Page 2 of 10

Re: DM42- clock losing time

Posted: Thu May 09, 2019 7:22 pm
by cappy
It's really a shame we can't get a reply from the developers on this issue

Re: DM42- clock losing time

Posted: Thu May 09, 2019 11:16 pm
by keithdalby
Issue?

Re: DM42- clock losing time

Posted: Fri May 10, 2019 4:30 am
by rprosperi
cappy wrote:
Thu May 09, 2019 7:22 pm
It's really a shame we can't get a reply from the developers on this issue
It's not clear that there is an issue, just some vague comments that you think it lost some amount of time over some other amount of time; you noted yourself that you didn't even record it.

I suggest you do this, to help characterize if there is an issue:

1. Update to the latest f/w
2. Set the time precisely and record the time that you set it (using a reliable, objective source)
3. Enjoy life for a couple months
4. Come back and check the time on the DM42 vs. the original source, and record the difference

Then come back here and report it. Either the result is there is not enough difference to matter, OR that you've provided a solid case that something is amiss and a basis for SM to investigate.

Re: DM42- clock losing time

Posted: Fri May 10, 2019 5:21 am
by cappy
keithdalby wrote:
Thu May 09, 2019 11:16 pm
Issue?
The clock losing time, after no use for an extended period.

Re: DM42- clock losing time

Posted: Fri May 10, 2019 5:23 am
by cappy
rprosperi wrote:
Fri May 10, 2019 4:30 am
cappy wrote:
Thu May 09, 2019 7:22 pm
It's really a shame we can't get a reply from the developers on this issue
It's not clear that there is an issue, just some vague comments that you think it lost some amount of time over some other amount of time; you noted yourself that you didn't even record it.

I suggest you do this, to help characterize if there is an issue:

1. Update to the latest f/w
2. Set the time precisely and record the time that you set it (using a reliable, objective source)
3. Enjoy life for a couple months
4. Come back and check the time on the DM42 vs. the original source, and record the difference

Then come back here and report it. Either the result is there is not enough difference to matter, OR that you've provided a solid case that something is amiss and a basis for SM to investigate.
Another user just reported the same thing, in this thread.
Given that I had the latest firmware when I reported it, and it’s been months since I posted the original message..
I think there’s enough information there for people with multiple DM-42’s to do the experiment and report.

Re: DM42- clock losing time

Posted: Fri May 10, 2019 6:18 am
by keithdalby
Are we talking a few seconds different per month? A few minutes? A few hours? A few days?

It's a calculator, not a precision time piece. A little lost time is not an issue.

Re: DM42- clock losing time

Posted: Fri May 10, 2019 7:30 am
by cappy
keithdalby wrote:
Fri May 10, 2019 6:18 am
Are we talking a few seconds different per month? A few minutes? A few hours? A few days?

It's a calculator, not a precision time piece. A little lost time is not an issue.
Many minutes over several months.

There are precision clock sources driving most microcontrollers (some use the cheesy resonators), there’s no excuse for that much loss of time especially in such a short period. I wouldn’t expect that much Lost time until several years have lapsed.

Re: DM42- clock losing time

Posted: Fri May 10, 2019 1:05 pm
by Walter
I can confirm that the DM42 clock has a precision in the order of minutes, not seconds. My DM42 is 11 minutes ahead of true time today - I adjusted it several months ago (sorry I didn't record that date).

I concur it's a calculator, not a clock, even less a precision clock. Nevertheless (and as mentioned in post #2 of this thread already), this precision is not conforming with a quartz-controlled clock which is standard nowadays even at low cost devices.

And it bears the name "SwissMicros" which certainly sets expectations ... ;)

Re: DM42- clock losing time

Posted: Fri May 10, 2019 6:31 pm
by cappy
Walter wrote:
Fri May 10, 2019 1:05 pm
I can confirm that the DM42 clock has a precision in the order of minutes, not seconds. My DM42 is 11 minutes ahead of true time today - I adjusted it several months ago (sorry I didn't record that date).

I concur it's a calculator, not a clock, even less a precision clock. Nevertheless (and as mentioned in post #2 of this thread already), this precision is not conforming with a quartz-controlled clock which is standard nowadays even at low cost devices.

And it bears the name "SwissMicros" which certainly sets expectations ... ;)
Thanks for confirming this Walter

Re: DM42- clock losing time

Posted: Fri May 10, 2019 8:04 pm
by Jebem
Despite using a 32768khz reference device (be it a crystal or a resonator or whatever they are using on it), the time depended on two factors:

1. Hardware oscillator circuitry precision. The hardware design, namely the load components used affect the frequency that will not be exactly 32768khz if the project was not covering this clock "feature". Also the crystal/resonator type will have its own precision tolerances, namely to temperature changes.
So to have a precision clock it requires adequate hardware design to support that "feature" in first place.

2. Software design. As already mentioned, documentation exists to be followed that covers the different processor states in order to maintain good accuracy, as long as the hardware oscillator precision can be trusted.

I don't remember to see any adverts from the manufacturer concerning the precision of the clock but i might be wrong.