Forum Replies Created
-
AuthorPosts
-
uRADMonitor
KeymasterThanks wolf, I hope to have the API ready as soon as possible. I will post all details here, and in a dedicated blog post as well.
uRADMonitor
KeymasterThat counter is a uint32_t, and it can go from 0 to 2^32-1, or 0..4294967295 .
4294967295 seconds means aprox. 136 years.
uRADMonitor
KeymasterThis can only indicate tube damage during transit, but it is somewhat unlikely, as it never happened until now. Unit data during pre-shipment tests is attached.
Attachments:
uRADMonitor
KeymasterCertainly – New coordinates in place. Congrats on running the first Russian unit. The SBM-20 has returned home 🙂
Thank you!
RaduuRADMonitor
KeymasterThat is a very nice solution!
However I am surprised you preferred to use the HTML content instead of the JSON available under http://ip/j
uRADMonitor
KeymasterCoordinates updated!
I will think about a way of implementing your requirement. Thanks!
uRADMonitor
KeymasterYour SQL query has been executed successfully (Query took 196.2649 sec) CREATE INDEX idx_db_devid ON uradmonitor( db_devid )
I was just creating 7 charts in parallel for some experiment with various tubes. It was hard for the server, but after adding this additional index all 7 were generated instantly!
Speed improvement: amazing!
Thanks
uRADMonitor
KeymasterHi Rick,
I didn’t notice this in my tests so it might be a particular case. Ignoring the first readings is not the best option, we would need to see the root cause of the spike instead. Ignoring would waste precious data, and we’ll never know how to set a “first interval” value. 1 minute? 10minute? Not reliable as if there is a problem it can happen at any time.
Looking at the chart I see the time axis is continuous, where did the power off occur?
On older firmwares there was a default duty cycle set for the inverter. At power up this could have produced higher voltages, potentially resulting in excessive counts on the geiger tube, however I see you run 110 already, where the duty cycle starts at an absolute minimum and is brought up to the required value, as controlled by the internal regulator.
Another cause might be the so-called “compensation capacitor” that might need to be replaced on your unit. This component is responsible for controlling any unwanted oscillations taking place over the tube. But I see no reason why this capacitor would fail, so this is not a real explanation either.
Power failures might still induce voltage spikes of various kinds. Can you run a few power off/on cycles and see if this is happening on a consistent basis? I suspect not.
But let me know.
Thanks,
RaduuRADMonitor
KeymasterHehe, I wrote a linear color divider like this myself some time ago, but in C. Thanks for the code!
uRADMonitor
KeymasterDone!
uRADMonitor
KeymasterMost likely when the cron called mysqldump backup script runs, some of the new entries are lost.
My plan was to archieve the old data periodically and keep a smaller DB on the run. This would have helped the backup process as well, yet some time would still have been used.
I will look on the alternatives you’ve indicated.
Thanks!
uRADMonitor
KeymasterThanks Vinz, I’ll add it to the prod server too since select by id is often used.
Would be very helpful if Ally or the other guys with expertise in DBs would step in for some more hints/comments.
uRADMonitor
KeymasterThank you , I was planning to add a colour gradient. Suggestion added to the todo list.
uRADMonitor
KeymasterCoordinates updated!
uRADMonitor
KeymasterThe location has been updated as per your request.
-
AuthorPosts