Samsung M.2 NVMe drive temperature reading problem


New Member
After an upgrade to 7.42-5030 I noticed temperature reading on Samsung NVMe drive is getting weird. To confirm I ran the old version 6.28-4200 and it performed as expected.
To illustrate I made 2 screenshots. Looks like reading of temperature sensor in the new version is getting lost (and previous reading used) most of the time.


  • Screenshot 2023-04-11 194748.png
    Screenshot 2023-04-11 194748.png
    473.3 KB · Views: 10
  • Screenshot 2023-04-11 195539.png
    Screenshot 2023-04-11 195539.png
    841.9 KB · Views: 10
Previous "Drive Temperature 2" is now "Drive Temperature 3", it's explained here:
Thanks for reply.
Let me elaborate. The test (Crystal Dist Mark) runs minutes. With the old version of HWINFO the reading value changes almost every time a reading is made. Look at the graph.
But if I run the latest stable version of HWINFO, there are only 3 changes of values during the whole time. That seems wrong.
I didn't include other graphs into screenshots, but CPU and MB temperature graphs update normally, not like a step function as is the case with Samsung NVMe.
Ah, I see. Frequency of SMART polling has now been reduced by default to avoid too frequent queries to the disk which might in some cases result in spin up/down.
You can adjust the polling rate by going into Sensor Settings and change the "Disk S.M.A.R.T. every" value. By default this is every 100 cycles, so 100*2 seconds.
SMART polling has now been reduced
Given the reason for this change is HDD specific, is it possible to do this for HDD only?
I would speculate that in the future systems with HDD become less and less frequent.
On the other hand, defaults should be both safe and simple. Introducing complexity [for default value] might not be a good idea.
Sorry in advance if it's been already discussed when this change was implemented.