Temperature control of 0.5c

[T82]

Need to maintain a small room for scientific research at 15c with only a 0.5 fluctuation allowed, any thoughts?

[mad fridgie]

This depends upon how small, (not really a size) so if small in my mind, then you have refrigeration sized to maximum possible load. this left on all the time. You then use a heater element and decent controller. To match the short fall on the refrigeration load under normal operating conditions.
If bigger than just small, then use refrig unit cooling a buffer tank, pump the water controlled by 3 port valve and decent controller

[dirk]

Mad Fridgie has given you the best and ideal soloution butIf trying to use standard cool room equipment and the load and ambient is such that heating is not required.Carry over from the evaporator coil is a important consideration, if you require a 0,5 differential. Make sure the system is pump down placing your solenoid as close to the expansion valve as possible, insulate the expansion valve bulb well ensuring that the valve is closed on the off cycle.Use a epr to regulate suction pressure.

[lana]

Hi there,

For this application it is better (or must !) to use a cap tube or hand adjusted throttling valve instead of TEV. The reason is this : TEV response takes time and this time does not allow to keep the temp. at a very close differential.
Second thing is you have to have some kind of capacity control. If the system is on and off, then you definitely will have temp. fluctuation.

Many years ago when I was student in London, there was a company who manufactured CNC equipment with high precision. The oil needed to be cooled and when oil temp. was varying even 2°C then CNC precision was out of range. They did a research with a well known Prof. at a university and the Prof. came up with a very simple and practical solution. He used a hand valve instead of TEV and adjusted the valve himself and took the head of the valve away so no one could touch it. They managed to keep the oil temp. constant at 0.5°C differential.

Of course control system plays a vital role beside the refrigeration unit.

Cheers

[Tesla]

Your Question is very open so I thought I'd put my five cents worth in. Depending on how much money is allowed, the size, and load variation. If a person walked into this small room the temp could shoot up over limits, to over come this problem a room within a room or a preconditioned entrance would be required. As for control - the closest you can get is constant ie not off/on. For this (DX) you would need an EEV, EPR, heater, Variable fan speed and VRV with a complex control system. High airflow rate to overcome temp difference in the room - This type of system was used for archives.
Another system I have worked on had two separate refrigeration systems where one cycles off then the other would cycle on via tstat - this system was not for close control but 100% backup for $millions of product. Mad fridgie's suggestion sounds better - chilled water is easier to control than DX.

[dirk]

My advise is not based on speculation but on a project that has been running for five years now, it was for agricultural research different soil types in room to check germination of seed at constant humidity and tempreature. the spec was for + 18c with 1c fluctuation we achieved fluctuation within .5c this monitored with computor data logging Danfoss controls used . the design of evaporator at 15c, control on a mixture of supply and return air . commissioning to achieve .5 fluctuation took two weeks. Room 4m by 4m 2.4 high. the load on the room remained constant same amount of soil each load although different soil types had different heat loads this became irrelevent as the mix gave a constant enough average, soil was precooled to 18c.
The project was cost driven.

[NoNickName]

The constant temperature in the room depends on how constant is the heat load. Cooling a rolling mill to such a temperature precision is impossible, because when the iron hits the cage, few thousands kW suddenly hit the chilled rolls.
In order to answer your question, you should first model the temperature profile of the goods to be kept at constant temp.

[sefuss]

From my experience, the hot gas or heater method built into your indoor unit works best, 3 way water valves can be ok but i have found they can take to long to react / adjust to sudden rushes of load

[chilled kiwi]

Hi T82.
Depending on budget etc I would reccommend using an Electronic expansion valve to be able to handle the varying load.
If the compressor is 3 phase: Have a Varible speed drive ons the compressor,contolled by one of the following options.
1: Via suction pressure. If the load is reasobably constant this will be OK but may wander at differant conditions.
2: Use a 4-20ma probe (or a PT100 via converter) mounted in the supply air, to control the drive.
3: Use a Carel IRDRV3000(or equivilent) to control the drive. Again, have the temperature probe in the supply air.
On all 3 options ensure you have the sleep modes etc set up with close tolerances. I use option 2 and maintain virtual flat lines on the graph (+/- 0.1k)
If the compressor is single phase:
The best option is to use a eletronic hot gas bypass valve, injecting discharge gas into the evaporator between the EEV and the distributor. The Carel EVO exansion valve driver has an option on it to control the bypass valve as standard. Place the control probe in the supply air.
Please note, I do not work for Carel, it is just the gear I am used to dealing with. Other manafacturers have similar products (Danfoss, Dixell, Sporlan)