Long term sensor reliability

[Dippy]

I am just about to start a PICAXE solar (heating) controller for water / pumps.
This is just a query on the temerature sensing.
Assuming:
A) The temp sensor will be encapsulated.
B) Any maths for temp conversion is sorted.
C) Everything is sorted (hopefully).

For cheapness vs longevity and trust/reliability, would you go for a DS18xx or Thermistor?

I only ask this as Rapid sell some neat pipe-clip-on thermistors.

PS. Has anyone come across reasonably priced flow sensors? Paddle or turbine magnetic pref. I'e been hunting high and low and can only find sutiable jobbies in excess of £80.
Note: internal optical type not really suitable as the water will have anti-corrosion (e.g. Fernox) and anti-freeze which tends to be a bit opaque.
[Needs to be connectable to standard copper 'central heating' pipe with/without adaptors and have a pulse o/p]

PPS. I'm in UK.
PPPS. Please don't spend ages searching, it was just if anyone knows of a source from memory. Just a casual enquiry.

 

[BeanieBots]

At £80, you've already found a MUCH cheaper flow sensor than any of the ones I know. I've found the ones which work like a water wheel to be the most reliable long term. Magnet in one or more paddles with pickup coil for sensing.

As for your temp sensor, I would go for DS18B20. I've experienced noticeable drift with thermisters over a period of several years. Also, corrosion on connectors can compromise readings. With DS18B20, the only problem you are likely to encounter is water ingress. Total sealing in epoxy cures that. I've had one immersed under 3' water for nearly 10years now and is still within 0.06C of it's original reading at 25C. It locked up a couple of times caused by emi from nearby flourescent tubes but it returned 0C rather than an incorrect value.
(requied power cycle to unlock)

 

[Dippy]

Thanks BB.
Re: DS. That test sounds good enough for me to make my decision. Good points re drift.
Appreciated.

Re: Flow sensor.
The only two 'low cost' candidates have come from GEMS RotorFlow.
I'm awaiting a reply from icenta. That looks good stuff, but I think I'll have to sit down when they tell me the price.

 

[Michael 2727]

I'd go with a DS18B20 also, just seal the wires well (for either version), Telcos'
use silicon grease on cables here (messy stuff) then epoxy/silicon the whole thing up.

Can't help with the flow sensor, they all seem to cost an Arm & a Leg, try irrigation
or farm suppliers or even landscapers for sprinkler system.
Don't spose they have much call for them in the UK ?

Does the flow sensor have to be accurate (L/min) or just sense, it is or is'nt folwing.

 

[evanh]

I've had horrible results with silicone grease. It always hydrates after a while. We eventually stopped using it in everything.

 

[Dippy]

Michael: I'm trying to get an actual water flow figure in lires/minute. I don't need fantastic accuracy. +/-0.001% should do

I'll order some DS18B20 shortly. I was going to insert them into either:-
1. Some 6mm Al or brass tube (thin walled model stuff).
..or..
2. A Battery/cable terminal thing. Looks like tube, one end squashed with an 'ole in it.

And filled with 2-pack something. Epoxy/polyurethane, whatever comes to hand. Time lag isn't too impotent.

Man from icenta says they already supply a company with Flowsensors for Solar Panels... hasn't told me price yet. I am bracing myself.

 

[BeanieBots]

I sealed my (actually two) DS18B20 in 1" of 5/8" stainless tubing with Araldite. The cable is tri-axe protected by 5' of silicone tubing which was a tight push fit into the stainless tube.

I'm about to build another one but was thinking of using polymorph instead of Araldite. Just drop into boiling water to dismantle should the need ever arise.

PS. It doesn't suffer from impotence!

 

[hippy]

DS18B20 every time, even if it does fail or need replacing more often or is more expensive.

Once replaced it will work, that's it. With a thermistor you'll likely have to re-calibrate and you'll have forgotten how to by then. Days later when you find the software it won't compile. When it does it won't do what the documentation says it will. Once it is calibrated you'll go through the same cycle in updating the control software. After a week or two of frustration you'll push the whole lot over a cliff edge and start redesigning from scratch.

Or something like that.

 

[Dippy]

Thanks all.

Well that, as they say, just about wraps it up for thermistors then. Saved me Steinharting too.

I can't use polymorph as my thing will be squeezed up against a hotwater pipe.

Polyurethane appeals as it has a little bit of give and still sticks well. But as I hae some Araldite kicking about that'll do.

PS. Are there different Polymorphs? I just looked at Rapid. Mouldable "between 30°C and 62°C." and "biodegradable" - so that doesn't sound too brilliant long-term on a hot water pipe.

 

[BeanieBots]

Are there different Polymorphs? I just looked at Rapid. Mouldable "between 30°C and 62°C." and "biodegradable" - so that doesn't sound too brilliant long-term on a hot water pipe.

Don't know but I would think so.
The stuff I have came in the BBC robot packs. I'd not describe it as mouldable until close to 60C. Almost liquid at 70C. Didn't know about the "biodegradable" bit though. Maybe not such a good idea especially as mine will get a fair amount of UV exposure!

ADDITION:-
I've just checked a 2" sphere I made from it about 5 years ago.
It's been sat on a windowsill in the spare room which gets the evening sun and gets up to temperatures of +40C in the summer. No signs of any degredation. Not even a dimple at the point of contact so I guess mine MUST be different to the Rapid version.

Late addition:-
I guess this answers the polymorph question.
http://www.mutr.co.uk/product_info.php?cPath=418_464&products_id=551