WatchUSeek Watch Forums banner

241 - 260 of 361 Posts

·
Registered
Joined
·
3,107 Posts
Discussion Starter #241
Small update ... and bug

It seems that I have stumbled on a small timezone-related bug which was there for quite some time but the lack of more serious testing was 'efficiently' hiding it.

The bug is present when you apply a large POSITIVE timezone correction - for instance when going from NYC to ATHN, but NOT on the reverse correction (where the negative correction made me do things 'the right way' from the start - plus that I am on ATHN timezone and I check very often the NYC timezone, so that one was seriously tested ;-)). The bug also results in wrong numbers for other formulas that involve such corrections - for instance calculating the sidereal time from NYC (but not from ATHN :-d) - since that involves calculating UTC from the local time.

For most of this week I will be away (unfortunately mostly 'job-hunting' :roll:) but I hope to have a fix for that bug over the weekend and to also be able to release some other stuff (that I was hoping to incorporate into the WorldTime watch but the lack of memory space made that impossible :-()
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #242 (Edited)
OK, the bug was bugging me

It seems that I have stumbled on a small timezone-related bug which was there for quite some time but the lack of more serious testing was 'efficiently' hiding it.

The bug is present when you apply a large POSITIVE timezone correction - for instance when going from NYC to ATHN, but NOT on the reverse correction (where the negative correction made me do things 'the right way' from the start - plus that I am on ATHN timezone and I check very often the NYC timezone, so that one was seriously tested ;-)). The bug also results in wrong numbers for other formulas that involve such corrections - for instance calculating the sidereal time from NYC (but not from ATHN :-d) - since that involves calculating UTC from the local time.
...
OK, the bug was bugging me - so yesterday I managed to have a workaround and also managed to squeeze in the code a way to edit the longitude for which Local Mean Sidereal Time (LMST) is calculated :-! (but I had to cut all the 'debugging stuff' I could find - for instance the 'baro alarm test' is no longer there - but the baro alarm should actually work better since I also fixed a small bug there).

Anyway - here is the latest Advanced WorldTime SportsWatch firmware:

ez430_chronos_worldtime2rc4.zip

Instruction from:

https://www.watchuseek.com/f9/ezchr...roject-start-498121-post4135648.html#poststop

and:

https://www.watchuseek.com/f9/ezchr...roject-start-498121-post4154803.html#poststop

still apply - and on top of that when you are in SIDEReal mode you can now enter the 'setting mode' (with a long-press on the corresponding bottom-left button) and edit the longitude (degrees East are 'normal' but West might appear strange since the arrows work 'the reverse way' - but that is just since mathematically those are negative numbers in our convention).

Please let me know if now the sidereal time matches (with at most 1-2 seconds of difference) the one displayed at:

Local Apparent Sidereal Time

(using either 0 / 0 / 0 for GMST or the same longitude you entered in the watch for LMST - you can switch the watch when in SIDER mode with bottom-right button).
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #243
A new flavor

Today we have another 'firmware branch' - the "CountDown and Interval Timer" b-)

Here is again a quick map on how the final 'flavors' seem to be heading - remember that ALL have the HEQ/TC part inside:

- the "Data-Logger" - only choice that can keep a log and be used to determine the TC curve; still has bluerobin inside;

- the "BlueRobin Base StopWatch" - it has the basic stopwatch (and all the features in the original stopwatch firmware from Texas Instruments) + timezones and manual DST; probably the default choice if you still MUST use bluerobin with some chest-band for heart-rate;

- the "WorldTime Advanced StopWatch" - it does no longer have bluerobin, but it has automatic DST and full-worldtime information (with the possibility of two timezones shown at the same time), plus minute repeater plus sidereal time plus stopwatch with split-time plus extended alarms plus altitude log and barometric alarm; probably the general 'best compromise' if you are not interested in very specific features;

