Can Cumulus damage the Instromet Data Logger.

[steve]

That's looking good, there were the correct number of logger entries available, so it looks like the logger is behaving now.

[fdhugo]

I have just found that I cannot access the PC remotely.
I suspect that it has slept again.
If so reset the data logger again?

[steve]

You shouldn't need to reset the logger. The problem you will have is that when Cumulus wakes up, it will carry on from that point and if it makes an entry in the log files, you will have a gap in your data. You can't do the usual Cumulus 'rewind' because Instromet won't let me use the command which moves the logger pointer backwards. If you can stop Cumulus immediately after the PC awakes (before it has chance to write the next log entry), you might get away without a gap in the data.

[fdhugo]

To be clear?
Cumulus running.
PC Sleeps.
Wake PC just after ##:10 interval.
Close Cumulus.
Open Cumulus, should transfer data for Sleep period, then continue at next ##:10 interval.

My memory is that Cumulus was frozen after Sleep, so I Closed and Opened it anyway, but that was before resetting the data logger.

[steve]

To be clear?
Cumulus running.
PC Sleeps.
Wake PC just after ##:10 interval.
Close Cumulus.
Open Cumulus, should transfer data for Sleep period, then continue at next ##:10 interval.

Yes, as long as it doesn't get to write a new logger entry and update today.ini, it should think that the last time it was running was before the PC went to sleep.

My memory is that Cumulus was frozen after Sleep, so I Closed and Opened it anyway.

It's possible that the PC going to sleep caused a problem with the connection to the station; this is one reason why handling sleeping reliably would be non-trivial.

[fdhugo]

The PCs have not Slept overnight, so perhaps SS U has solved that problem.
We shall see.
Thank you.

[fdhugo]

Spoke too soon.
PC Slept.
Woke PC.
Closed Cumulus.
Opened Cumulus.
Longish data transfer shown.
No record in log of data during Sleep.
Yesterday items on display showing values later than last log entry.

[steve]

You started Cumulus at 16:53, and it had last logged an entry at 10:10

30/07/2015 16:53:16.357 : Read today file: Date = 30/07/2015, LastUpdateTime = 30/07/2015 10:10:00, Month = 7

The logger said it had 40 entries available, that seems about right for approx 6.5 hours at 10 minute intervals.

The first entry the logger sent is correctly the first one after 10:10

Code: Select all

30/07/2015 16:53:20.030 : Logs available = 40
30/07/2015 16:53:20.030 : Number of history records = 40
30/07/2015 16:53:20.030 : Get next log - RDLG,1
30/07/2015 16:53:21.030 : response=RDLG,1
rdlg,218885,1,600,10:15:48,30/07/2015,13.7,14.4,13.9,,,,50.9,54.7,52.7,1021,1021,1021,0.9,6.6,3.3,351,122.13,130.54,219


>
30/07/2015 16:53:21.030 : data=rdlg,218885,1,600,10:15:48,30/07/2015,13.7,14.4,13.9,,,,50.9,54.7,52.7,1021,1021,1021,0.9,6.6,3.3,351,122.13,130.54,219
30/07/2015 16:53:21.030 : Logger entry zero: Y = 2015, M = 7, D = 30, H = 10, M = 15, S = 48
30/07/2015 16:53:21.030 : Earliest timestamp 30/07/2015 10:15:48
30/07/2015 16:53:21.030 : Actual number of valid history records = 40

It then sent entries for 10:25 to 11:05 at 10 minute intervals, all fine. But the next entry it sent was for 08:21 on the 29th July:

Code: Select all

30/07/2015 16:53:34.468 : Get next log - RDLG,1
30/07/2015 16:53:35.468 : response=RDLG,1
Erdlg,218729,1,600,08:21:09,29/07/2015,12.4,13.4,12.8,,,,63.8,66.9,65.6,1014,1014,1014,1.4,6.5,4.0,315,114.67,129.02,199


>
30/07/2015 16:53:35.468 : data=rdlg,218729,1,600,08:21:09,29/07/2015,12.4,13.4,12.8,,,,63.8,66.9,65.6,1014,1014,1014,1.4,6.5,4.0,315,114.67,129.02,199
30/07/2015 16:53:36.468 : Processing logger data entry 6 for 29/07/2015 08:21:09

It continued sending logs from that time, in sequence, up until 13:30 on the 29th:

Code: Select all

30/07/2015 16:54:39.010 : Get next log - RDLG,1
30/07/2015 16:54:40.010 : response=RDLG,1
rdlg,218760,1,600,13:30:05,29/07/2015,15.0,15.7,15.3,,,,54.4,60.0,57.3,1015,1015,1015,2.0,8.4,4.8,339,115.79,129.07,207


>
30/07/2015 16:54:40.010 : data=rdlg,218760,1,600,13:30:05,29/07/2015,15.0,15.7,15.3,,,,54.4,60.0,57.3,1015,1015,1015,2.0,8.4,4.8,339,115.79,129.07,207
30/07/2015 16:54:41.010 : Processing logger data entry 37 for 29/07/2015 13:30:05

