[RASMB] Time-stamp checks for SV data

[Borries Demeler]

Hi Glen,

As Walt pointed out, we all need accurate w^2t integral values for
accurate analyses. In an effort to guard against the problem Rodolfo
reported we routinely check the time elapsed in UltraScan to see if we
get *reasonable* values, you can make a plot of t vs. w^2t integral for
each scan and see if you get a reasonable intercept. So, UltraScan users
with ProteomeLab 6, please check the time correction value in the run
details window. However, this approach only works of course if the w^2t
value is correct, and that cannot be checked with a time server. If you
are relying on time from filestamps you run the risk of messing things up
if you copy files and they end up getting assigned new filestamps :-)

A related issue is perhaps how accurate the acceleration setting is.
If you set the acceleration to 400 rpm/sec and accelerate to 60 krpm,
you should reach 60 krpm after 2.5 minutes. Our instrument pretty
much is right on, and that gives you the ability to calculate the w^2t
independently to make sure you get the right w^2t at a given time (our
machine gives the right values). Perhaps someone with Proteomelab 6
could try that? A stopwatch is all you need for this.

Thanks, -Borries


In-Reply-To: <883495.48409.bm at smtp101.vzn.mail.bf1.yahoo.com>
> 
> On Fri, Jan 25, 2013 at 01:14:00PM -0500, Glen Ramsay wrote:
> Greetings All:
> 
> An underlying assumption is that the computer clock is accurate.  Usually
> this is the case, and when it fails, it is because the mother board watch
> battery has died while the computer was turned off (not an experiment
> problem).  Computers connected to the network check their time via a time
> server, but this function only sets the time, it doesn't log time errors.
> Hopefully any small time changes aren't going to be a factor.
> 
> NIST has time servers that can be polled.  In principle the data
> collection software could validate the elapsed time by this means.
> 
> Is the current situation adequate, or do we need to use the time servers?
> My feeling is we don't need the time servers, but I'll throw out the
> idea for discussion.
> 
> Glen
> Aviv  Biomedical, Inc.
> _______________________________________________
> RASMB mailing list
> RASMB at list.rasmb.org
> http://list.rasmb.org/listinfo.cgi/rasmb-rasmb.org