- the "CountDown and Interval Timer" is a new module that I was hoping to add to the above branch, but when finally ready that one was just too big - so here we have a new "sports and regatta-oriented" flavor which 'trades' the minute repeater and the sidereal time from above for (up to) 9 countdown timers which can be chained with one another or with the start of the advanced stopwatch and can be configured to generate sound while counting or at the very end of the interval - see below for more details;

- it seems that there will also be an "Astronomical Watch" - that one will drop the stopwatch, bluerobin, the full worldtime (but manual timezones and DST will still be there) and all the other new features for a number of advanced astronomical features that I now hope will include also (on top of sidereal time) Equation of Time, Local Solar Transit Time (actual local noon), SunRise and SunSet (on two world-wide locations that can be user-configured and which are also used for sidereal time) plus a Moon Age (that will shatter the Ulysse Nardin current accuracy claims :-d) plus eventually last new moon, next new moon and the full moon in-between!



Now back to the "CountDown and Interval Timer" - the module is displayed on the top line, which is now:

TIME -> ALaRM -> CouNTeRs -> TEMPerature -> ALTitude/Pressure -> TILT

while the bottom line is:

DATE -> WORLDtime -> CHRoNOgraph -> INFO -> SYNC -> ACC -> PPT -> RFBSL

When a counter is active (possibly in the background) you will see the special icon with an R in a circle blinking (in a somehow similar way that you also see the special icons for alarms or chronograph when those are activated in the background). If you are on display with the special timer that is active you will actually see it counting down.

The settings for "CountDown and Interval Timer" are entered with a long press on the corresponding left button (top-left, since it is on the top line) - on that you have on the top line the minutes and seconds (up to 99:59) and on the bottom line a number of 'flags' - the rightmost one (but first activated) is ON/OFF, then the leftmost digit (second activated) is the timer number (1-9 since there are up to 9 timers), then a flag for the sound - which can be set to None, a sound at the End, a beep every second for the full time while the timer is counting or All (meaning both the beep and the sound at the end); the next flag is configuring what the counter will do/chain at the end - that can be off (just stop), re-start another counter (1-9) (that kind of chaining is essential for interval training) or start the Chronograph (the letter S for Stopwatch could not be used since it looks just like 5) - that is important in certain regatta-timers!

After reset all counters are off but counters 1-4 are pre-configured as one-time countdown timers with a sound at the End and intervals of 5 / 10 / 15 / 20 minutes; the only extra difference is that the very first one is also configured (in 'regatta mode') at the end to start the Chronograph - while 2-4 will just stop. The timers 5-9 are preconfigured in an 'interval-training loop' - first one in the loop (timer 5) is set for 30 seconds with a beep every second (which should keep you going on a very intense exercise), then it chains to timer 6 which is set for 10 seconds of rest (and no beep), and after that it chains to timer 7 which is set for another 40 seconds of exercise with the beep, then chains to timer 8 which is first half of rest time which Ends with a specific sound and chains to timer 9 which is apparently doing nothing (but the sound from timer 8 however is sounding during this one with a different frequency so you know that the cycle is restarting!) and then chains in an endless cycle to timer 5!

A one-time counter will stop itself (or at most start the Chronograph), BUT to stop such an endless cycle you need to be in CNTR mode on the top line and then enter 'Settings Mode' (long press on the corresponding left button) - that will SUSPEND all timers (however if you leave the watch there it will exit after a number of seconds of inactivity and the active counters will resume!) - the actual way to stop things is in 'Settings Mode' to move between timers 1-9 and see which one is ON and just turn that one OFF !!!

The link to the WorldTime firmware with repeater and sidereal time can be found in the post below:

https://www.watchuseek.com/f9/ezchr...ect-start-498121-post4168218.html#post4168218

And the actual WorldTime firmware with "CountDown and Interval Timer" is:

ez430_chronos_countdown+interval2rc4.zip

Please let me know if you see any problems/bugs!
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #245 (Edited)
A new release candidate - and the most accurate astronomical wristwatch in the world?

A new release candidate (maybe the final one) - and the most accurate astronomical wristwatch in the world?

