[time-nuts] Temp/Humidity control systems?
kb8tq at n1k.org
Thu Oct 27 10:37:06 EDT 2016
Ok, take this with a bit of caution ….
Your “time cave” does not have a specific spec on temperature or on humidity.
You get to pick a number for either one. Anything in the “non condensing” (let’s
call it < 80%) range for humidity is likely ok. Temperature up to 40C is probably
ok for any gear that I can think of. As long as you *never* go in and out of the
cave, comfort in the cave is a non-issue.
If I load up the cave with gear that runs up the electric bill, the cave will self
heat to some degree. How much it self heats is unclear. The number will be highly
dependent on your situation. Based on things like racks full of gear in out of the
way places, it’s a good bet that it will be 10 to 20 C above the temperature in the
rest of the basement.
Assuming the humidity in the basement is under control (if not, fix that), all I need
to do for humidity in the closet is to exchange air with the basement. Since the
volume isn’t very large and (hopefully) the water flow is modest …. a small fan
should take care of that.
If the closet is bound by the outside (buried) wall and the basement, both should be well
controlled temperatures. If I plug everything in and let it run, the heat rise in August should
be the heat rise in February. Let’s say that’s a 20C rise. The basement is at 20 C and
the closet is at 40C. Maybe its more, If it’s less, you need to buy more stuff:)
All I need to do is knock the 40C down to 35C with a small amount of “cooling” to
keep things under control. My “humidity” fan might do that. If not, a very simple
water based heat exchanger with a controlled fan will do the job. You need some sort
of fan(s) in the closet anyway. Without them you will never get the gradients under
control. A pump, 10’ of tubing and two heat exchangers are not a lot of money. There
also isn’t a lot to break if you do it right.
Why do it this way? No compressors to mess up the local power line and break
every X years (the fridge in the kitchen died Sunday ….. they do break). No super
cold surfaces to mess up humidity. No crazy heat flows to create clod drafts and
fast transients in the closet temperature.
The obvious downside is that you *do* go in the closet. When you do, everything
goes a bit off. My guess is that you are actually better off letting it recover slowly
than you are trying to move it back in under 10 minutes …. who knows. One solution
to that would be to sub-divide the closet. The gear to do this is cheap.
> On Oct 27, 2016, at 10:06 AM, jimlux <jimlux at earthlink.net> wrote:
> On 10/27/16 6:30 AM, Poul-Henning Kamp wrote:
>> In message <C1109A57B22F4DAB86B0BB172981F264 at Alta>, "David J Taylor" writes:
>>> You can buy the smallest "window" airconditioner and "plumb" it to your
>>> chamber (I used dryer vent hose, cardboard, and lots of duct tape)
>>> Attached is a plot temperature and RH of an insulated box about 1.2
>>> meter wide, 2 meters tall and 60 cm deep, filled with 100 or so 750 ml
>>> bottles of liquid.
>> You can do *much* better with an old fridge and a small waterpump to
>> circulate water in the cooling loop.
> I tried that first...
> Advantage is that the mass of the water (serving as a thermal transfer medium) is much greater than that of air. You wind up with kilos of water at a relatively constant temperature.
> 1) It leaks
> 2) It grows stuff (even with additives to prevent it)
> 3) you've increased the number of thermal transfers: refrigerator coils to air to water to air to contents of box. Both of the air:water transfers are not particularly efficient in a "cobbled together in the garage" sort of scenario.
> I think your suggestion of just adding (solid) mass to the system is better - the advice to people with wine cellars is to fill empty slots with bottles filled with wine (actually, the advice is to buy more wine, so your cellar doesn't have any empty slots.. but if that's not possible, fill the empties with water, recork and stow them)
> The challenge with a refrigerator as chiller is that getting decent coupling from the cooler coils to the water is tough: you're pretty much restricted to air as the transfer medium. In a "real chiller" they put the evaporator coils in the water so there's good thermal contact. Refrigerators/freezers aren't made for this - I tried 3 different approaches of varying complexity:
> 1) Put a 5 gallon plastic bucket of water in the refrigerator - the bucket of water does get cold, but--- it also evaporates inside the refrigerator, and the water condenses on the coils, freezes, and then eventually is lost to the air when the unit is defrosted.
> 2) Put an array of copper tubing in close contact with the cold plate in the freezer, weighted down by bricks to make close contact - well, let's just say I found all sorts of interesting galvanic reactions can occur, even at low temperatures - the other problem is that if the circulation rate slows, the water in the loop can freeze, and once it starts to freeze, it has positive feedback - the flow rate slows even more, and pretty quickly, you have tubes full of frozen coolant. - it is a good thing I was doing this in the garage.
> 3) trying to make a cold plate by using two sheets of aluminum, some aluminum spacers, plenty of silicone, and some hose barbs is a lot of work, and doesn't seem to work much better.
> I think one problem is that the refrigerator/freezer control system is designed to work off two sensors: one is a air temperature sensor and the other is a sensor on the actual evaporator unit (basically, if it gets too cold, it shuts off the compressor to prevent low pressure damage). The "design point" for all of this is also probably not optimum for moving heat out of your equipment closet.
> And that's just the challenge on "improvised water chiller"
> Then you have the other "water to air heat exchanger".. serpentine tubing would seem to be the best way, but it turns out that this is non-trivial to design so that you get even flow rates in multiple loops, if you have multiple paths. And, arranging the tubing effectively is hard. There's also all the fabrication/leakage/hose connection issues. I tried making serpentines out of copper and aluminum tubing that would be part of the shelf on which the bottles are piled. That cools the bottom bottles nicely, but the thermal transfer among the bottles is slow.
> The best approach to "get cold to all bottles" (or, more correctly, take heat from warm bottles" is to have a fan to circulate air among the stacked and racked bottles. Well, once you are rigging up a fan to push air through cold tubing (a re-purposed car heating core - more fabrication of adapters from one tubing size to another - I had a big box of hose clamps, between size adapters, and pieces of hose of all sizes). Remember that this is all wet at one point or another, either from leaks or condensation, so stuff corrodes, rusts, etc.
> Yep - a $99 window airconditioner bought on sale (about this time of year is good, in the Northern Hemisphere) worked just fine. Plumbing air is a lot easier than plumbing water or glycol coolant. You won't get down to <10C because window air conditioners aren't refrigerators - the choice of refrigerant and internal components and set point range isn't compatible with that (and putting a small heater on the AC's temp sensor, which is in the return air flow to the evaporator did not allow me to "bias" the set point)
> However, for a time-cave - I think it would work great - cheap airconditioner, large thermal mass buffer, well stirred air. Concrete blocks with holes and have the air blow through that would make a fine low pass filter - keeping the humidity "reasonable" is probably straightforward - most of time-nuts stuff isn't humidity sensitive - as long as it's "non-condensing". You'd just want to watch and make sure that you don't have cold air blowing on a metal surface intermittently - I worked in a screen room (solid walls) that had the AC blowing on the walls.. when the AC cycled off, the walls instantly started to sweat as our respiratory moisture condensed on what was now the coldest thing in the room.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts