Sven clock vs mechanical clock in accurace

More
2 years 9 months ago #6688 by Pacman223
Let me start off with some back ground. My lovely wife decided to give me a new watch for my birthday - a Laco Aachen.

shop.laco.de/en/product/Aachen.html?info=82

(a replica of their WWII German pilots watch).

VERY NICE watch Absolutely love it!! If anyone is thinking of owning a nice mechanical watch, this one is really nice, and it comes with some great history. Laco built Pilot watches for Germany during WWII.
(Other companies that make good WWII German pilot versions are Stowa and IWC)


So I thought about seeing how well it keeps time vs my Sven clock. I have never seen anyone document how well a Nixie clock actually keeps time.

So, I kept track for a couple of weeks of how each did (averaging the weekends while I wasn't at work).

In short..
New mechanical Laco Watch's two week average = +8.27 sec/day
Sven Nixie clock = +.06 sec / week (supper accurate in my book)

I found the Nixie clock was STUNNINGLY accurate. I didn't actually realize how well it was doing.

I read that Mechanical watches, may need about 2 weeks of running before they settle down a bit.
(I will continue to track this and see if thats really true or not)

Each morning I compared the time a web page that claims to be very accurate.


(on a side note... the jump for Sven to +1 sec on 7/1/2015 - I believe this was caused by my office being really hot that day (around 83 F). That day was REALLY hot outside 105+F and it felt like our buildings AC was struggling to keep up, My Sven clock reported temperatures of around 84-86 F. Most days its usually around 78 F)

thought I might pass this info along, might be interesting for someone wondering how accurate Nixie clocks /Sven clock really are. I might try this with some other clocks.

Also, has anyone else ever thought to record how accurate their Nixie clocks are? do temperature compensated quartz crystals really improve the accuracy? Do GPS connected clocks really stay perfectly accurate?
Attachments:

Please Log in or Create an account to join the conversation.

More
2 years 9 months ago #6690 by Ty_Eeberfest
Very nice watch! I like...

Funny coincidence that your Sven gained 1 second on July 1, since June 30 -> July 1 was a leap second. Your clock did an "anti-leap" second instead. :laugh:

I check accuracy against WWV once in a while and if I'm off by more than 1 second a month I'm dissatisfied and start playing with the "tweak" factors in code. My ITS1-A clock from Mr. Nixie has the tweak factor available as a settable parameter. My original designs have an internal tweak factor that I change by recompiling and reflashing, not ideal but it's just for me - not a commercial product - so it's fine.

My personal opinion is that TCXOs are a good thing and I use them in all my originals. Since they are one-offs, not commercial products, I don't consider the extra cost to be a problem. Are they more accurate? We could probably have a good argument about that question and debate the definition of accuracy, What I can say is that they are very consistent - the frequency does not wander around like plain crystals do. I find that I still need to set a tweak factor with TCXOs but once the factor is correct the clock runs accurate to within about 2 seconds per year.

Look into it later when the dust is clearing off the crater.

Please Log in or Create an account to join the conversation.

More
2 years 9 months ago #6692 by Pacman223
a leap second would explain the jump for a single day. :)

I will have to fix my numbers to reflect that.

Please Log in or Create an account to join the conversation.

More
2 years 9 months ago - 2 years 9 months ago #6693 by Ty_Eeberfest
I had to think about that one for a second... lol. I guess if your source of "precise time" adjusted for the leap second and your Sven did not (it's not GPS disciplined, right?) then it would look like the Sven gained a second.

BTW, I wonder how precise your "precise time" source actually is. There are methods that compensate for the inevitable net-lag but they aren't perfect. I'm sitting here looking at NIST time ( time.gov ) right now, it says it is compensated for net-lag, and yet it's nearly a second off from my GPS disciplined clock. My clock's Sync LED is telling me it has received a GPS update something less than 2 hours ago (probably much more recently but the LED does not begin signalling "GPS fail" until 2 hours have passed with no updates) so I think it can be trusted. Besides, this clock does better than 1 second per week when running with no GPS at all.

Look into it later when the dust is clearing off the crater.
Last Edit: 2 years 9 months ago by Ty_Eeberfest. Reason: Typos

Please Log in or Create an account to join the conversation.