Well, you might remember the post from here where I was describing the amount of variation seen around the 'mean cycle' for sidereal time, solar time and the moon phase. The amount of calculations needed to eliminate all those variations and achieve even remotely-precise numbers is quite big and as such it is a small miracle that good results have been possible with the ezChronos - probably in 25-50 years from now when an advanced wristwatch might have about the same computing power as our current notebooks this records will be vastly surpassed - but let me say that for the moment I strongly believe ezChronos is the clear leader ;-)

The 'feature distribution' among the five 'branches' for our HEQ firmware is now also pretty clear - and all five pretty much are filled very, very close to the maximum amount of memory - so if we discover serious bugs in some cases it might be needed to remove something from that branch in order to create space for the fix :-(

The 'Data-Logger' is pretty much unchanged since rc3 (and remains the only branch that can be used to determine the initial TC curve).

In the 'BlueRobin StopWatch' the only change since rc3 (other than fixes common to all stopwatch-related branches) is the addition of Day-of-Year (DY) and Week-of-Year (WY) as sub-modes in the date mode - the BlueRobin stack is just too big to leave enough space for anything else (after I already added ThermoCompensation, TimeZones, altitude-lock, repeater, the split-time and the 'info extensions' ;-) ).

Once the BlueRobin part was removed more space becomes available - so all the above plus the full world-time (with full automated DST change) together with the advanced (and now 5) alarms and the full barometer log/alarm are now present in two branches - the recent 'Worldtime plus CountDown and Interval Timer' one (where basically only two small fixes have been made since rc4 - both the 5 alarms and the 9 counters were showing one last extra 'false item') and the 'WorldTime plus Sidereal plus MoonAge' branch.

One quick word on the 'WorldTime plus Sidereal plus MoonAge' - the SIDER mode was consolidated in a more similar way to the full ASTRO mode - and the commands from the buttons are also more like in that one. There are now two sub-modes - the already-seen sidereal time and the new 'simplified moon age' sub-mode - and you can switch sub-modes like always with the corresponding right button. The MEAN moon age is displayed in DAYS.FRACTIONS (up to 29.53, but internally a value around 29.53058888 expressed in special-optimized integer arithmetic is used) and the '24 hours MEAN error' (currently seen in non-relevant claims of accuracy from inaccurate wristwatches) should only take place after more than 10000 years - take that Ulysse Nardin / Lange / IWC :-d (however the actual new record-holder is not this branch, but instead the new ASTRO firmware - keep reading below). As before there are still two possible longitudes which you can now 'toggle' with a long left-button press when sidereal time is shown. The visible indicator of which one is selected is the [AVG] symbol at the bottom - if that is blinking you are on the second (default 0.0 longitude - giving Greenwich Mean Sidereal Time GMST) one, if not blinking you are on the local one (with Local Mean Sidereal Time LMST) - which is the one that can be read from the flash INFO (if you have programmed one there). The actual manual user-editing of the longitude for the selected one (meaning that both can be edited in turn) is done with the same long left-button press but when the second sub-mode (MoonAge) is on display.

And now to the 'new kid on the block' - the full 'Astronomical Watch' - this one required so much space for the astronomical algorithms so that everything possible was removed - the full-auto DST/worldtime, the stopwatch and most of the advanced sub-features in other modes had to go (but timezones and manual DST are still present), and there are now only 4 modes on the top:

TIME - ALaRM - TEMPerature - BAROmeter

and only 5 on the bottom line

DATE - ASTRO - INFO - SYNC - RFBSL.

Even some internal features had to be toggled-off to make space, including the ability to WRITE the HEQ parameters - those written in the INFO flash still are used for TC, but if you want to further adjust them you need to first load some other firmware, write the new TC parameters into the flash memory, then load back the ASTRO firmware! There is only very limited memory left, so any bugfix will probably result in the removal of the names of the modes and then of the INFO mode (which here is only showing the battery level anyway).

