So, if you want refrigeration, that evap temp, needs to be lower than the refrigeration range.

Which is lower than AC range. AC with a 40 degree evap, the air never reaches 40 degrees.

Using an AC for refrigeration, is like using a boiler, to melt steel. An AC just doesn't get cold enough, like the boiler never gets hot enough.

Unless, >40 degree cold source, is the coldest temp you need. Like an AC.

The FDA recommends keeping your refrigerator at or below 40° F (4° C). AC can not do that.

And there are other things that must be taken into account, like the super heating needed at the end of evap......before going into compressor.

Lots of engineering goes into controlled cooling.

Both you and Ylli are saying that the AC can not produce <40 deg. F in a walk-in. My neighbor was saying that he wanted a walk-in that cooled to a temperature only a few degrees above freezing. To boot, he has proposed cooing with a very small AC about 18 inches long.

What would you say about the AC sizing chart that I am providing a link to (below). This is a chart for a walk-in cooled by an AC. One temperature range of the chart is only a few degrees about 32 deg. F.

https://www.storeitcold.com/build-it/ac-selection/

 

Worst case there may be a need to replace some of the freon with R134. Of course the cooler that uses R134 also froze the evaporator into a block of ice. So it will be interesting.

 

I did not say...............that you can not modify and operate an AC to produce temp > 40.

I said it was designed for 40. It's tuned. It can be mis-tuned.

I am saying that mis-tuning it can have consequences. It's like a circuit. Changing one parameter, changes another.

It could decrease compressor life. And efficiency.

If the units works as expected and wanted, I will be happy for him.

Please let us know how it does turn out.

 

AC with a 40 degree evap, the air never reaches 40 degrees.

While I concur with your implied assertion that 'metering' (i.e. the ID of the capillary lines -or- expansion valve parameters --as equipped--) applicable to A/C units does not lend itself to efficient low temp operation - I'm bound to say that I've had very good experience with R22 'window units' down to cooled-space air temps of 37°F --- The 'trick', such that it is, lies in implementation of an automated 'frost-prevention' scheme (e.g. temporary de-energization of the compressor upon and during detection of slightly reduced airflow [which being indicative of incipient 'frosting' of the evaporator])...

Very best regards
HP

 

One way to get an evaporator to operate at a temperature lower than designed is to bleed out a bit of coolant. That will reduce the low side pressure and therefor the boiling point of the refrigerant. Reduces efficiency too.

 

One way to get an evaporator to operate at a temperature lower than designed is to bleed out a bit of coolant. That will reduce the low side pressure and therefor the boiling point of the refrigerant. Reduces efficiency too.

Indeed! It is for that reason that a slight 'undercharge' results in evaporator 'frost-back' under 'design-standard' operating conditions

Best regards
HP

 

Thanks to all for your comments.

If the room temperature sensor of an air conditioner has a two-wire connection, can I safely assume that the sensor is a thermistor? Some time in the next few weeks I will be helping my neighbor to select an AC. Unfortunately he is saying that he wants to buy a used one.

Regards,
Pete

 

Thanks to all for your comments.

If the room temperature sensor of an air conditioner has a two-wire connection, can I safely assume that the sensor is a thermistor? Some time in the next few weeks I will be helping my neighbor to select an AC. Unfortunately he is saying that he wants to buy a used one.

Regards,
Pete

I'd expect the room temperature sensor to be a simple bi-metallic switch.

A used AC might be a blessing in disguise - Old style coolant (R22) may get the temperature down a bit lower than the modern R410A, and if the charge is a bit low that will further reduce the lowest achievable temperature.

 

The 'trick', such that it is, lies in implementation of an automated 'frost-prevention' scheme (e.g. temporary de-energization of the compressor upon and during detection of slightly reduced airflow [which being indicative of incipient 'frosting' of the evaporator])...

Very best regards
HP

The pricey device that my neighbor had discovered on the web has a temperature sensor in contact with the fins of the AC. Presumably this device has circuitry to shut down the AC when it detects the temp. of the fins at 0 deg. Celsius.

 

If a used device that can be verified as working correctly can be purchased then it may be a good choice, if the price provides enough savings. presently many air conditioners use electronic temperature controls, with onkly the very cheapest using a thermostat. Also, 2 wires to a sensor do not mean anything, because the sensor could be a thermistor, a tempistor, or a diode. Or some other kind of device.

 

Installed in my neighbor's walk-in refrigeration room, the AC was set to cool to 60 deg. F. Then when 34 k Ohms of resistance is connected in parallel with the NTC thermistor of the AC, the AC brought the room temp. to 47 deg. F.

The measured resistance of the thermistor of the AC with respect to temp. equal to 60 deg. F and 32 deg. F. was respectively 14.6 k Ohm and 18.9 k Ohm. Assuming a linear decrease of resistance with increasing temp., that means that the resistance of the thermistor at 47 deg. F. would be equal to about 16.6 k Ohm.

The thermistor at 47 deg. F. with the paralleled resistance then has a total resistance of 11.2 k Ohm which is considerably less than the resistance of the thermistor at 60 deg. F equal to 14.6 k Ohm.

My approach was to get the AC to "see" resistance of the thermistor and paralleled resistance equal to that of the unmodified thermistor at 60 deg. F. to get it to cool to some temp. <60 deg.

My question is why apparently the total resistance must be about 75% of the measured resistance of the thermistor at 60 deg. F. to achieve cooling to a chosen temperature less than 60 deg.? Over the range between 60 deg. to 32 deg. it would not seem that the themistor would have resistance vs. temp. that is terribly non-linear.

