[ntp:questions] How bad is USB?

David J Taylor david-taylor at blueyonder.not-this-part.nor-this.co.uk.invalid
Fri May 1 13:11:07 UTC 2009

Richard B. Gilbert wrote:
> You can try it and see what happens!  The results may be sufficiently
> good for your purposes.  You will almost certainly not get microsecond
> accuracy.  If you are willing to settle for +/- 10 milliseconds you
> can almost certainly get that.
>  There are (at least) two groups of people who hang out here:
> a. The "chimeheads" in pursuit of the "one true time", and
> b. The "synchs" who don't really care what time it is as long as all
> their machines have the SAME time.
> Both groups use the same tool, NTPD.  It turns out that having a
> stable and accurate source of time makes it much easier to get a
> whole herd of machines to agree on what time it is.


Yes, I /could/ try it, but I was rather hoping that someone else already 
had.  I also think that it will be somewhere between 1 microsecond and 10 
milliseconds, but just where?

I'm in a third category (probably) - as long as my PCs "sound" to have the 
same time, I'm happy.  Actually, I want UTC as well, but as I have a 
speaking clock working on two PCs it's "nice" when they both speak at the 
same time.  I wonder how close that needs to be - a few milliseconds, 
perhaps?  Here's what I actually get - all with Windows.


- within about 250 microseconds on my serial/PPS ref-clock PCs
- within about 2 milliseconds on my "Synced from ref-clock" PCs
- within about 25 milliseconds on my Vista PC.

Yes, I used to get better on the ref-clock PCs when running FreeBSD, but 
I've since been able to shut down that PC and just use a Windows box 
instead.  It's good enough for me.


More information about the questions mailing list