Login
You're viewing the mstdn.social public feed.
  • Jul 9, 2026, 11:27 AM

    I'm going to buy either model H19122 or H19242 from this supplier aliexpress.com/item/1005012330

    H19122: 12V, 85W
    H19242: 24V, 150W

    as far as i can tell, both models come with a 3-phase inverter, meaning the power rating is a maximum and i can probably connect them directly to a protected Li-Ion pack (or even USB type-C trigger) then regulate compressor power with a PWM input or something. but i'll ask the vendor about this first

    💬 5🔄 0⭐ 6

Replies

  • Jul 9, 2026, 12:29 PM

    asked the vendor, they provided usable documentation, the motor speed/power can be set with a resistor (so I can probably just PWM it). the compressor must remain horizontal, so this will have to be assembled in a particular way

    here's a very crude diagram of what i want to build

    Image attached toot
    💬 1🔄 3⭐ 16
  • Jul 9, 2026, 12:31 PM

    both the air intake and air exhaust will be on the back; exhaust might go somewhat towards the top, but i don't want the back of my head baked in hot exhaust air

    the water tank is at the very bottom in case it leaks
    the battery pack is kept away from the hot air and compressor (i don't expect the baseplate to get hot)

    💬 3🔄 0⭐ 10
  • Jul 9, 2026, 12:35 PM

    this system will run on distilled water with algacide in it as i can't imagine wanting to go below zero (i think it's not a big deal if the heat exchanger has some spots that are below zero, ice slush can still move through it)

    it will need interlocks. i can think of two:

    • tilt switch (for the compressor orientation)
    • QC hose switch (to avoid pump overload)

    although the latter might make sense to replace with an overflow valve if i can find one that opens at the right pressure

    💬 4🔄 0⭐ 8
  • Jul 9, 2026, 12:38 PM

    i don't know how to pick a circulation pump. if i go for a positive displacement pump, i'll absolutely need that second interlock; if i go for a centrifugal pump, those can operate with the output plugged though i don't think they like it very much

    do any of you have experience here?

    💬 5🔄 0⭐ 3
  • Jul 9, 2026, 12:47 PM

    now that i think about it, "overflow valve" is a product that probably only exists in factorio. but what i can use instead is a flow restriction or a metering device. if the jacket is connected, there is a low-resistance path through it. if the jacket is disconnected, that path is broken and water is forced to flow through the restriction. i think that's good enough?

    Image attached toot
    💬 6🔄 0⭐ 4
  • 💬 1🔄 0⭐ 0
  • Jul 9, 2026, 11:40 AM

    @G_glop excellent question, actually. the main problem with being non-vertical with this type of compressor is that they eventually run out of oil. however this is such a small system that it might not matter? i'll ask the vendor

    💬 3🔄 0⭐ 1
  • 💬 0🔄 0⭐ 1
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 1
  • Jul 9, 2026, 6:51 PM

    @whitequark
    However if you shut the compressor down because it tilted, I think you have to lock it out for some magic length of time to prevent compressor stalls.

    What's a compressor stall? *shrug* No idea. How long does it need to reset? Dunno. What is the mechanism for this reset? Beats me.

    💬 1🔄 0⭐ 0
  • Jul 9, 2026, 6:57 PM

    @mdwyer a compressor stall is an event that occurs if you try to start a compressor with its outlet being at a high pressure (which is the case right after you turn it off). it doesn't have inertia yet so it cannot turn, and the rotor locks (different from mechanical and hydraulic lock). this is bad for the motor

    usually pressure equalizes after less than 5 minutes. so the way this is normally handled is a PTC thermistor in the starter winding circuit, once it cools down the motor starts again

    💬 1🔄 0⭐ 1
  • Jul 9, 2026, 8:58 PM

    @whitequark
    Ah! Today I learned! And I didn't have to Google it myself, he said in sheepish shame...

    Still, point holds: it would kind of suck to lose cooling for 5 minutes because you had to tie your shoes.

    Anyway, I'm excited about your project. I wanted to do something similar, but for sleeping. EightSleep has a mattress-based system, but somehow a wearable sounded more interesting. You can close-mount thermometers and add other sensors, too.

    💬 1🔄 0⭐ 0
  • Jul 10, 2026, 1:24 AM

    @mdwyer I don't think such short misalignment matters? issues only begin when the compressor runs out of lubricant. it can probably run for a minute or two even upside down

    💬 0🔄 0⭐ 0
  • Jul 9, 2026, 11:50 AM

    @whitequark

    I like that they're plumbed with a service port - but it seems an unnecessary addition, unless they expect you to pump it down, verify leak tightness, then charge with what I can only imagine is an obnoxiously tiny amount of refrigerant.

    Shame international shipping costs are so obscene these days - lots of scientific & medical surplus around these parts - chillers going for a pittance...

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 2🔄 1⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 1
  • Jul 9, 2026, 12:54 PM

    @whitequark hmm yeah that might be harder
    Also the pressures there might be quite low (as in the difference between working and jacket disconnected)
    So a thing that turns the pump off when the pressure rises to far might be better

    💬 1🔄 0⭐ 1
  • 💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Jul 9, 2026, 6:52 PM

    @whitequark @4censord Current sensor on the motor. If it’s bunged up the current will go too high; if it’s sprung a leak, too low; detection of a dead motor and motor on fire are bonus features you get for free.