But now let's see what the 'Astronomical Watch' can actually do:

- there are still TWO locations (and the flashing/solid indicator at the bottom, just like in the simpler Sidereal branch) - in that Sidereal branch locations were longitude-only (since that is the only location part used for sidereal time, the moon-age is using none); in the full ASTRO module the locations have both longitude and latitude and also a timezone used to adjust the times that are shown - the first location is read from the INFO flash (and as a timezone the local one set for the normal TIME module is used upon reset), the second location defaults to the Greenwich Astronomical Observatory (pretty much London) BUT with UTC time (meaning no DST!) - everything described above for those two locations can be individually configured in the corresponding 'Setting Mode'!

- the ASTRO sub-modes (changed with the right-button as always, for 1 second the name of the sub-mode will be displayed on the other = top line) are:

SIDR - SIDeReal time (hH:MM:SS)
EOT - Equation Of Time (+/- MM:SS)
TRNS - solar TRaNSit (actual local solar noon HH:MM:SS)
SRSE - Sun-RiSE (HH:MM)
SSET - Sun-SET (HH:MM)
MOON - MOON percentage illuminated (0.00 to 100.00, the % icon will also blink)
MAGE - Moon AGE (precise DAYS.FRACT, can be as high as about 30.25)
- (date for current new moon)
- (date for current full moon)
- (date for next new moon)
- (date for next full moon)

- the last 4 sub-modes are somehow special - on the top line instead of a name the precise hour and minute (HH:MM) will be shown for 1 second - so you must be quick and look at that part first if you are interested in it :-d

- the precision involved is different in sidereal/sun/moon - in sidereal time it still is the same 'better than 2-3 seconds' (plus whatever the watch time error is), the equation of time and solar transit are also precise to 2-3 seconds, the sunrise and sunset represent the 'official' values (not civil or naval sunrise/sunset) and seem to be within 2-3 minutes on 'normal' latitudes (below 65 N or S) and under 500 meters of altitude, going to within 10-15 minutes on extreme latitudes or altitudes; the moon times are also within 10-15 minutes or so from the corresponding astronomical values; the moon percentage and the age are slightly different in this regard - those use as a base the last new moon but also do a small amount of basic recalculation when the display sub-mode is changed - so for instance every about 15 minutes or so the age might change with 0.01 day and something similar could take place on the percentage;

- those precision values could have been within 1-2 minute for sunrise/sunset and under 3-5 minutes for moon - unfortunately the floating-point library that I can use for this CPU is simple-precision (around 6-7 digits) and HUGE efforts had to be made in adjusting the implementation of the algorithms so that the final accuracy would be decent (just as a very simplified technical example - for instance Julian Date JD 2455773.2773 is a good approximation for a new moon on Jul 30 18:40 UTC, however only the first 7 digits can be accurately represented in a simple-precision number - and that immediately would result in more than 12 hours of error - which of course is avoided by calculating from a different offset - which is also much closer to us than the '4713 before our era' JD starting point);

- but even with all those simple-precision limitations the accuracy for instance for moon-phase is 10-100 times bigger than what the previous record was claimed in wristwatches - remember that the solar mean times can vary with 14-16 minutes from the mean and the precise moon phase has variations up to 18 hours from the mean :roll:

- for extreme latitudes a special value can be seen for sunrise/sunset - during the 'polar day' you will see DAY (no sunset on current day) and during the 'polar night' you will see NIGHT (no sunrise) - for instance if you are among the over 300000 citizens of Murmansk you will see DAY from about May 22nd to about July 23rd and NIGHT from December 2nd to January 10th :)

Some other usage-level details:

- the same unusual trick seen in sidereal time is still used for the actual display of values over 19:59:59 - the bottom line does not have a full digit for the very first position, just one single segment which can be either off (a zero), on (1) or can blink - and that marks in this version an ultra-economical display of the value 2 :-d ; however that 'blinking 1 as a 2' should only be seen on sidereal time - if the solar transit is over 19:59:59 there is something seriously wrong with either the timezone or the longitude that you are using ;-)

