POU Hot Water Source (beer brewing)

Not open for further replies.


Advice, please, general and specific. I want to use a tap (cold) water source and a siphon from a boiling stock pot to a mixing valve to source me up to 8 gallons of water at an exact (like, +/-2 F) desired temperature (from ~150F to ~180F). Point Of Use (POU) hot water heaters abound but not so much for these temperatures and besides, I want it small enough to put away with the brewing equipment. And besides I wanna build it myself.

Seems like this project would start with a Teensy & user interface, two solenoid valves, a cold water under-pressure source hose, a hot water siphon hose coming from a pot boiling on the stove, and a temperature sensor dangling below the output hose. Operation would be to set a gallon container below the outlet with the temperature sensor in it, set the temperature, hold a button down until the container is full of the correct temperature water.

General advice:
1. Any obvious problems with the above (for example, I might have a problem in the boiler, with liberated absorbed gas collecting at the top of the hot water siphon tube preventing siphon.) ?
2. There's a TON of homebrewing toys out there and I haven't found this anywhere. There're two tons of homebrewing arduino projects out there and I haven't seen a POU hot liquor tank. This makes me nervous; again, what am I missing?

Specific advice:
I need a valve that'll take ~boiling water. I figure a clothes washer dual supply solenoid valve really ought to withstand boiling temperatures so I took one apart (the electrical parts) and boiled it. Seems to stay perfectly intact, yay. A plus would be one that is rated for drinking water, but I'll test this one for odors/taste and I'll pretend that flushing it well will reduce any bad plastic stuff sufficiently. Bad idea? Thoughts on more suitable, extremely cheap, boiling-rated valves?
Hi Frank. Well, I maybe *could*, but the first beer I'm aiming for is Urquell (which being a pilsner (THE Pilsner) will use my conscripted old fridge kegerator cum lagerator, making it a double Teensy-beer). Depending where you are, Pilsen is a lot closer to you and if you were me you'd be headed over there every month or so anyway for R&R (Research and Refortification). I'll lechyaknow how it goes.

Noone had actual good ideas wrt the POU, I guess. I ran into a problem with the washing machine valves--They utilize the non-zero line pressure to help them operate against the built-in spring. A few feet of pressure from stovetop to near floor gives ~1/2 psi/ft and that's not enough to run those valves. Haven't picked it up from there. Maybe a metal ball valve, but I haven't worked on that yet. ___Marty
Syphoning boiling water sounds like something to avoid - the risk of vapor locks, and more importantly
super-heated water explosively boiling in the tube would have to be addressed - you might want to consider
removing the heat source beforehand, and using anti-bumping granules.

For temperature control the only really fast responding sensor I know of are the thin flex-pcb mounted thermistors
which aren't necessarily designed for immersion, but will respond in a second or so. Oddly enough I have worked
on a PID loop for temperature control of a hot tap using proportional stepper valves before - latency in temperature
measurement is to be avoided if possible. With bang-bang control using solenoids you might want to consider slow PWM
as the control input.

There are other approaches like measuring the two temperatures, and using a height or weight sensor to control filling
the output container sequentially from cold and hot.
Let's look at the energy balance. Assume you want 8 gallons of water at 160F. Your cold water tap yields water at 60F. Boiling water is 212F.

The boiling water is 52 degrees away from the desired temperature. The cold water is 100 degrees away from the desired temperature. A SWAG tells me you'll need about 2.6 gallons of tap water and 5.4 gallons of boiling water. Higher desired temperatures mean more boiling water.

These desired temperatures are in the range where just a little bit in the wrong place will produce significant scalding injuries. I would try to minimize the amount of hot water in storage as much as possible for safety's sake.

You might be better off starting with hot water from your water heater. Hot water is usually about 120 to 125F, so you only need to heat it by 40 to 60 degrees instead of 80 to 100 degrees.

For a large instrument calibration bath, we would heat the water by running it through a copper coil immersed in boiling water. You could do the same thing with a heat exchanger coil in your pot of boiling water. You control the temperature by adjusting the flow rate through the heat exchanger. You could do that with a pump or valve that can handle hot water heater (125F) temperatures and use the Teensy to control the flow rate in response to the output temperature. With this method, the boiling water never leaves the pot--it just has to raise the temperature in the flowing water by 35-55 degrees. Of course, the 160F water coming out of your hose is still something to treat with caution!

The for this to work, you need to know the desired flow rate of the heated water and the heat exchanger energy transfer rate.

For MarkT: There are small glass-encapsulated thermistors with response times down to about 25 milliseconds. I've used them for oceanographic sensors. Unfortunately, if you want fast response, you can't surround the active part with a lot of protective material, so they are pretty fragile.
For MarkT: There are small glass-encapsulated thermistors with response times down to about 25 milliseconds. I've used them for oceanographic sensors. Unfortunately, if you want fast response, you can't surround the active part with a lot of protective material, so they are pretty fragile.

Interesting, they must be really tiny to get that performance - do you have a part number?
Many years ago I did a student 'work-week' for a quartz crystal (oscillator kind) manufacturer. My first task was to measure the resistance of these tiny bead thermistors that were to be placed inside the crystal case. Well, wouldn't you know that the resistance kept changing due to self heating from the Wheatstone bridge current. This was their gnawing problem. I eventually solved it by using an cup oil (heatsink) to immerse the thermistor to get a stable reading. Felt good as a student to contribute to real world problems.

Interesting, they must be really tiny to get that performance - do you have a part number?

The one used on the oceanographic sensors was the FP07DA103N when I last worked there in 2018. That is a 10K Ohm at 20 deg C unit. It's not in stock at DigiKey, but the 8K Ohm version is available for $195.84 qty 1. That's not a typo! Not a sensor you buy just to play around with. The Time Constant (TC) in still air is 100mSec. The TC for a plunge into water is 7 milliseconds. We used them to measure the small temperature changes in seawater due to turbulence in the 5 to 40Hz band.

There are other glass bead thermistors for about $2 that have still air TCs of about 5 seconds. In flowing water the TC would probably be under 1/2 second. That is probably fast enough for a control loop that runs about once per second. With mechanical valves or pumps you probably can't go much faster.

For the OP's application, it is important to remember that the TC in flowing water will be about an order of magnitude lower that the TC in still air.

KOC62's comments about self heating are probably not a problem in flowing water for the OP's application, but on the oceanographic sensors we did have to make sure the bridge currents for the thermistor amplifier were kept small (by using a low voltage across the bridge.) We discovered that when we did a new circuit design with modern op amps. If we powered the bridge off the 5V supply, with the sensor in a glass beaker of water, we would get different readings in still water and when we stirred the water. We figured it out and fixed it before the system was deployed. Deploying oceanographic moorings isn't as expensive as putting instruments into orbit---but then we didn't have a NASA style budget, so we did a lot of testing before sending instruments to sea.
Ah, at that price I'm not surprized I haven't seen one - the thin plastic ones I have used (I think its basically sandwiched in
a bit of flat-flex pcb - not likely to last forever) These work fast enough epoxied into a bit of pipe for PID loop in a hot-water
temperature controlled mixer valve setup.
The nice thing about designing systems with a few really expensive sensors is that the funding agencies don't quibble about the cost of upgrading to an MPU with more memory or better peripheral systems. It's not just the cost of the parts, but the need to support talented engineers and technicians to build and calibrate a system that becomes a major budget item. The old baseball adage applies: "If it was easy, everybody would do it."
Not open for further replies.