Thanks if you have any idea,
Pete

 

Most thermistors are not linear over a wide range, and in addition, the control circuit is not designed to cool to that temperature. When using a system well outside of it's claimed performance area there is no reason to expect the operation to be a linear extension of the claimed range. Also, the physical capability of the refrigeration hardware is not assured outside of the claimed capabilities.
I like the approach that you used, simple and minimally intrusive. Good thinking indeed!

 

The AC that is being used to get the room temperature below the normal range that the AC would cool to is set to cool to 60 deg. Fahrenheit. The resistance of the AC's room temp. sensor, a NTC thermistor, with respect to air temperature of 60F equals 14.6k Ohm.

The idea behind the controller that I put together is to put resistance in parallel with the thermistor so that the compressor of the AC will run for a longer time than usual and thereby increase the resistance of the thermistor to point where the total resistance of the thermistor and paralleled resistor equal 14.6 k Ohm at which point the compressor should stop running.

However in testing the AC with connected controller, in achieving a room temperature of 42F, by trial and error I found that the resistance of the paralleled resistor must equal 27.6 k Ohm. The resistance of the thermistor at 42F equals about 17.7 k Ohm. This works out to a total resistance of thermistor // resistor equal to 10.8 k Ohm. That is, the required resistance is considerably lower than what it should be which is 14.6 k Ohm.

This does not make sense to me as it would seem that resistance lower than 14.6 k Ohm should result in the AC "seeing" a temperature higher than 60F and result in the compressor continuing to run rather than stopping.

 

Thermistors are not linear devices, although over small ranges the slope of resistance change per degree may be close enough to linear to be useful. So that is probably why the anticipated resistor value does not match the experimental results. A circuit intended to compensate for the lack of linear response over the normal temperature range may not have the same response outside that range. It is impressive that you are able to reach the low of 42 degrees with a room air conditioner, as I recall earlier comments that it could not be done with the refrigerant that is in the system. So your success shows that not every bit of advice is always correct.

 

So I removed the thermistor from the coil to outside the innards, and mounted a small lightbulb next to it to keep it warm. The A/C now runs constantly.

@wayneh
Can I ask what wattage of lightbulb you used? Did you wrap the lightbulb and the thermistor in aluminum foil?

My poor neighbor needs a refrigerated room for his bread recipe to work properly, and we are having difficulty in getting my technique of parallel resistance to get the temperature low enough.

Thanks if you can help,
Pete

 

My poor neighbor needs a refrigerated room for his bread recipe to work properly, and we are having difficulty in getting my technique of parallel resistance to get the temperature low enough.

Please be advised that A/C units aren't 'metered' for 'cold side' ambient room temperatures much below 65°F (i.e. mean evaporator temps < 45°F)...

Best regards and good luck!
HP

 

Please be advised that A/C units aren't 'metered' for 'cold side' ambient room temperatures much below 65°F (i.e. mean evaporator temps < 45°F)...

Best regards and good luck!
HP

In October of last year, I was able to get a room temperature of 40 deg. F by paralleling resistance with the thermistor of my neighbor's fairly large A/C. Whether or not I could have kept the room at that temperature for several hours, I can't say, as I didn't try that.

There is a company claiming to get cooling down to 34 deg. F with a device that is a heater in contact with the thermiistor of the A/C and the two wrapped in aluminum foil. See the link.

https://storeitcold.com/product/coolbot-walk-in-cooler-controller/

Probably you are correct that using an A/C to cool a room to close to freezing is not really appropriate.

Regards,
Pete

 

In October of last year, I was able to get a room temperature of 40 deg. F by paralleling resistance with the thermistor of my neighbor's fairly large A/C. Whether or not I could have kept the room at that temperature for several hours, I can't say, as I didn't try that.

There is a company claiming to get cooling down to 34 deg. F with a device that is a heater in contact with the thermiistor of the A/C and the two wrapped in aluminum foil. See the link.

https://storeitcold.com/product/coolbot-walk-in-cooler-controller/

Probably you are correct that using an A/C to cool a room to close to freezing is not really appropriate.

Regards,
Pete

While indeed, the thermostat may be modified (or merely bypassed) to achieve lower temperature operation (although frequent defrost [i.e. compressor off/fan on] periods will be required is maintenance of a clear evaporator heat-exchanger to be realized) -- The issue addressed in my previous post regards (application-specific) engineered/manufactured 'metering' of refrigerant flow to the evaporator (which being achieved via 'capillary tube{s}' or an 'expansion valve') -- Then too, is the issue of temperature-appropriate lubrication...

Bottom line being that while an A/C unit may indeed be caused to operate at significantly lower than design-minimum temps -- It'll be neither efficient nor 'long lived'...

Very best regards and again best of luck!
HP

 

Actually, the idea of heating the temperature sensor is the best one yet, because it does not connect into the control system at all. It just adds a bias heat source. So you have little chance of breaking the system I like that a lot.

 

@wayneh
Can I ask what wattage of lightbulb you used? Did you wrap the lightbulb and the thermistor in aluminum foil?

I'm currently using a 7W C9 Christmas bulb, one of the big old traditional ones. I've quit using them outdoors and have bulbs leftover, and they happen to fit the fixture I have.

I've used a little duct tape to hold the sensor closer to the bulb. They're held slightly apart by a small block of wood. Wrapping them completely would make me nervous. I've considered wiring in a resistor held next to the sensor but frankly the light comes in handy. Easy to spot across the dark basement.