- different 'Settings-mode' are (as always) started with a long press on the corresponding bottom-left button; however since there are so many settings you get 3 'entry points' - if you have the long-press starting in SIDEReal time (which is the default sub-mode anyway) you only toggle between the two possible locations - this will also force a recalculation (you will see later why that is handy);

- if you have the long-press starting in the second sub-mode (Equation Of Time) you can set the timezone used for displaying ASTRO times/dates - after reset the timezone for the first location is copied from the local timezone configured for the watch, but is NOT later set automatically if you change the local timezone (or for instance if you SYNC the watch) - so you should check that setting if something seems wrong by full hours! Also the second location is using UTC by default - that is handy when you want to compare UTC moon-phase times but for sunrise/sunset remember that during the summer London is using DST - but of course you can change that second location to whatever you want;

- in all other sub-modes a long-press will enter the settings for longitude and latitude (for the CURRENT location - you can know if you are on the first or on the second if the AVG symbol at the bottom is solid or flashing); settings are in degrees-minutes-seconds with East/West and North/South;

- you can calculate the astronomical values from some other location by simply entering the longitude, latitude and eventually timezone - when that is set the astronomical values are re-calculated automatically;

- also an automatic recalculation is triggered at midnight after date change; (technically that takes place immediately ONLY if the ASTRO mode is active, otherwise the CPU-intensive calculations are delayed until the user gets to the ASTRO mode);

- the date used in all ASTRO calculations is the 'current date' as set in the date module - and that can be used to calculate things for some arbitrary date (also based on the fact that the precise seconds in the watch are not affected by any simple changes in date or hour or minutes) - HOWEVER after MANUALLY updating that date/time an automatic recalculation is NOT triggered - and here is where the toggle between the two locations can be very handy since it forces a recalculation! Same result can be had by entering ASTRO settings mode and exiting with save - however keep in mind that every recalculation is very CPU (and battery) intensive!


A special acknowledgment goes to our moderator Ronald - not only he was the first to suggest the idea of an astronomical firmware with sunrise/sunset, but also he provided a very helpful coding example for that sunrise/sunset algorithm - in old Fortran code, but it still helped me a lot in understanding certain algorithms from other sources!


Testing things - the very first step is to be 100% certain that you have your longitude, latitude and timezone set correctly in the ASTRO module - remember that the ASTRO timezone is separated from the one in the local time and that you have two locations - probably leaving the second (the one with blinking symbol) untouched on Greenwich and UTC is simpler but on the very first location you should certainly set things right before checking any numbers!

A very good site to check the precision of the displayed sidereal time would be:

Local Apparent Sidereal Time

- it is using atomic time and the 'full astronomical correction' so we should test the watch against that site (and please report back if you see more than 2-3 seconds of error while your time is correctly set)!

A site I was using to check some of the other astronomical values is:

Sunrise and Sunset Calculator

but there might be other/better sites - please let me know what you find/use and how accurate the watch is compared to those!

The 'PC companion programs' have not been changed and can still be found at:

ezc_pc_v2rc2.zip

The actual firmware for each watch model (now 5 files for each frequency) can be found at:

433MHz_v2rc5.zip

868MHz_v2rc5.zip

915MHz_v2rc5.zip

I will be now testing for about two weeks the timing with one of the above (probably WorldTime+Sidereal+MoonAge, in the hope that Ronald will get Astro in tests, kjell the BlueRobin one and Webvan the WorldTime+CountDown) and maybe if no serious bug is seen then in 2-5 weeks we will have the final versions for everything b-)
 

·
HAQ and AW moderator
Joined
·
7,494 Posts
Re: A new release candidate - and the most accurate astronomical wristwatch in the world?

