[Dxspider-support] Frequency Rounding Error

[Ron Stordahl]

Lee Sawkins wrote:

> I entered several Dx Spots this morning.  All came back 0.1 KHz higher 
> than what I entered.  If I put in a spot for 3507, it came back as 3507.1.
>  
> This is with the latest version of Spider .
>  
> Lee
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
>
>------------------------------------------------------------------------
>
>_______________________________________________
>Dxspider-support mailing list
>Dxspider-support at dxcluster.org
>http://mailman.tobit.co.uk/mailman/listinfo/dxspider-support
>  
>
I asked Dirk to correct an error in rouding, which he did.  I tested thus:

28001.01 thru 28001.04 (50% of the range) rounds as it should to 28001.0
28001.05 thru 28001.09 (50% of the range) rounds as it should to 28001.1

This reduces the average frequency error to 25 Hz.  On the other hand 
AR-Cluster simply truncates by dropping the second digit following the 
decimal point creating an average error of 50 Hz.

I never tried, until now:

28001.0  rounds to 28001.0  (correctly) but
3501.0 rounds incorrectly to 3501.1  !!
3501.01 rounds incorrectly to 3501.1  !!

So there are problems with this when the whole part of the frequency in 
KHz is less than 5 digits apparently.

Ron, N5IN
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.tobit.co.uk/pipermail/dxspider-support/attachments/20060114/75373171/attachment.html