The next log it sent was for 08:21 on the 29th:

Code: Select all

30/07/2015 16:54:41.137 : Get next log - RDLG,1
30/07/2015 16:54:42.137 : response=RDLG,1
Erdlg,218729,1,600,08:21:09,29/07/2015,12.4,13.4,12.8,,,,63.8,66.9,65.6,1014,1014,1014,1.4,6.5,4.0,315,114.67,129.02,199


>
30/07/2015 16:54:42.137 : data=rdlg,218729,1,600,08:21:09,29/07/2015,12.4,13.4,12.8,,,,63.8,66.9,65.6,1014,1014,1014,1.4,6.5,4.0,315,114.67,129.02,199
30/07/2015 16:54:43.137 : Processing logger data entry 38 for 29/07/2015 08:21:09

The 40th and final log was the one following, which was for 08:31 on the 29th:

Code: Select all

30/07/2015 16:54:43.285 : Get next log - RDLG,1
30/07/2015 16:54:44.285 : response=RDLG,1
rdlg,218730,1,600,08:31:07,29/07/2015,12.8,13.6,13.3,,,,61.3,66.2,63.7,1014,1014,1014,1.4,8.3,4.4,302,114.72,129.02,225


>
30/07/2015 16:54:44.285 : data=rdlg,218730,1,600,08:31:07,29/07/2015,12.8,13.6,13.3,,,,61.3,66.2,63.7,1014,1014,1014,1.4,8.3,4.4,302,114.72,129.02,225
30/07/2015 16:54:45.285 : Processing logger data entry 39 for 29/07/2015 08:31:07

It looks like those are the actual entries for those timestamps, i.e it's not that the logger was using incorrect timestamps for the latest data. So perhaps whatever it is is that happens when the PC sleeps is causing the logger to go awry, but that's just speculation based on this set of evidence. Note how the sequence numbers of the logger entries aren't in sequence.

Could you attach the diags file from just before the one you uploaded, please?

By the way - as you don't seem to able to stop the PC sleeping, you could set the 'close on suspend' setting (make sure you don't also have 'confirm shutdown' set). Then it will attempt to close itself tidily when the PC goes to sleep, but this isn't guaranteed. You really need to stop the PC sleeping in the first place.

[fdhugo]

Today's attached.
Data has stopped transferring when Cumulus is closed normally.
I'm going to try resetting the data logger again.

[fdhugo]

It has worked...

Cumulus Closed.
Logger reset.
Close on Suspend set
Cumulus Opened.
Logging...
>11:30 UT Cumulus Closed.
>12:00 UT Cumulus Opened.
Data transferred.
12:10 UT Logging...
PC Sleeps with Cumulus running.
Cumulus Closing Down message shown.
>12:30 UT PC Woken.
Cumulus window displayed, frozen.
Task Manager, Cumulus not present, not running.
Cumulus Opened.
Data transferred.
Data present in file for Sleep period.
12:40 UT Logging...

I'll try it again later...

We're required to shut PCs down over the weekend for network maintenance.
I'll Close Cumulus normally.

[fdhugo]

The Diags file

[fdhugo]

2015-08-03.
Started PC.
Opened Cumulus.
Data transferred, record complete, though with unusual times, not precise ten minutes'.
Closed Cumulus, fearing Sleep.
2015-08-04.
PC had not Slept.
Opened Cumulus.
Data transfer box open for a long time.
Closed Cumulus while box still open.
No new data in log.
Opened Cumulus.
Data transfer box open for a long time.
Left it so and did other things.
Box closed, Cumulus display normal, two new entries in log, nothing for the period of closure.
Diags attached.

[steve]

Could you zip up the entire diags folder and attach it please, it's difficult to work out exactly what's going on without seeing what happened previously.

But from this file, it does look the same as before, the logger is sending old data, and it's sending it out of sequence.

[fdhugo]

OK.
I will when I can.
I suspect that Closing it while the data transfer box was open was not a good idea, but it had been open for several minutes with no sign of transfer.
I have lost remote access to the PC, I suspect that it is Sleeping.
I had closed Cumulus while I still had access.
I suspect a reset is required.

[steve]

One thing I've noticed is that when the logger suddenly changes sequence and starts sending older data, there is a spurious "E" in the response:

Code: Select all

04/08/2015 06:38:52.700 : response=RDLG,1
Erdlg,219035,1,600,11:10:35,31/07/2015,15.5,16.8,16.1,,,,35.1,58.6,50.3,1020,1020,1020,0.0,6.3,4.0,162,136.87,130.55,245


>

I don't know what this means, and Cumulus ignores it - the fact that the logger echoes the command isn't even mentioned in the documentation, so Cumulus has to find the actual start of the reply, the part from the lower case rdlg onwards. But the spurious "E" is apparently in some way related to the problem.