I will work with Catalin on the Astro firmware beginning tomorrow, ET.
I thank Catalin for his acknowledgement. I do discuss algorithms with him and check calculations from the watch's output, but the firmware coding and preposting testing was all done by him
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #247 (Edited)
Minor update ... but we still have to call it RC6

After the RC5 version described in detail at:

https://www.watchuseek.com/f9/ezchr...roject-start-498121-post4184942.html#poststop

a really small bug related to how some flags were being read from the customized INFO flash (which I believe are anyway only present in 2 watches so far) have surfaced - that had a very minor impact on 4 of the 5 firmware branches but since in the most recent ASTRO version things were ultra-tight I also needed a little more 're-organizing' in that code to provide the extra space for the bugfix.

Long story short - new (and this time hopefully final) versions are available at

433MHz_v2rc6.zip

868MHz_v2rc6.zip

915MHz_v2rc6.zip

Those contain just the 4 of the branches that were affected, and from those the 'BlueRobin StopWatch', the 'Worldtime plus CountDown and Interval Timer' and the 'WorldTime plus Sidereal plus MoonAge' versions contain just the bugfix and no other change.

In the full 'Astronomical Watch' I could not resist and since 99.999% of the code was already there I have also added the nautical (NRSE, NSET) and civil (CRSE, CSET) values for sunrise/sunset - basically the nautical value corresponds to the time when the horizon becomes visible (and outside which no 'sight-based orientation' is possible at sea) while civil values tend to follow the time when artificial light is normally needed for certain activities. Also the 'day of year' and 'week of year' are back in the date mode. We have also noted that the 'percent of moon illuminated' might at certain times be less accurate than suggested by the display with .01 precision, but that is just a result of using the same display routine as in the moon age so for the moment things were left in that form due to space limitations :-d The only other detail that might be worth mentioning is that in the ASTRO firmware we have just one single 'special buttons combination' - top-left plus backlight (in this order) will toggle (on/off) the 'buttons sound'! (the other 3 branches have the full repeater plus some other quick-navigation combinations, but don't need this one since in INFO mode after the battery voltage there are settings to control that sound and a few other similar settings).
 

·
Registered
Joined
·
25 Posts
Re: Minor update ... but we still have to call it RC6

To be sure the testing is usefull, I was wondering if you could make a short list of option I shoould test using the BlueRobin.
rgds
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #249
Re: Minor update ... but we still have to call it RC6

To be sure the testing is usefull, I was wondering if you could make a short list of option I shoould test using the BlueRobin.
rgds
Anything that you previously used on the BlueRobin part to see if it still there and still works - just to test that we have not introduced any strange problem.

From the 'new part' - the timezones and DST are now no longer so new, maybe see if something else from it has any problems - the chronograph with split time, the extra settings in INFO mode, the repeater and generally any 'buttons combination'!
 

·
Registered
Joined
·
25 Posts
Re: Minor update ... but we still have to call it RC6

Was playing with the BlueRobn version. I am now in a loop on lower part. Was playing with repeater - long press on light, and use the INFO to change between S 10 and S 15, was going to the next by pressing #. (I did not write down the keys I pressed before the loop started) Pressing # now show 23624 , 24423 and this repeats, after ab 1 min it show S 15 a loong pres on # goes to S 10 (the 2 left "symbol" show up, one is the heart")a pres on lower right show 23624 Have tried long press on R and L lower part, but can not geet out of the loop.
The uper part work fine.
Update: Pressing both left key (upper and lower) bring up rFbSL, but when I press R-dows to start download the second symbol from left show up, and next press show 25049 and then after a min rFb5L show up again. Can not start the update on the watch
I asume that the only option now it to open up the watch and use the other USB programming and debugging interface.
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #251
Re: Minor update ... but we still have to call it RC6

Was playing with the BlueRobn version. I am now in a loop on lower part. Was playing with repeater - long press on light, and use the INFO to change between S 10 and S 15, was going to the next by pressing #. (I did not write down the keys I pressed before the loop started) Pressing # now show 23624 , 24423 and this repeats, after ab 1 min it show S 15 a loong pres on # goes to S 10 (the 2 left "symbol" show up, one is the heart")a pres on lower right show 23624 Have tried long press on R and L lower part, but can not geet out of the loop.
The uper part work fine.
Update: Pressing both left key (upper and lower) bring up rFbSL, but when I press R-dows to start download the second symbol from left show up, and next press show 25049 and then after a min rFb5L show up again. Can not start the update on the watch
I asume that the only option now it to open up the watch and use the other USB programming and debugging interface.

