Normal compressor cycle time

[Jim_H_WY]

We have several True T-49 "reach in" refrigerators.

The factory thermostats seem to like to fail with the contacts welded closed such that the samples we're storing in the coolers end up freezing which costs us many thousands of dollars.

Also, these thermostats are psychic and evil. They know that it's always most damaging to stick closed over a holiday weekend such that the maximum time will elapse before the problem is discovered, thus assuring that every sample vial in the cooler will be frozen and broken.

One failed over last new-years, and most recently, one failed over Thanksgiving. Very sneaky!

Not to be outfoxed, I've designed and installed freeze protection devices which operate completely independently from the units' thermostats and controls. That has saved us from several of these episodes being catastrphic already.

But to further enhance things, I've decided to replace the thermostats themselves with electronic controllers, thus eliminating the badly designed or underrated units that they come with.

Also, the standard replacement thermostats are quite expensive (far more than the cost of a very good electronic controller, solid state relay, and enclosure).

But here's the question:

The factory thermostats cycle the compressors on and off quite frequently. And the one I've designed and installed to replace the factory unit also cycles the compressor fairly quickly (although not as fast as the factory thermostats).

Is there any way to know what the minimum allowable "off time" is for these units so that I can set things up to prevent the compressor from turning on until at least that length of time has elapsed since it last turned on?

With my current settings, the unit comes on for about two minutes, then shuts off for about five minutes. The factory units cycle a bit faster.

But in both cases, if you open one of the doors, the unit will come on almost immediately because the warm outside air is "seen" by the temperature sensor very quickly.

So the potential exists for the compressor to be turned on very quickly after it just turned off if someone opens the door right after the compressor just turned off.

These are small units with 1/2 horsepower 115VAC compressors that the nameplate claims will draw 9.5A full load (I measure about 8 amps when running steady). They have a 17 oz charge of R134A if that helps.

What would the minimum "off time" be for such a unit? Or are these designed to prevent problems if they short cycle? It seems to me that they'll short cycle just about any time someone opens a door if they're not already running. And yet, we've had zero problems with the compresors.

It's just those malevolent thermostats that fail frequently!

[Magoo]

Sounds like the stat is switching the total system power/amps. Try installing an interposing contactor that will handle the power/amps, and control the contactor with the standard stat.

[mad fridgie]

Hi, first check ratings of the contacts within the controller, make sure that they are rated for inductive load, not just resistive load (if rated resistive, then use a external switching device.
Most modern electronic refrigeration controllers have an inbuiilt anti-cycle timer (stops short cycling).
Max starts per hour on these small machines is 10, as you get bigger, then the starts per hour generally reduce.
So your anticlycle time of 5 mins would seem OK.
The question is; are these short term temp swings exceptable?

[Jim_H_WY]

Thanks, guys. But that's not the question.

I know that the OEM thermostats are obviously not up to the task or they would not have their contacts welding. And if I wanted to stay with the OEM thermostats, I could install an interposing relay to better handle the load of the compressor.

But since we need better precision anyhow, and since the thermostats with the burned contacts need to be replaced even if we DO install an interposing relay, we'd be faced with the expense of buying new, replacement thermostats.

And quite frankly, for what they are, they're outrageously expensive.

So it's no problem, and cheaper, actually, for me to just install a good control instead of a new OEM thermostat of dubious accuracy, and then need to put in an interposing relay to make it reliable.

Thus, I've replaced these "expanding fluid" type thermostats, for less money, with modern electronic systems which are more accurate, easily programmable, and I'm using solid state relays which are well up to the task. I've had ones like this running for ten years continously with zero issues. It's the OEM t-stats that were the problem.

BUT, the real question I've got is:

What is the minimum "off time" for a compressor of this sort?

My electronic controllers can cycle these units fairly fast if I let them.

The factory OEM thermostats cycle the units very fast, and you cannot adjust them at all.

Yet we have had no problems with the compressors. So I'm curious as to why this is. Are we just lucky so far?

So since I worry that the compressors are being "short cycled" by the OEM thermostats, and since I've got the ability to set the electronic ones to do whatever I want, I would like to get it right.

Allowing shorter cycles provides tighter temperature control which is important to us. But again, I worry about what this does to the compressors.

Are these compressors (or something else in the systems) designed, somehow, to bleed off the pressure very quickly so that they're ready to go again in a very short time after being switched off?

Maybe I'm just overestimating the amount of time you really need to let this kind of compressor rest before powering it up again.

If I can find out the minimum "off time" for this type of compressor, I can program that delay into my microcontrollers so that they'll protect the compressors from short cycling. But I'd like to allow the system to kick on as quickly as possible while still being safe for the compressor, to keep the control as tight as possible.

Thanks again for your help, guys. This is a very helpful forum!

