[Nath!]

Hello all! I am now the proud owner of the Grand Seiko Snowflake SBGA211! And I love it!

Something that is concerning me though. This being my first spring drive watch, I am not sure how it works or anything about it, and the quoted accuracy is +/-1 second per day.

More or less 24 hours after I purchased it, it was 2 seconds ahead. I reset it back so it was accurate again. For a day/day and a half it was 0.5-1.5 seconds slow, kind of going between that range, until yesterday when it was 1.5/2.5 seconds behind. Woke up this morning and it is now 1 second ahead.

Is this normal? Because that means in the 7 hours I was asleep, it gained 3.5 seconds. But on the whole, that means it gained 1.5 seconds in 3 days, or if I count the day I bought it when It was accurate, that means it gained 1 second over 3 days. It also doesn't help me worrying about it when I see posts that other peoples Snowflakes are still on the dot accurate after a week, if not 0.2 seconds fast or slow. just not sure if it should be fluctuating like this or not.

Just looking for some insight and what have you as I haven't owned a spring drive watch before! Many thanks!

 

[BrianBinFL]

By what means are you checking the sync?

 

[Nath!]

By what means are you checking the sync?

Just an app on my phone that is the correct time down to the seconds. I am, checking it daily, sometimes more than daily, but I have been told to not check it daily and to wait a week or a month and then check. Hodinkee. So it isn't an actual watch app that tracks how accurate it is, but I set the watch to be exactly right on sunday and this is the result.

When I got home from work around 2 hours ago it was still 1-1.5 seconds ahead, but at the time of posting this reply, it is now under a second ahead. I know the mechanism inside slows it down or speeds it up (?) as and when it is needed, I just didn't think it would fluctuate this much on a day to day basis.

I did see a post where it is better to check the accuracy of a spring drive movement over a longer period of time, like a week or month.

 

[BrianBinFL]

Just an app on my phone that is the correct time down to the seconds. I am, checking it daily, sometimes more than daily, but I have been told to not check it daily and to wait a week or a month and then check. Hodinkee. So it isn't an actual watch app that tracks how accurate it is, but I set the watch to be exactly right on sunday and this is the result.

When I got home from work around 2 hours ago it was still 1-1.5 seconds ahead, but at the time of posting this reply, it is now under a second ahead. I know the mechanism inside slows it down or speeds it up (?) as and when it is needed, I just didn't think it would fluctuate this much on a day to day basis.

I did see a post where it is better to check the accuracy of a spring drive movement over a longer period of time, like a week or month.

You would think this would be a reliable method, but it is not. The time on your phone is wildly inaccurate - at least for this purpose. If you compare the time your phone has to the actual time you will find that sometimes it's right to within a tenth of a second, and sometimes it's off by 1.5 seconds, or more.

If you want to know how wrong your phone is then open a browser on your phone and go to Time.is - exact time, any time zone. It will tell you how "off" your clock is.

Doing precise accuracy assessments is only possible if your reference time source is both accurate and precise. The phone clock is neither.

 

[Nath!]

Oh wow, that I did not know! Thank you for the useful information!

You would think this would be a reliable method, but it is not. The time on your phone is wildly inaccurate - at least for this purpose. If you compare the time your phone has to the actual time you will find that sometimes it's right to within a tenth of a second, and sometimes it's off by 1.5 seconds, or more.

If you want to know how wrong your phone is then open a browser on your phone and go to Time.is - exact time, any time zone. It will tell you how "off" your clock is.

Doing precise accuracy assessments is only possible if your reference time source is both accurate and precise. The phone clock is neither.

Oh wow, thank you! I didn't know this! So the accuracy of mobile time and time related apps can fluctuate. How interesting! Well according to that time on Time.is, it is almost completely spot on. I shall see how it goes for the next few hours or the next week using this website and report back! I wonder how inaccurate my phone time was when I set the watch on sunday.

I would hate for this watch to not perform as it should, as I love it so much, I would hate to have to send it away!

 

[BrianBinFL]

Oh wow, thank you! I didn't know this! So the accuracy of mobile time and time related apps can fluctuate. How interesting! Well according to that time on Time.is, it is almost completely spot on. I shall see how it goes for the next few hours or the next week using this website and report back! I wonder how inaccurate my phone time was when I set the watch on sunday.

