Crazy Usage Hours
Open Source CPAP Research and Review Software
Status: Beta
Brought to you by:
jedimark
Menu
▾
▴
#38 Crazy Usage Hours
I just installed the new 0.9 beta version. After initial review, things are looking really good. However, I\\\'ve got 3 days with crazy usage data.
10/22 - 500:17 hrs
10/30 - 497:51 hrs
11/9 - 510:08 hrs
I have a ResMed S9 Autoset machine.
Because of these strange readings, my graphs for Usage and Session Times are distorted as a have 3 very tall bars and the rest of the days, the bars are barely visible. See attched screen shot.
Can you help figure out what\\\'s going on? Any data you want me to send to you that might help?
Love Sleepyhead!
Thanks!
Graphs
Ouch..
Can you perhaps send me the EDF data for those days (and your STR.edf file)
I've got some good excuses to test out the auto-updater now :)
I forgot, does it mention the number of sessions on these days?
(Does the Sessions chart down below spike like this?)
I'd be happy to if you can give me some step by step instructions on where I find this data and how you want me to send it to you. Yes, sessions are shown and there are multiple sessions on each of the days with issues. (I don't see any chart below?)
The sessions are stored by date in the DATALOG folder of your SD card.. (You find all this in "My Computer" with Windows Explorer)
EDF files all begin with a date and time..
There are 4 different EDF files making up each session, PRD, EVE, BRP and SAD..
If you sort by filename, (by clicking on the header in Windows Explorer) it will order them so you can easily find the dates..
I don't need the .crc files with the same names.
STR.edf is in the SD cards root directory.. (in fact, that alone would be a good start if the rest is confusing, as it gives clues on session size)
If you create a temporary folder and copy them into it, you can right click on it and create archive/ compress folder / or whatever it's called in your right-click menu..
There aren't like 10 or 20 sessions listed on these days?
Not sure if I still have the data you need. I copied the STR.edf file from the SD card and also records from DATALOG entries for 11/9 and 11/10/2011. See attached file below. I didn't start backing up detail data until 12/10 so don't have detail entries for anything before 11/9. But 11/9 was the last of the 3 days where I had excessive readings. There were 3 to 4 sessions for each of the 3 dates. Hope this helps. If you need anything more, let me know and I'll try to get it for you.
Mark, I just looked at 11/9 on the prior v.0.8.8 pre-beta and see that it came up with 10:42 hrs for that date with 4 sessions starting at 5:36 PM and ending at 5:05 AM. Looking at the ResScan records, I had a 5 minute and a 7 minute session on 11/9 starting at 5:36 PM where I think I was trying out a new mask. I didn't actually go to bed that evening until 10:38 PM.
That data was the right files.. thanks..
It could be the combine close session slider in preferences causing the differences with ResScan..
SleepyHead by default tries to sticks sessions closer than 4 hours apart on the same day.
Try sliding it all the way to the left (off)
I uploaded a screenshot from showing how your data is rendering here.. I need to check this data over on windows notebook to make sure it's not a platform bug..
I tried adding a screenshot of my graphs but don't know if you can see the image or not. For some reason, when I click on it and try to open it, it says file is unsupported, damaged or corrupted.
Moving the slider didn't change anything on my graphs. It looks like your graphs are rendering correctly. I can't figure out why I have those 3 distorted days. If I change the min/max settings for the usage and session times, nothing changes on those graphs. Still getting 3 days with very long bars and the rest of the days hardly visible. The session times scales are completely different too, going from 12:00 AM to 10:00 PM instead of 3:00 PM to 11:00 AM. Any other ideas?
This ones got me miffed..
Going by the previous screenshot your running Vista..?
Does the same thing happen if you create a new profile?
No, I'm running Windows 7.
I'm at work now so I can't try creating a new profile until tonight.
One other thing I noticed this morning was that my graphs start on 9/29/11 and yours starts on 9/30/11. 9/30 was the first day that I started using CPAP.
Created a new profile and imported the data with the slider to the off position before importing the data and still the same issue with those same 3 days. Any other suggestions?
Another thing I'm noticing about the way your graphs rendered is that it's showing 2 days with low usage, and it looks like those 2 days are adjacent to 2 of my 3 problem dates 10/22 and 10/30. And my graphs should reflect 100% usage because I've used my machine at least 4 hours every night since 9/30 (ResScan graphs show 100% usage).
One more odd observation. If I look at the daily tabs for the 3 bad days and then the details tab below in each case, at the bottom where it shows the Session ID's, in each case the first record has a start time of the evening in question and an end time of 13:03 which is faulty as I would never wake up at 1:03 PM.
Interesting.. Something else is going on with your data..
The question is why am I not seeing this..
SleepyHead has a special way of sorting it's dates into sessions.. If both the Combine & Ignore sliders in Preferences are fully to the left (switched off) and the split date is set to 12noon, it should work the same way as ResScan.
Every time SleepyHead starts it resorts the sessions into days, so you shouldn't have to reimport..
I'll try importing again with just your STR.edf data and see if it's a glitch there..
I have never seen this happen before for ResMed data.. :-/
You might need to try
I might need to try...what?
Both the combine and ignore sliders are set to off and split date is set to 12:00PM.
Frustrating. I've really grown to like the Sleepyhead software, but this bad data has made the v0.9 beta unusable for me. I'm willing to try any ideas you can think of.
For now a workaround to make the program useable, is to go into those days in Daily view, and flick the session switches down the bottom Details tab off..
Done and at least I can see the usage and session times bar graphs normally again now. Reduces my compliance slightly from 100% reality, but I guess that's the lesser of the two evils and my overall statistics aren't skewed now by those 3 days. Thanks for attempting to fix.
Still some inaccurate data for usage hours. Comparison of ResScan versus Sleepyhead data for last night:
ResScan:
Went to bed at 10:42
Woke at 5:07
Usage hours 6:26
1 mask event
Sleepyhead:
Went to bed at 10:42
Woke at 5:06
Usage hours 8:53
3 mask events
1. 22:42 to 2:31
2. 2:33 to 5:06
3. 2.34 to 5:06
Combine and ignore sliders set to off position
In reality I did wake up at 2:31 and shut the machine off for about a minute
But how does Sleepyhead come up with 8:53 usage hours?
Ah Ok.. There is a clue there..
I see it's not merging two sessions that are close together that should really have been just one session, and it's counting the hours twice..
If you look at the date and time of these sessions in sleepyhead, and then look at the filenames in the DATALOG folder matching these times, it might pinpoint the answer
Can you possibly extract a list of all the *.edf filenames on this same date..
I have only ever observed these "drift" sessions 2 seconds apart.. But you may have exposed the problem goes deeper than this.. (resmed didn't help the issue by choosing a fresh (and potentinally different) timestamp eachtime they save a session file belonging to the same session to the SD card..)
At work now and going out tonight. But, I'll try to get you the data sometime tonight. Thanks!