[Jim_H_WY]

Well, clearly I cannot read, or another post got in there before I posted.

Thank you very much for the information. I'll make sure I've got things set up to keep the starts to 10 or less per hour.

I can't see how to edit a post on here...

Anyhow, the OEM thermostats in these units cycle them quite rapidly, but close to your figure of ten per hour. Looking at the data acquisition system, it appears that the OEM units cycle between ten and twelve times per hour.

With the electronic ones set up the way I'm using them, I get more like 8 cycles per hour.

The temperature swings are acceptable for our purposes at these rates.

So I guess we're good to go and I won't worry about it.

That's great information because I was sort of thinking that this might be far too fast.

Thanks again, guys! I'll try to read better before posting in the future.

[Brian_UK]

Hi Jim and welcome to the forum.

Further to the advice given above and your additional pressure question....

The reason for the time delay on the compressors is because the motors are cooled by the returning refrigerant gas. Obviously, every time the compressor starts the motor heats up and takes a certain amount of time to be cooled down again.

Short cycling of the motor can cause it to overheat and eventually burn out, hence the liit on the number of starts.

Regarding the pressure thing...

Some compressors can handle starting against a pressure differential and some require a low differential otherwise the compressor can stall, overheat and away we go again.

With respect to the sensitivity of the electronic control sensors; was the old thermostat sensing air temperature or was it in contact with the evaporator pipework?

[Royal241]

If you are so tight on running a low differential I would replace the cap tubes with an expansion valve and selenoid (pump down), add a few more years on the compressor. But that's some project on a little cooler, is it worth? Then do it!

[Jim_H_WY]

Thank you Brian and Royal.

I did not take the thermostat out of the offending unit that I worked on most recently. As it turns out, it's really easy to simply abandon it because the condensing unit simply plugs into an outlet supplied by that thermostat. So I didn't need to even remove the old thermostat to eliminate it from the system.

And I was not around when another one was replaced about a year ago in a similar unit. So I do not know if the sensor for the OEM thermostat is in contact with the evaporator tubing or if it just measures air temperature.

But my guess is that it measures air temperature. I don't think it's one that can check for evaporator freeze-up. And that's fine in this application because there's almost no source for water inside the coolers and the humidity is very low here normally.

We've never had one of this type of cooler freeze off in the 20 or more years that I've been associated with this place.

I installed my controller's temperature sensor so that it measures the air temperature as it's drawn into the evaporator enclosure. So it's seeing only air temp.

Thus, I achieve the differential by directly programming the turn-off point and the hysteresis (deadband) values into the controller.

The control has been excellent with this set-up, and even with the dead-band set wide enough to achieve an acceptable cycle rate, our samples do not experience much of a temperature swing themselves.

We must keep the samples at 4 degrees C plus or minus 2 degrees. So I set the turn-off point to be 2C and the hysteresis to be 4 degrees C. That cycles the air temp between our acceptable upper and lower temperatures but the worst-case samples see a swing of only about 0.15 degrees peak to peak due to their heat capacity to surface area ratio.

We actually have our data acquisition system set up to simulate the worst-case samples which are small, long-skinny glass vials with about 40 milliliters of water-based sample inside.

We use temperature probes housed in machined housings that achieve the same thermal time constant, and thus correct for air flow, etc.

I kept the sensor for the temperature controller very small with high surface area so that it responds very quickly.

I could set the sensor up to have higher thermal storage and some insulation if I want it to have a longer time constant, but I figure I've got more possibilities if I use a fast sensor and then control the cycling with the controller.

In this case, actually setting the limits to be our desired limits is working very well. The samples stay right in the center of the desired operating band and from what you folks have told me, it sounds like the compressor is not being cycled too rapidly. So it's working well.

I plan on laying out a PC board for a controller that has inputs for three sensors for future use. Two sensors will go to the main controller and the third will be available to a completely separate "safety" or "anti freeze" system that will act as a fail-safe. Freezing of samples is a major concern for us with these coolers and some incubators that also have the ability to cool.

It's really expensive and embarrassing when you freeze a bunch of client samples

In some systems, the dual probe controller will let me do automatic defrosting of the evaporator with one sensor while doing the actual control with the other. In other (heating) setups, that second sensor will allow for better control using a method that I hope to patent.

The third sensor will not always be used because it's only needed for the fail-safe anti-freeze circuit which will be completely independent.

Right now, I'm just using an off-the-shelf controller. The good thing about it is that it's cheap and readily available. The bad thing is that it doesn't directly accept my preferred semiconductor temperature sensor ICs, so I'm actually using a thermocouple for this application which is not ideal.

But hey! It works just fine and is a LOT better than the OEM setup.

I really appreciate your help, folks. I know a bit about electronics and programming, but not so much about cooling systems. It's an amazingly intriguing field and I'm pretty clueless about most of it.