I wonder how the RTC and associated 32kHz crystal interacts with the system time and main oscillator? I assume the time base is driven by the 32kHz crystal not the system clock. In fact I think they are separate; when I run my M100 at 5MHz, I don't count seconds twice as fast.
Steve On Mon, Nov 16, 2026 at 11:06 AM B 9 <[email protected]> wrote: > On Sun, Nov 15, 2026 at 10:04 PM John R. Hogerhuis [email protected] > <https://mailto:[email protected]> wrote: > > >> >> On Sun, Nov 15, 2026 at 6:33 PM B9 <[email protected]> wrote: >> >>> >>> I had never noticed it! It is less than 1kHz off, making the clock 0.03% >>> fast. It takes nearly an hour to be off by one second. >>> >>> >> Wow, that seems like a lot! A second every hour is 2.8 minutes a week. >> But maybe no one ever relied on it, I guess. Why would you when you can >> just call the Time Lady? >> >> I wonder if people just didn't use it or they set it a lot. >> >> Maybe a fun project, add an atomic radio sync to the Model 100 :-) >> >> -- John. >> > You’re right, that does seem like a lot. I think for me, perhaps it’s that > I set it often. I often experiment with programs and accidentally cause a > cold start, time-warping my Tandy 200 back to midnight, January 1st, 1900. > > I checked my second Tandy 200, and it shows f = 2.45836MHz. > [image: f=2.45836MHz. An oscilloscope showing a square wave with a > frequency measurement.] > > > > Name Measured MHz %err > Nominal 2.4576 0.000 > b9 T200 A 2.45841 0.033 > b9 T200 B 2.45836 0.031 > > So my two machines are measuring nearly identically. I recalibrated my > oscilloscope and got the same numbers. > > Did people treat time more casually back in the 1980s? Do crystals drift > over time? Could there be some sort of wave propagation delay because I’m > measuring it at the external system bus connector? > > —b9 >