I would hate for this watch to not perform as it should, as I love it so much, I would hate to have to send it away!

That's the problem with using the time on your phone. Sometimes it will be spot-on, and sometimes it'll be off by a good bit. I suppose it probably depends on how long it has been since the phone refreshed its time sync with the network.

I have a live-feed-GPS clock that's accurate to 100 micro-seconds or less, and that's what I use for my sync checking. I have photo-compared that source against time.is and nist.time.gov and have found those two sources to be reliably accurate.

If you have a rooted Android phone there are apps out there that will allow you to force sync the operating system's clock to an NTP time source. You could do that just prior to taking time measurements with your app and that might make the app usable. Generally I think those apps have A LOT of problems because they rely upon you having a consistent reaction time in pushing a button, they rely on phone time (which is unreliable), etc.

For a mechanical watch that will be wrong by several seconds after several days those differences might not matter much. But for a quartz or Spring Drive movement those differences are a problem.

 

[Nath!]

That's the problem with using the time on your phone. Sometimes it will be spot-on, and sometimes it'll be off by a good bit. I suppose it probably depends on how long it has been since the phone refreshed its time sync with the network.

I have a live-feed-GPS clock that's accurate to 100 micro-seconds or less, and that's what I use for my sync checking. I have photo-compared that source against time.is and nist.time.gov and have found those two sources to be reliably accurate.

If you have a rooted Android phone there are apps out there that will allow you to force sync the operating system's clock to an NTP time source. You could do that just prior to taking time measurements with your app and that might make the app usable. Generally I think those apps have A LOT of problems because they rely upon you having a consistent reaction time in pushing a button, they rely on phone time (which is unreliable), etc.

For a mechanical watch that will be wrong by several seconds after several days those differences might not matter much. But for an quarts or Spring Drive movements those differences are a problem.

Thank you for your help! I was superbly worried that it wasn't as accurate as it should be, but it turns out my phone and PC clock is innacurate. That is quit the compliment to the watch! I love this watch so much!

 

[BrianBinFL]

Thank you for your help! I was superbly worried that it wasn't as accurate as it should be, but it turns out my phone and PC clock is innacurate. That is quit the compliment to the watch! I love this watch so much!

You're welcome. Pass the wisdom along to someone else when you can. I think a lot of folks fret about their watch being inaccurate when in fact the problem is with their reference source that they mistakenly thought was accurate.

 

[dealta66]

Brian is absolutely correct. Your computer and phone generally sync to time.gov or whatever every so often. I don't use those apps...I just references the time to time.gov. Enjoy the Spring Drive, the accuracy is incredible.

 

[Trog7074]

Time.is only on my iPhone which has been excellent.

My Snowflake has no been running continuous for 13 months at +0.25 s/d which is precisely what it was out of the box. I set it -30 secs and allow it to run to +30 secs thus only requiring resetting every 8 months.

 

[Nath!]

Brian is absolutely correct. Your computer and phone generally sync to time.gov or whatever every so often. I don't use those apps...I just references the time to time.gov. Enjoy the Spring Drive, the accuracy is incredible.

Oh I will enjoy it, thank you! I have a few other watches that I also love, and have worn nothing but the Snowflake since I bought it 5 days ago. And turns out yeah, it is still spot on accurate 5 days after the purchase. I must have caught my phone time when it was also spot on accurate as I used that to set it on sunday. Time.is literally saved me a heap of worrying!

 

[Trog7074]

Resurrecting this topic rather than create new.

My Snowflake had been running at +0.25 s/d for 16 months when I decided to cease winding it while other projects were addressed. Since mid November 2022 up until mid January 2023 it was stopped. 5 days ago I fully wound it and it's been on my WW since where I'm maintaining at least 90% full power. I would have expected it to have gained 1 second at least but it's still smack on the second. Let's see what it's at after 30 days.

13 days on and the seconds hand is still hitting the marker. If anything I would say -0.25 seconds after 13 days.

Ohhh and my Secret Santa request was for a double Klarstein WW which I have to say is far better than the first version. This one actually stops with both watches vertical 12/6 whereas my first one both winders had different rotational speeds and both stopped anywhere around the 360.

 

[Trog7074]

30 days and +1 second.