You’ve got your monitoring system installed: Great!
Now you just need to set up your alert set points, and then you won’t need to think about it anymore.
But how do you do that? What happens if you don’t choose the right alert set point? How do you know you chose correctly?
If you’ve never done this before, it can be hard to know what to do here - under-sensitive alerts might mean you’ll miss a problem with your equipment, but over-sensitive alerts are annoying and will cause alert fatigue.
Alert fatigue happens when you receive so many alerts that you learn to ignore them. If you receive 99 alerts that are false alarms due to normal usage, it is pretty likely that you will ignore the 100th alert. Of course, that 100th alert is the one that indicates a freezer failure, or something equally important.
At TetraScience, we’ve monitored more than 3,600 pieces of equipment over the last four years. We've collected lots of data about how freezers and incubators perform and what alert settings work for the Lab Managers and Ops/Facilities folks who rely on our Lab Monitoring system. One of the most common requests we receive is for alert setting recommendations. We decided to gather our data and anecdotal experience and write it down - so here it is:
Alert variables: threshold and delay
When setting alerts, the variables we have to work with are alert threshold and delay. The way alerting generally works is this - once the temperature (or other data feed) has crossed the set alert threshold, a timer starts. That timer continues for the duration of the set delay, and if the temperature is still out of bounds after that period of time, you receive an alert.
As an example, if you have a -80? freezer, where you have set an alert threshold of -65? and a delay of 30 minutes, you would receive an alert if the freezer temperature went above -65? and remained above -65? for 30 minutes. The figure below shows what this looks like on your TetraScience Lab Monitoring dashboard.
Setting alert thresholds
The purpose of setting an alert threshold is pretty straightforward: if your equipment exceeds a set temperature, you want to know so you can take corrective action.
When selecting a threshold, the important thing to remember is that temperature naturally fluctuates during normal usage.
For example, a primary cause of temperature fluctuation is users opening and closing the door during the course of a normal day - this will cause excursions that last for a few minutes before your equipment recovers. Freezers will also have a sinusoidal temperature variation even when nobody is using them. This is caused by the compressor cycle and can be a noticeable variation (+/- 5? would be very typical).
Because these variations are part of normal operation, we typically recommend thresholds that are a few degrees outside of normal operations. Below are some examples of reasonable thresholds:
Note that we do not recommend low temperature alerts for -20s or -80s: these generally are not useful and can cause alert fatigue. We do, however, sometimes recommend low temperature thresholds for 4? refrigerators, as you may have material that you do not want to accidentally freeze. However, risk of hitting freezing temperatures does vary by type of refrigerator. For example, low temperature thresholds can be helpful for refrigerators that are likely to have frost on top.
Setting alert delays
The purpose of setting a delay might be a little less obvious - wouldn’t you want to know immediately if your equipment is out-of-bounds? In some cases, this is true (if you’re monitoring a gas manifold, and it goes into alarm, there’s no reason for a delay).
But for cold storage and incubators, the time it takes to return to the temperature set point after opening the door can be 20-30 minutes or more - especially if users open and close the door several times in rapid succession.
Because of this, we typically suggest alert delays in this range:
We can use a shorter delay for incubator high temperature alarms because the incubator temperature does not generally rise with normal use. High incubator temperature normally means there is a problem with your HVAC system or the incubator itself.
The best monitoring system is one that you use, not one that you tune out.
Alert setup shouldn’t be difficult, and it shouldn’t take a lot of fine-tuning down the line. Though lab equipment can show some variation in typical operating ranges, we usually find that you should be able to set universal alert thresholds and delays for each type of instrument. The important thing to remember when you are setting up your lab monitoring system is to avoid the extremes of alert sensitivity - the best monitoring system is one that you use, not one that you tune out.
Do you have alert threshold/delay setting best practices to share? Additional questions? Let us know!