[Renamed] S9 graph display problem

[idamtnboy]

Dreamdiver,
Here's the latest. I did a complete clean uninstall and reinstall of 3.12 and created a new patient record from saved data copied back to an SD card. The time ruler on the nav window was messed up same as before. I then did a complete clean uninstall of 3.12 and a clean install of 3.11, and once again created a new patient record from the SD card. The time ruler in the navigation window has exactly the same problem as with 3.12. I rebooted between the uninstall & reinstall.

I then installed 3.12 on my Thinkpad lap top. Almost agonizing because it has only a 1 Ghz processor and 376 Mb ram! The time ruler on the lap top copy is correct! Go figure. I wonder if the problem is related to video drivers on my PC and how they interact with MS graphing libraries that are part of Resscan. Both computers are running WinXP SP3.

Going back to the data skew issue. I discovered that the event time in my Oct 25 to Oct 31 data is off by about 20 seconds. The card had data on it beginning Aug 4, which makes about 95 days of data. Looking at the EVE_header.txt files (extracted for the EVE.edf files) I see the record start time is about 20 seconds later than what is in the BRP_header.txt. The time value in the filename in all cases is consistent among all 4 edf file types, which means that the start time in EVE.edf is the same 20 seconds later than the filename time. The event onset time value, recorded as seconds from beginning of recording, for each event as recorded in the EVE file is the correct increment, i.e., if it is added to the BRP start time it falls dead on at the end of apnea time, plus first breath after. The only data value that I see that is screwed up is the record start time inside the EVE.edf file. Resscan uses that to place the event against the time ruler. Now, in my Nov 1 to Nov 12 data the times are correct, but the card was erased Nov 1 and Nov 7. Is this consistent with what you determined about the skew?

[idamtnboy]

Are other 3.12 users also seeing this?

I'm not.

I tried a couple of dozen random samples of nights from both my S8 and S9 data, and the bottom graph set aligns correctly with the selected event on the upper graph set for any clicked-on event on all the sampled nights.

Cheers,

Bill

When I click anywhere on the event or flow graph in the nav window, the corresponding zoomed in part of the detail graph lines up with the red line in the detail window. The time ruler in the detail window is correct. But the time ruler in the nav window is off, up to 1 1/2 hrs if the time span for the nav window is set to 24 hrs. That's on my PC. The lap top copy is correct.

[idamtnboy]

Going back to the data skew issue. I discovered that the event time in my Oct 25 to Oct 31 data is off by about 20 seconds.

I looked at the EVE files for Oct 1 and Oct 15. There are 3 for Oct 15. The time discrepancy between the filename time and the start time in the EVE file varies. One of them is almost 1 minute difference. There appears to be no solid correlation between how much data is on the card and the time shift. It's weird that the filename time is correct, consistent among the four edf files within 1 or 2 seconds, but the recording start time is not. It sure looks like there is a glitch in the S9 firmware somewhere.

[DreamDiver]

Dreamdiver,
Here's the latest. ...
Is this consistent with what you determined about the skew?

idamtnboy,

First let's clear the S9 firmware skew off your plate. Let's stop looking at .edf files altogether. We know they are not the cause of this problem. You've already shown they are valid, as well as we can validate them considering the skew problem. Firmware issues are not responsible for your ResScan-specific graphing scope skew.

Because you see no problems with your data on your second windows machine, we can further determine that the problem is specific to one computer you are using and not with anyone else's computer. For purposes of discussion we'll refer to the machine that won't install ResScan properly as 'Wonky'.

The diagnosis may be one or a combination of the following:

• Wonky's windows registry is corrupt. That may mean you need to reinstall windows on this machine.
• There is a .dll conflict between one application and ResScan on Wonky. In other words two programs have put a .dll with the same name in the same place that do slightly different things, but not enough to totally fry either application after install. That's not supposed to happen on modern windows machines, so I'm not sure this is as likely as windows registry corruption.

Either way, since the problem is isolated to one machine of yours, it is less likely a problem of ResScan, and more likely a misconfiguration of that one particular machine that's affecting ResScan.

The remedy for either diagnosis will unfortunately likely involve either an attempt at System Restore (to the earliest version available before you installed ResScan the first time) or at worst, a clean reinstall of Windows on that machine. If it still continues to happen even after a clean reinstall of windows, there is the uber-remote possibility that Wonky has a hardware flaw for which you won't be able to compensate.

