CO2 Readings drop after power cycle

If I cycle the power on the CO2 sensor when it's reading over 1500ppm it drops to around 500ppm when it reboots. Not sure which reading is the true CO2 level.

Do you mean power cycle as in total power off and ON? or just a regular reboot?

Yes, I mean a total power off and on. A simple reboot does not affect the CO2 reading.

The CO2 sensor needs time to start up with a minimum value of 480-520. This because the way it detects CO2. Inside the sensor is a heater and it takes more than 5 min to stabilize. With power off shuts down the sensor’s internal heater; a reboot does not.

Here's the problem, whenever the sensor has been reading over 1500ppm of CO2 for several hours. I cycle the power buy unplugging it, which only takes a couple seconds. When the sensor comes back on, it does take 5 minutes to get a CO2 reading, but it's reading around 500ppm. I don't think the CO2 levels in the room drop from 1500ppm to 500ppm every time I cycle the power.

how long does it stay at 500ppm? or it doesn't go back to 1500ppm?

It will stay at 500ppm for several hours and then slowly climb back up to 1500ppm.

May I ask why do you need to do power cycle.
As I've tested my unit here, seems stable compared with other CO2 meter.
But I do not do power cycle. By the way why would you power cycle the units?

When I first install the CO2 meter I noticed when I came into the office the CO2 level was around 500ppm. As the day progressed the meter climbed to 1500ppm and even as high as 2000ppm. Those levels are considered unhealthy so I was concerned that I either had a HVAC problem or the meter was not reading correctly. That's when I started unplugging the meter to reset it and notice the considerable drop in CO2 levels in a matter of minutes. I don't have another CO2 meter to compare it with but this meter doesn't seem to be working correctly.

I did experience a climb around that value but that is due to the fact that using isopropyl or ethyl alcohol or sanitizers causing to have the CO2 level rise.
Although after a while it usually goes down to normal levels.
If you could try and check every hour the reading so that you can see if it is really defective or not.
but don't power cycle it as I did my test without power cycling it.

Here is a chart on our TVOC sensor on our work room.
https://ibb.co/mB0fPg6
As you may see the spikes are caused by using sanitizers and rubbing alcohol. I think when you actually read a spike, you shouldn't do power cycle as I think the sensor has a correction in itself. it usually returns to normal value.

I tested my sensor with isopropyl alcohol and got similar results. The readings spike but return to normal in a matter of minutes. My sensor also fluctuates throughout the day, it slowly climbs and falls back down. My concern is when it reaches 1500ppm to 2000ppm and stays there for over an hour or more and doesn't come back down. That's when I power cycle the sensor and get readings of 500ppm after it comes back online. I don't know which readings to believe. The 1500ppm readings, that have remained constant for several hours, or the 500ppm when I cycle the power.

I am assuming that the 1500 to 2000 ppm are real and thus maybe there is something that triggers that reading from your area.
you'll get false reading when you do power cycle.
If you think that the sensor is defective, you might as well file it for an RMA to them.
Though on my test most of my readings are fine.