[time-nuts] Maxim DS1342

[Russ Ramirez]

> hmurray at megapathdsl.net said:
>
> A 32 KHz clock is also useful for running (well, crawling) in low power
> mode.
>  That can be important when running off batteries.
>
> I suspect the reason that chips don't have an option to drive the RTC from
> a
> 1 PPS input is that there is no volume in that corner of the market.  Does
> anybody other than a time-geek even know what PPS means?  (There is a wiki
> page, actually 2, a disambiguation page for PPS and a another page for
> Pulse
> Per Second.)
>
> The product planners for single chip CPUs systems have a list of important
> customers.   If you are buying a million chips a year, they will make sure
> that one of their next generation of chips is a good fit for your
> application.  "Good fit" means it gets the job done and has a reasonable
> cost.
>
> If you are a startup with a sensible business plan that turns into a
> million
> chips a year, they will help you select one of their chips to get off the
> ground and discuss the good-fit/price part when you get closer to placing a
> large order.
>
> My million-a-year may be off by a factor of 10 (or more) either way, but I
> think it's the right ballpark.
>

I think the volume argument is sound for sure, but OTOH what is required to
implement a single multi-purpose I/O as an external 1 Hz input is a handful
of transistors. It could be a differentiating feature to some vendors, and
in fact it is not unique to Maxim in terms of RTC chips. For CPU chips
specifically, volume would be an issue and lack of big client requests for
this as an embedded function is likely.

I also agree that the required precision is an issue, and this is manifest
in Microchip's RTC chips where they have calibration circuitry to
compensate for slight 32 kHz crystal differences if so desired. Obviously
the time nuts market is even smaller still :-)

Thanks for your thoughts.

Russ