[idamtnboy]

idamtnboy,

First let's clear the S9 firmware skew off your plate. Let's stop looking at .edf files altogether. We know they are not the cause of this problem. You've already shown they are valid, as well as we can validate them considering the skew problem. Firmware issues are not responsible for your ResScan-specific graphing scope skew.

Because you see no problems with your data on your second windows machine, we can further determine that the problem is specific to one computer you are using and not with anyone else's computer. For purposes of discussion we'll refer to the machine that won't install ResScan properly as 'Wonky'.

The diagnosis may be one or a combination of the following:

• Wonky's windows registry is corrupt. That may mean you need to reinstall windows on this machine.
• There is a .dll conflict between one application and ResScan on Wonky. In other words two programs have put a .dll with the same name in the same place that do slightly different things, but not enough to totally fry either application after install. That's not supposed to happen on modern windows machines, so I'm not sure this is as likely as windows registry corruption.

Either way, since the problem is isolated to one machine of yours, it is less likely a problem of ResScan, and more likely a misconfiguration of that one particular machine that's affecting ResScan.

The remedy for either diagnosis will unfortunately likely involve either an attempt at System Restore (to the earliest version available before you installed ResScan the first time) or at worst, a clean reinstall of Windows on that machine. If it still continues to happen even after a clean reinstall of windows, there is the uber-remote possibility that Wonky has a hardware flaw for which you won't be able to compensate.

Let's make sure were clear and agree on the issues. One is a time ruler display problem I am seeing in Resscan on my PC. I've pretty much come to same conclusion as you concerning problems with Wonky (I call her Hagatha!), that something about Hagatha is upsetting Resscan displays. Edf files are not at the root of that problem.

Now, issue two is the data skew you originally brought up back in April, and which I mistakenly mixed in with my display problem last week. That is the issue of an event being shown in Resscan at a time significantly different than the end time of an apnea, right? That is what I'm referring to in the latter part of my post above. What I am seeing when looking at the edf files is a discrepancy in the data recording start time between corresponding BRF and EVE files, and a discrepancy between the recording start time and the time stamp of the EVE files. That time discrepancy is equal to the time shift between the event marker and the apnea that I see in some cases in Resscan. I'm going to take a look at more files later today and tabulate the times. I'll post my findings when I'm done.

[DreamDiver]

For my part, I find it's easier to discuss similar but unique hardware challenges in separate threads in order to give participating members a clearer focus. I know it would help me and perhaps others if you would start a separate thread or continue on a previous thread about the S9 Firmware skew rather continuing to post about it here. It's just making it more difficult for future readers to parse whether we're discussing one issue or the other.

Let's make sure were clear and agree on the issues. One is a time ruler display problem I am seeing in Resscan on my PC. I've pretty much come to same conclusion as you concerning problems with Wonky (I call her Hagatha!), that something about Hagatha is upsetting Resscan displays. Edf files are not at the root of that problem.

Bingo. Windows system restore at least - clean reinstall at worst. Let's give Hagatha McWonky an overhaul.

Now, issue two is the data skew you originally brought up back in April, and which I mistakenly mixed in with my display problem last week. That is the issue of an event being shown in Resscan at a time significantly different than the end time of an apnea, right? That is what I'm referring to in the latter part of my post above. What I am seeing when looking at the edf files is a discrepancy in the data recording start time between corresponding BRF and EVE files, and a discrepancy between the recording start time and the time stamp of the EVE files. That time discrepancy is equal to the time shift between the event marker and the apnea that I see in some cases in Resscan. I'm going to take a look at more files later today and tabulate the times. I'll post my findings when I'm done.

I'm guessing you've searched on this forum for quite a number of threads on the S9 firmware skew problem. If you have, you'll have noticed that we've pretty much gummed it to death. There are quite a number of well-versed coders on this forum who have put their minds to a considerably number of hacks, tweaks and workarounds for this issue. I hope you are successful in figuring out what so many others could not. Please consider promising yourself this: if you cannot make it work in three reasonable attempts, divert your new-found energy and efforts to another Herculean task. Expecting different results from the continually repeating the same procedure isn't anywhere near sound reasoning. In three months someone else will pick up where you left off and maybe shed some light -- or maybe not. Good luck with it.

Please consider renaming the first post of this thread the 'ResScan graph display problem'. It probably has nothing specifically to do with the S9.