ResMed AirSense 10 Time as shown in OSCAR

[BruQu]

I keep notes about start and end times for my ResMed AirSense10 CPAP sessions, and add them to the Notes section in OSCAR. There is always time lost at the beginning as pressure builds (typically 6-7 seconds), and also more time lost at the end, which is recorded in OSCAR as 0, 1 or 2 seconds past the Start time's seconds.

Of late, I've noticed that what is reported in the Welcome screen differs from what is on the Details page by a second (lower), and that both start and end times are affected by whether I hit the "ON" button to start the session right on the minute, or at some later time within the minute, and whether turn-off occurs more or less than 30 seconds past the start time's seconds.

Can someone tell me how this ResMed model handles time recording, and how OSCAR reads that data into the record?

My assumption is that there is a "rounding" algorithm, and that these records are decimal minutes, not seconds.

Thanks!

Using OSCAR 1.4.0+e35d47b3, ResMed AirSense 10 (14.0/19.0)

[KeepSmiling]

I do not know how small increments are done.

On the ResMed 9 am using the time on it was different from the actual time.

In OSCAR under the following folder

File
Preferences
CPAP

There is a section: CPAP Clock Drift.

I adjusted that and the timing matches. (I never checked for the exact minute.)

[BruQu]

Thanks, but I'm not concerned about clock drift.

When a user changes clock drift, it affects the entire data set, though I guess it makes somebody feel better to see the latest day seem to coincide. As I indicated, I do keep tight and accurate records. I shifted from a Respironics machine to ResMed 1,040 days ago, and have tracked my data for the past 9 years on a daily basis. Some folks are content to be casual about it, but I pay very close attention ... obviously.

One reason for my interest is that insurance coverage for CPAP devices and supplies is tied to "compliance" that, for the insurer is an absolute value. For Medicare, that magic number is 4 hours plus, and even 3 hours 59 minutes 59.99999 seconds falls short. ResMed and OSCAR both appear to take liberties with recording time values, and they are always on the debit (short) side. ResMed imposes the day starting and ending at precisely noon, but it NEVER records time in a session PAST 11:59:50 and it NEVER records a time earlier than 12:00:08. In ResMedLand, days are only 23 hours 59 minutes and 42 seconds long.

I'll hope for a technical  answer from one of the OSCAR gurus.

[pholynyk]

The first thing I will say about the clocks in the ResMed machines is that they are not particularly reliable. The clock in my AS10 runs reliably slow; over a period of weeks it loses some minutes, I set it forward 8 minutes last night, but I don't recall when I last set it. I have not heard reports of the AS11 clocks gaining or losing time, but they are not set accurately, and can not be reset by the user.

In terms of the times recorded in the files, time in the STR.edf file is minutes past noon; whether rounded to the nearest minute, or the previous or next minute, I have no idea. In the other edf files time is recorded as strings "dd.MM.yyHH.mm.ss" - again, no idea if that is rounded, or previous or next second, but I would suspect previous 'tick'.

I am not sure what happens when a sleep session runs over the machine's noon. I would expect there to be a data gap as files are closed and new files are created.

The 'clock drift' referred to in the Preferences is a misnomer. It is a static correction applied to all times when the data is displayed, and has nothing to do with the machine clock running fast or slow.

[BruQu]

You're my hero! Your response tells me what is known, which is what I wanted, and I've saved that.

My experience affirms yours, especially that built-in clocks lose time. It's particularly acute if one unplugs the machine to remove the data card, and takes more than a few minutes before "restoring" power. To minimize that, I restore power to the now-cardless machine, cut power to reinsert the card when I have the data. I reset the clock periodically with no data card to a reliable source ("atomic clock" that syncs to the Naval Observatory every night - still off a few thousandths of a second but ...) by backing it up several minutes, then resetting it again to the current time. I'm still working on the algorithm used by ResMed, but you've given me an important clue.

I seem to have lucked into a very savvy sleep doc, who is more interested in my actual experience than in blindly taking data and assuming it is accurate and complete. I provide comprehensive info in as much detail as I can, and we work together to get me the best sleep. Some docs, who do know lots more about medicine than I do, still respect a very engaged patient enough to listen. Those are the ones whose patients thrive!

I'm not sure the "League of DME's" is doing themselves a favor by making it harder for patients to understand or to access the equipment they use, and I'm pretty sure I can find ways to not do business with the worst offenders. I am specifically responding to your comment about the "11's" not allowing patients to reset their clocks, which means I'll never have nor use one. I encountered a local DME who wouldn't even answer questions until they had billed Medicare "for the new machine" ... so I told them to tear up the Rx. I had a sleep lab tech who "knew more than I" and provided what my sleep doc thereafter described the resulting "guidance" (an hour into the appointment) as the worst sleep study she had ever seen. I now will not visit that sleep lab. But I digress ...

Thanks VERY much for your response! I appreciate you!

[Crimson Nape]

The time loss while it's unplugged sounds like your onboard battery is dying.

- Red

[BruQu]

That's what I wondered (though it's been a constant "feature" for over two years) ... but I haven't seen anything to suggest I could replace it. Searching online seems to produce results for adding an external battery pack as an alternative to household A/C power supply. Is replacing an internal battery feasible on an AirSense 10? Is it something like a personal computer's internal CMOS battery, one that's not soldered onto a circuit board (I'm a pretty good mechanic, but suspect that tackling that chore might result in a "paperweight")? Do you happen to know of a "how to" resource without compromising your need to be "arms length" from such instructions? My device was not new when I acquired it, so a failing battery would not surprise me.

Thanks much for your input! You all really are heroes! (and I'm not confusing you with sandwiches at a NY delicatessen!)

[Crimson Nape]

It uses a CR 2032 and is mounted on the main board. I can't remember if the A10's or the S9's battery is soldered in, one is.
- Red

[pholynyk]

My recollection from the post which had pictures was that the battery had welded tabs that were soldered on to the board.

I think they said it would be difficult to replace.

[BruQu]

I think I'll just hope the battery lasts longer than I do ;-)  As is, it's LOTS better than a paperweight!

Thanks!