to the bottom to see the temperature profile and if there is any stratification of temperature.

If there is, the algorithm can figure out the actual temperature of the water if it was well mixed.

It can also turn the pump on to mix until the water temperature has a coefficient of homogeneity of 1.0 +/- 0.001.

 

[plnewb]

Since a pool light is not on all that much, and only at night, when a temperature reading would arguably be of less value, the problem, if any, of the light affecting the temperature reading could be considered acceptable.

Some light niches, like mine, are full of pool water. The temp sensor could be housed inside a light niche. The heat from the light would definitely affect it there, but as I mentioned, maybe that would be acceptable. WIth the light off, the water temp inside the light niche should be pretty close to the water out in the pool, but it's bound to be affected somewhat

Precisely, but I did make a strong assumption, that you clarified to not always hold:

Just like how we advise where to take your water testing sample, the niche should be mid pool, between shallow and deep end, 18-24" deep

This is a design detail thats literally frozen during the pool construction, so unless I designed the pool or someone reasonable did (and most pools are designed for looks than functionality) this would not be satisified

I know how you hate battery operated wifi controlled devices, and there's tremendous pain and suffering behind that realization, but I must emphasize that those experiences are due to poor and substandard implementations of both battery and wifi - that both these issues can be adequately addressed

The challenge is addressing them at a price point that can return enough profit while keeping a pricepoint that most pool customers don't balk at.

But we don't care about profit in a DIY forum - if our battery operated wifi controlled devices work reliably and keeps work reliably with parts that are quick and cheap to buy, we are happy

 

[plnewb]

temperature sensors

Do you have part recommendations James?

 

[plnewb]

The other way to wire it would be to have, or convert to, a low voltage lighting system, and then it would be OK to use an existing light niche conduit for the sensor wires. Sharing that conduit with more than one cable might pose a challenge, as fishing even just the one light cable can be hard to do. But adding a second, small cable is probably possible

There wouldn't be a need to run a second wire - the same pair that powers the light would suffice using multiplexing. This is a solved problem.

 

[JamesW]

Do you have part recommendations James?

I use a robot to go around the pool checking the temperature and chemistry at different depths and looking for debris and algae.

This is the A.I Pool Supervisor Robot and it watches over the pool and reports on all pool activity and it supervises the cleaning robots to make sure that they do a good job.

There are 7 cleaning robots and sometimes the Pool Supervisor Robot makes them fight like on Battlebots.

 

[Dirk]

There wouldn't be a need to run a second wire - the same pair that powers the light would suffice using multiplexing. This is a solved problem. On rockets there's a severe weight constraint, so NASA has developed techniques to be efficient and taught those to their staff

That's how my home automation system can piggyback my existing Romex wiring. Though interference then becomes a problem. My HA manufacturer tried to address this with filters, but eventually redesigned to add a second path using RF. Would a pool's LED powerline affect a sensor array?

The problem with multiplexing would be that the light and sensor array would have to come from the same manufacturer, which in effect would further limit the number of buyers, making the likelihood of such a combo coming to market near zero. There just wouldn't be enough profit in it. The bigger problem is adoption. Some large percentage of pool builders are still building the same pool they learned how to build 30 years ago (or how their mentor built pools 50 years ago). They just don't want to adopt new tech. Half of them still claim an SWG is bad for your plumbing. They still insist on main drains (which lost their usefulness decades ago). Etc. Convincing a pool builder to install a "new fangled" sensor array would be next to impossible. Most of them don't even know how to test pool water properly, or provide a proper chemical startup process. This is just not in their periphery.

So no one is going to develop this thing. Not in my lifetime, anyway. Which is why I had to butcher my pool myself!

It does make @JamesW chuckle, so the project was not without added benefit!

 

[Dirk]

There are 7 cleaning robots and sometimes the Pool Supervisor Robot makes them fight like on Battlebots.

Battlebots! Underwater?! IN MY POOL!!??!! Where can I get these? I will pay any amount!!

 

[JamesW]

I have them custom made by Boston Dynamics.

They are only $250,000.00 each, so a bargain, really.

 

[plnewb]

Would the LED powerline affect a sensor array?

Definitely not if the light is off when the sensor array is being read, which is almost always the case unless you want readings at night or on an overcast day - what are the chances that happen to you?

Remember - we don't have to build a product that works for everyone - a strong constraint a commercial manufacturer has to worry about - we are doing this for ourselves - you and I.

The problem with multiplexing would be that the light and sensor array would have to come from the same manufacturer,

Correct but in the DIY case, the manufacturer is you

I urge you to imagine a waterproof sensor attached to the robot or vacuum head that cleans your pool. James has a sophisticated (and expensive) robot that has that built in but you can put your own hand built waterproof sensor attached to the robot or vacuum head and get the exact same functionality.

We can even rag on James for being sophisticated when his manufacturer obsoletes the product and he can't buy parts or replacements, but we can, because we built things using commodity, easy to source and buy parts

 

[Dirk]

I have them custom made by Boston Dynamics.

They are only $250,000.00 each, so a bargain, really.

Thank you. I just placed my order. They'll be here in 6 to 8 months. Worth it. But I could only afford two. BD only offers a five minute warranty. That's a red flag and seems a little iffy. We'll see.