More
2 years 9 months ago #6699 by Pacman223
I used

utctime.org/

as the correct time. but you are correct, I really don't know the delay to my computer, or how reliable it is.

but I figured that is my best bet... I thought about trying to use one of our servers at work to measure against, but then I remember one time where one of our servers was a few minutes off causing all sorts of problems. :)

Please Log in or Create an account to join the conversation.

More
2 years 8 months ago - 2 years 8 months ago #6809 by Pacman223
Update...

I decided to continue tracking the accuracy of watch and now my newer nixie clock (with the txco).

for week 1. it hasn't lost/gained any time, at least as far as I can tell (just using the web page and looking at the time of the clock vs web page). It appears to be at least as accurate as the sven clock, we will see if over time its better :)

I'm also keeping track of my watch, and seems to have stabilized to around +7 sec/day.

I'll keep track and post my observations.
Last Edit: 2 years 8 months ago by Pacman223.

Please Log in or Create an account to join the conversation.

More
2 years 7 months ago - 2 years 7 months ago #6872 by Pacman223
just a small update.

I had the new clock running for about two weeks (new clock uses a temperature compensated crystal), calculated out how much slower it was running, and sped it up ( changing the clocks speed in its software). I have now recorded almost two full weeks after being calibrated, since the calibration, its kept absolutely perfect time. its really appears to be perfect as compared to the reference time that i'm using. I find that amazing.

so I thought that my Sven was really good, it seems to me, its likely going to be around +6 sec / year. but TXCO clock might go less then +2 sec/year. I'll keep recording its time difference and see.

Answer to the question, is the TXCO worth the extra cost.. I think so. I think very accurate clocks are more interesting then less accurate clocks... with that said.. I think Sven is plenty accurate enough for almost anyone and my ITS1 clock seems to be just as accurate, which makes sense, as its similar. I record the same thing for my QTC clock, but I suspect its very similar to the Sven clock. I really appreciate that you can speed up or slow down the clocks via a setting in its software.

I will attach a graph of the new clock when I remember to get it from work :) (Every morning when I start work, I record the time difference from the reference time on the internet).
Last Edit: 2 years 7 months ago by Pacman223.

Please Log in or Create an account to join the conversation.

More
2 years 7 months ago #6957 by Ty_Eeberfest
Not too surprising to me that the TCXO clock is that accurate. In my experience TCXOs are extremely stable over long periods of time. I share your interest in "super" accuracy so of course I consider them to be worth the money when designing & building for myself. If I was making a product for sale I might not use TCXO since I'd be competing with products that have been engineered to meet a price point rather than engineered and then priced.

Look into it later when the dust is clearing off the crater.

Please Log in or Create an account to join the conversation.

More
2 years 6 months ago - 2 years 6 months ago #7008 by Pacman223
I'm really impressed with TXCO now.. So far I have been running the clock for 1 1/2 months, and it seems to be within 1/10 of a sec. I would guess at this rate its not going to be off more than .2 sec a year.

I honestly can't decide if its any different from the time I'm checking it against.

I had no idea I could make a clock this accurate. at this level of accuracy, I'm not sure that a GPS receiver would help, maybe after a year or so of running it might be a tad closer to perfect time.
Last Edit: 2 years 6 months ago by Pacman223.

Please Log in or Create an account to join the conversation.

More
2 years 4 months ago - 2 years 4 months ago #7044 by Pacman223
I haven't seen much posted, I thought I would give a small update on my TXCO clock.

At 100 days since its been running its now about .25 sec fast. Its really difficult for me to eye exactly how fast it is, but something close to that. that makes it about 3/4 sec - 1 sec fast per year.

I think that is REALLY good.

attached is a small

picture of it on my desk at work, next to a brand new Sven clock that I made as a Christmas present
Attachments:
Last Edit: 2 years 4 months ago by Pacman223.

Please Log in or Create an account to join the conversation.

Moderators: AccutronTy_EeberfestIan
Time to create page: 0.045 seconds

Search

Tube Suppliers

Go to top
JSN Boot template designed by JoomlaShine.com