Re: Weibull Distribution but with a varied cycle time

Just from experience, a minimum of 12 - 16 data for creating a Weibull
plot. This truly is a minimum, but it's a good place to begin.

One point I wandered around but did not dig into. The slope of the
line on a Weibull plot can be used to tell us whether the failure rate
is increasing or decreasing as the device/system ages. This is
important. If the failure rate is decreasing with age then there's a
sense of "burning in" by removing the weak ones. If the failure rate
is increasing with age then there's a sense of "wearing out". For
many electronic devices (but not necessarily total systems) the
failure rate is constant or almost constant after burning in (or
burning out, if you prefer) the weakest ones. Weibull paper provides
an easy way of finding the slope of the line along with printed
guidelines to help in determining whether the failure rate is
decreasing or increasing. This is one reason I suggested going
directly to valid Weibull graph paper.

Re: Software for plotting. By this I assume you mean not just creating
blank Weibull graph paper, but plotting the points on the Weibull
plot. There's probably commercial software that does this. If this
sort of activity is to be done just now-and-then, it's hardly worth
buying software and learning to use it.
If it's something you propose to do daily, then software may be
useful.

I use Kaleidagraph for making plots of data, especially probability
plots. Weibull plots are one of many kinds of probability plots. My
Kaleidagraph software (on an old Macintosh) is fairly old. It doesn't
do Weibull plots. Perhaps the most recent version of Kaleidagraph will
do that...?? There may be other choices.

There are many kinds of probability plots (different scales, different
axes) other than binomial, normal, log normal, Weibull, etc. It's
unlikely that anyone has created software that makes plots for all of
these.

Good luck... OMU

On Jan 27, 6:32 pm, "googlinggoog...@xxxxxxxxxxx"
<googlinggoog...@xxxxxxxxxxx> wrote:
Thank you for your kind offer, I will be in touch if i need further
reference material. I believe I have all of this in my undergrad
notes or at least the basics and so I will have a dig for that first
before bothering you!

One last question really concerns how many times is a good amount of
times to test to get suitable data? Obviously the more data I collect
the more representative the results will be. But realistically there
needs to be a cap on the amount of time I run my tests for.

I'm quite looking forward to this now. Another thing can you recommend
a way to produce weibull plots on computer? as you mention excel is
not the way to go. I will do all my analyse by hand as you mention to
start with how ever.

Thanks

David

.

Relevant Pages

  • Re: calcuating goodness of fit in the statistical tool box
    ... Peter Perkins wrote ... (I used normal, lognormal, and Weibull ... make plots of the ... Thank you for the confirmation, ...
    (comp.soft-sys.matlab)
  • Re: points on a trendline
    ... If it's a straight line, you can use INTERCEPT and SLOPE ... you can use LINEST according to Bernard ... >> Some of the plots are not very smooth so I added a trend line. ...
    (microsoft.public.excel.charting)
  • Re: Buying land: What should you look for?
    ... > For the plots I'm considering, the local power lines either border the ... > slope would hide the house from the road. ... > my previous gardens have been on southern slopes. ... not me - still very slowly terracing the parts that aren't ...
    (misc.rural)
  • Re: Buying land: What should you look for?
    ... For the plots I'm considering, the local power lines either border the ... The property tends to slope north (with a few gullies ... my previous gardens have been on southern slopes. ... anyone had any experience in terracing wooded country? ...
    (misc.rural)
  • Re: Compression and PSNR
    ... > Plot PSRN against BPP and compare the plots. ... > my first post even in the curves. ... as differentially (looking at the slope). ... test and rate the streams according to a scale from ... ...
    (sci.image.processing)