Where can I find a very accurate water heater?

October 11, 2005 8:33 AM

Where can I find a very accurate water heater?

I'm working on getting an old test system from 1979 up and running and one of the requirements is that it needs a water heater that can keep a water supply at approximately 150 deg F and pressure of 40 psig. I can install a reducing valve to keep the pressure down to what I want.

The main thing is I need to be able to set it to a particular temperature and get little variance from it. Any recommendations? I'm thinking 20 to 50 gallons should be sufficient. I wish I knew more, but I really don't.

The water heater that is suppose to go with the testing system is no longer available. If you want any more information check out the 1241 Adiabatic Oxygen Bomb Calorimeter over on Parr's web site.

Thanks

posted by nickerbocker to technology (13 comments total)

You may want to look into adding a PID (TLA for: Proportional, Integrated, Derivative) controller to the heater as a way of controlling the temp.

TTI Global is one supplier.
posted by Icky at 9:03 AM on October 11, 2005

I also thought of a PID when I read this. I've installed one in my espresso machine (using parts from TTI) and I know there are lots of resources on the interweb describing how to hook one up.
posted by exogenous at 9:21 AM on October 11, 2005

MSC carries temperature controllers, some with build-in PID. Search for "temperature controller" (not thermostat).
posted by kc8nod at 10:07 AM on October 11, 2005

Reading the manual for the 1241 it appears you just need consistently close to 150F and the unit will make the micro adjustments by mixing hot and cold water. The easiest/cheapest way to do this may be with a temperature controlled mixing valve. They don't require any kind of power just access to hot and cold water and a few minutes adjustment with a thermometer. If you don't want to plumb them in direct any plumber could hook you up with adaptors to screw onto a hose bib. You'll need to spec a commercial unit or one for radiant heat as the residential types usually top out at 120-125F.
posted by Mitheral at 11:02 AM on October 11, 2005

So far this has all been very useful. Thank you for your replies.

Are there any water heating units with built in PID controllers out there? I'm not coming up with anything through google.

I'm not sure if modifying the temperature control unit (most likely a thermostat) in a water heater is the route I should take. At the same time, I can't really have more then a 5 degree variance in the water from day to day. The more consistent the water temperature, the more consistent the testing results will be.

Surely there are water heating systems available that provide accurate consistent temperatures. Seems like that would be some what in demand for a number of applications.
posted by nickerbocker at 11:08 AM on October 11, 2005

Mixing valves will be within 1 degree all day, everyday. Assuming both input temperatures don't fall below or rise above the set point. IE: if you feed it 130F ot water it won't be able to mix to the set point.

All the critical temperature stuff I've ever dealt with (photography, radiant heat, solar water heating, scald protection) used mixing valves.
posted by Mitheral at 11:24 AM on October 11, 2005

Laboratory water baths hold temperature very accurately. If you can live with putting your materials to be tested in a bath that hold temperature well, instead of having a supply of water through a heater, you may be able to save some money.

These heaters are used in fancy restaurants that cook food sous vide, i.e., sealed in plastic and then immersed in hot water at precise temperatures for long periods. See this thread
posted by KRS at 11:42 AM on October 11, 2005

Those temperature controlled mixing valves look great. You should probably give a hint about the price - here is one on ebay, but it doesn't go hot enough.

What about a tankless water heater (these are electric)? The ones on that page still don't go hot enough, but it might be worth pursuing. There might even be some point in adding a control loop around a tankless model... I wouldn't want to try to control a complete hot water tank unless outgoing flow rate was very regular.
posted by Chuckles at 11:47 AM on October 11, 2005

You should probably give a hint about the price - here is one on ebay, but it doesn't go hot enough.

Money is no object.

I think the mixer valve looks like the best and easiest solution, if I can just find a water heater that gets hot enough I should be good to go.

Hooking up a PID to an expresso machine sounds like fun.

Thank you for all your replies. As usual Ask.Metafilter proved to be the best resource for information.
posted by nickerbocker at 12:20 PM on October 11, 2005

Isn't "PID" a name for a control systems algorithm? How can you have "a PID"?
posted by phrontist at 2:33 PM on October 11, 2005

Apparently 'a PID' is a turnkey box that applies a PID algorithm with minimum implementation effort. It does the ADC on the input, applies the PID based on programed constants, and performs the DAC to drive the output. Many/most seem to have adaptive algorithms built in which will set the constants for you, which would help the uninitiated use them with even less effort.
posted by Chuckles at 3:21 PM on October 11, 2005

I guess it is worth adding... Presumably most of these products assume a single input single output system, although the concept might extend a little further. Also, for any stable system - you just want to improve performance - an adaptively tuned PID will work pretty well.
posted by Chuckles at 3:26 PM on October 11, 2005

Apparently 'a PID' is a turnkey box that applies a PID algorithm with minimum implementation effort.

I actually studied PID algorithms back in my Control Systems class in college. Wrote some in Matlab before too. Still, this is all application level, so a turnkey box is more of what I'm looking for ;).
posted by nickerbocker at 8:58 AM on October 12, 2005