Temperature controlablilty of Aspen compressor

I have no experience with the Aspen system so I can't comment on it, but why not try the solid-state alternative, thermoelectric cooling modules? They are infinitely variable, silent, quick to respond, compact and use no refrigerant (you still have to remove the waste heat from the hot side of the cooler). You have the option of controlling the fan speed, cooler output, or both simply by having a controller that measures the enclosure air temperature and varies the output voltage/current to the devices. The real trick is to determine where to place the sensor.

Hi RAM. The little brother of this machine does use Peltiers, but it has a much smaller heat output. So we came up with a Peltier solution for cooling this machine, but by the time we had enough heat transfer capability we had so many Peltiers and so much heatsink surface area that there wasn't really enough room to fit everything else that we needed to. And the Peltiers were still much less efficient than a refrigerant cycle should be. We seem to be stuck in the no-mans-land between smaller heat loads where Peltiers are optimal and larger heat loads where refrigerant cycles work well. The Aspen compressor looks like it should run in this range though -- if it is sufficiently controlable.

I`d suggest utilising the standard set up for driving the compressor but forget +/- 0.1C stability. Over cool the air, and re-heat with an electrical element finely tuned via PID controller and thyristor driven.

If you ensure steady state for the refrigerant system by forcing "full load" condition for 100% of operation, it will flat line providing condensing pressure/ temperature is maintained accurately and provide reasonable accuracy thereby giving the electrical trimming little to do.

Hi Leon. Your response of "forget +/- 0.1C stability" is very much the same as that which is making us shy away from this approach. What I am having difficulty with is understanding what underlying physical process makes this kind of control not possible; is there some kind of fundamental long wave instability in the evaporator? Obviously the boiling process is violent, but these instabilities should be random and on the order of a second or two. Nick alludes to the system mass below; it seems to me that that mass should smooth out any transient short term thermal variations. What am I missing?

That +/- 0.1C stability requirement is on the temperature sensitive components down stream. There is a fairly long air path through metal from the evaporator to the temperature sensitive components so I should think that any short term temperature variations in the air leaving the evaporator will average out from contact with the metal passageways. And the critical electronics themselves have some mass.

At this point, we have some physical space limitations which make adding a duct heater problematic.

In all honesty if I were to design a system to give the results of air off temperature you seek, I would shy away completely from a direct expansion system where you are looking for such tight control of the secondary media. As you suggest the boiling process of refrigerant is erratic, couple this with the variable cooling effect of the evaporator due to the drying effect upon the air, oil content and collection points, plus the decision as to whether it is best to use capillary, expansion device whether mechanically or electronically driven.....the list goes on unfortunately.

I have had the best results at achieving +/- 0.1C utilising a water coil served by a water chiller fitted with hot-gas and liquid injection which is capable of such tight control. The water can run freely through the coil if the load is constant and a relatively tight air off temperature achieved. Alternatively fit a 3-port valve in a "tempering loop" in order to maintain the air off temperature control. Finally, the electrical heater which I mentioned earlier with PID control, this will "iron out" any swings in stability. It`s the only element in circuit where you will have infinite control. The mass of water in circuit is also important, in order to maintain stability, the more the better.

This may seem an expensive way of doing the job unfortunately, but it does just that........it does the job.

If water is > 23°C it will not work. So that an other circuit has to maintain water temperature low enough.

When I mention the mass I thought about an application I know. A measuring room quite big was controlled within ± 0.3 °C. The goal was ±0.5 °C and the result was better than expected. This is not as easy if the mass is very small since no thermal inertia is available to smooth.

What I think if you are afraid of the erratic boiler behavior is that you can intercalate a mass between evaporator and cooling circuit. This mass being the "heat sink" for the cooling air. This mass (thermally speaking) can be a water reservoir since water has the highest Cp so that you get the lowest volume for same heat inertia.

You will have in fact 2 circuits : one to cool the thermal inertia which can accept a quite crude regulation and the one to cool your component which will be fine tuned.

Hi Leon,

The remote chiller idea is indeed our fall-back which I think that we are going to go to. The 3-way valve was our first approach to temperature control with that, but it doesn't look like those valves will be precise enough either, so we have been researching the heater control idea. This does seem like an expensive way to go about it, but maybe it is the only way.

Your choice is correct. It depends on the environmental temperature if it is possible to have it over 20°C then it is the only solution. If the temperature will never be > 23°C then you could use a simple convection surface which can be important due to the low temperature difference. As well your choice of an important "I" is correct but do not forget that a too high "I" can lead to instability (phase shift). The tolerances you set (± 0.1°C) are very tight and depend on the mass you want to maintain a constant temperature and on the capacity of your cooling equipment.

Hi Nick. Unfortunately, it is possible that ambient could be set above the 23C target. Thanks for the caution on "I" as well. With positioning motors it is possible to tune the PID; hopefully that would be true with this type of system as well -- if controlling it is possible at all!

I can't offer experience with the proposed system, but I would suggest that you consider a cascade control scheme with your primary temperature control setting a secondary loop. The secondary loop could be evaporator or suction pressure control (preferred) or temperature control on the evaporator itself (possibly). This would help eliminate variances due to condenser temperature changes, compressor 'hunting', etc.

Very good tip Bigg. It has been a long time since my "Controls Engineering" class, so I had to look that up. I think that you are right. directly measuring the suction pressure should work -- now I just have to go find an appropriate transducer.