    💬 0🔄 0⭐ 1
  • 💬 1🔄 0⭐ 0
  • Jul 9, 2026, 5:38 PM

    @4censord @whitequark

    Depending on the pressure differential a non-return valve can serve the same purpose. It will take some trial and error to find one that opens at the right pressure. Try aquarium supply stores.

    💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Jul 9, 2026, 12:54 PM

    @whitequark if it’s just water simply point a pressure relief valve at operator’s face and rely on pavlovian conditioning

    💬 0🔄 0⭐ 2
  • Jul 9, 2026, 1:17 PM

    @whitequark If it's just a restriction then there'll always be some flow, and there's some efficiency loss from entropic mixing of cold bypass and hot return.

    Idk if that really matters given the temperature rise through the jacket is quite small, but putting a relief valve on the bypass would avoid it.

    💬 0🔄 0⭐ 1
  • 💬 2🔄 0⭐ 1
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Jul 9, 2026, 11:22 PM

    @whitequark in .au they’re called “pressure over release valve” and found on tank storage style hot water systems to prevent the tank turning into a pressure vessel due to over-heating. Completely passive device, just works on pressure differential, and available in wide range of activation pressures.

    If you do this, you’ll need a different interlock to stop the pump from running if the reservoir runs dry (which you probably should add anyway)

    💬 1🔄 0⭐ 0
  • Jul 9, 2026, 11:32 PM

    @whitequark oh, and for quality of life, you want a pair of series stop valves with a quick-disconnect in between them between the jacket and the refrigeration system, and put the pair on both the supply and return lines. Close both valves, disconnect, almost zero coolant comes out.

    Jacket —— x || x —— system

    💬 0🔄 0⭐ 0
  • Jul 9, 2026, 12:54 PM

    @whitequark only pumps I know a little about for wearables are peristaltic and maybe that would have insufficient flow rates.

    💬 0🔄 0⭐ 0
  • Jul 9, 2026, 1:56 PM

    @whitequark I have a couple of nice little 24V centrifugal pumps with BLDC motors with a little speed controller in a box with a pot to adjust it. A Chinese friend ordered them, about 12 years ago. It won't self-prime, so it should be located below the water level (or below any possible air bubbles, even if there isn't supposed to be air).

    They are about fist-sized, and waaay too powerful for what you want, I think. They make a fountain comparable to a garden hose at full power.

    💬 0🔄 0⭐ 1
  • Jul 9, 2026, 2:03 PM

    @whitequark In open-center hydraulics, they solve this by using 3-way valves for all the actuators, with the normal positions going to the next actuator and then the last actuator goes back to the tank with no restriction or further valves. With no actuators engaged, no pressure gets built, and a fixed-displacement pump will run fine into zero load (and the prime mover will basically idle). Not sure if they make "3-way quick connects" or something though

    💬 1🔄 0⭐ 1
  • Jul 9, 2026, 2:12 PM

    @whitequark If I were making this I'd be tempted to use metal quick-connects, add a bonding wire between the jacket-side quick-connects, and then wire the pack-side connects to the power supply and enable input such that plugging in the quick-connects starts the pump after a few second's delay. You'd want a plastic bracket for the pack-side connects to ensure they can't complete the circuit by touching each other

    💬 1🔄 0⭐ 0
  • Jul 9, 2026, 2:20 PM

    @whitequark alternately a pressure switch with an activation point below the maximum pressure of your pump would work (I think the ones sold for reverse osmosis units would be a good match for your system). If your pump doesn't have check valves you could add one before the switch to keep pump backflow from causing constant short-cycling

    💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 1
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Jul 9, 2026, 12:46 PM

    @whitequark something to be aware of is that at least with larger compressors not only do they break when running not upright, after they tip it can take several hours for the oil to settle in its proper place and if you turn them on too early they still break

    💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Jul 9, 2026, 12:54 PM

    @whitequark My understanding is that compressors should stay upright because gravity is the only thing keeping oil out of the refrigerant lines and in the motor. With that in mind, I'd worry that the tilt switch might be too sensitive for moving around, but also I think it'd need some sort of (timed?) lockout to give the oil time to settle.

    💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 1
  • Jul 10, 2026, 8:49 AM

    @whitequark
    I doubt that you want the water through the jacket to be much below 25C, so you might not have any water even in the heat exchanger below 0C. But I'm only guessing. If it ends up that there is, adding propylene glycol or glycerin as antifreeze might extend the life of the pump.
    Direct contact with 25C water would chill you a lot, enough to cause hypothermia in an hour, but I have no idea with the jacket, which obviously has non-trivial thermal resistance.

    💬 1🔄 0⭐ 0
  • Jul 10, 2026, 8:51 AM

    @brouhaha it has rather thin plastic tubes, thermal resistance across those can't be low. and the surface area coverage is very low, maybe 5% for the upper torso, even that could be optimistic

    💬 1🔄 1⭐ 2
  • 💬 1🔄 1⭐ 1
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 1
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 2🔄 0⭐ 0
  • 💬 1🔄 0⭐ 1
  • 💬 2🔄 0⭐ 2
  • Jul 9, 2026, 1:37 PM

    @whitequark lol yea seems sketchy but it sounds its less regulated, I also see a lot of listings in Germany. Which is odd because it seems to be covered under EU F-Gas regulation.
    Normally regulated refrigerants need to be sold in specialized shops to license holders only.

    💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0