Hmm, that is pretty strange - how was the battery level before the problem?

Before opening the watch you should also try some of the other 'buttons shortcuts' - unfortunately I believe since RC5 or so I did remove the 'special reset' (which was the two buttons on the left plus the backlight button pressed at the same time) thinking that we are now in the 'safe stages' - however I believe you might want to try "left + right" - first on the top line (if that works it should take you to TIME), and then on the bottom line (it should take you to DATE).

If that fails the easiest way to restore over USB is to use the "eZ430-Chronos Watch Firmware Update Tool" from:

BM innovations GmbH - eZ430-Chronos
 

·
Registered
Joined
·
25 Posts
Re: Minor update ... but we still have to call it RC6

Tried "left + right" on top, then on bottom, got rFb5L, but downloan did not start. Tried it many times, also some other compination. Had started wireles updata on pc. Sudenly the watch look dead (nothing on display) , tried right download a few times and sudenly the download started. Now its working again. The v is 2.94 Rgds
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #253
Re: Minor update ... but we still have to call it RC6

Tried "left + right" on top, then on bottom, got rFb5L, but downloan did not start. Tried it many times, also some other compination. Had started wireles updata on pc. Sudenly the watch look dead (nothing on display) , tried right download a few times and sudenly the download started. Now its working again. The v is 2.94 Rgds
Good that it worked without opening the watch - on Monday I will try to install the Bluerobin version and I will try to repeat the bug!
 

·
Registered
Joined
·
1 Posts
Re: Minor update ... but we still have to call it RC6

For what it's worth: tideals.com is giving out 50% off coupon codes until August 30th, so I just picked one up for $25, free shipping. I'm excited to join in on the fun.

Cheers!
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #255
Re: Minor update ... but we still have to call it RC6

For what it's worth: tideals.com is giving out 50% off coupon codes until August 30th, so I just picked one up for $25, free shipping. I'm excited to join in on the fun.

Cheers!
Wow, now I am a little tempted to get one in the US frequency and some heart-band :-d
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #256
Rc7

New versions with some bugfixes are available at:

433MHz_v2rc7.zip

868MHz_v2rc7.zip

915MHz_v2rc7.zip

Those contain just the 4 of the branches that were affected = the 'BlueRobin StopWatch', the 'Worldtime plus CountDown and Interval Timer' and the 'WorldTime plus Sidereal plus MoonAge' and the full 'Astronomical Watch'.

The fixes involve more resiliency on buttons combinations, a fix for an non-initialized split-time and better precision in some places (especially in the moon percentage) in the full 'Astronomical Watch' (plus that now things are automatically updated when you change the date). Also the simplified moon-age 'WorldTime plus Sidereal plus MoonAge' now has a 1-second label to be more 'user-friendly'.
 

·
Registered
Joined
·
1 Posts
Re: Rc7

Hi, I've just learned of this development. i am a longtime developer of time and astro algorithms. right now im working on a world calendar clock for embedded microcontrollers like the PIC and AVR. i am interested to get the chronos and try myself.

to compare, my plan is for a clock that operates in different calendars for example, Islamic, Hebrew, Chinese etc. In addition, i am supporting the astronomical calendars such as Persian. for this i need to compute all the usuals; sunset/rise, equation of time, apparent solar position, sidereal time etc.

in addition, i need the accurate apparent solar longitude. this is to calculate the real start of the year (ie the first point of Aries). also i get the accurate times of seasons and solstices etc. the persian calendar for example starts at the real new year!

i have hammered the equation of time into simple form, but this one is a lot more complicated. right now im coding a super accurate version for reference. i have some code from a while back that is simplified which i plan to use, but i must check it first.

are your calculations for the chronos open source. if so, i will compare your values to mine. it will be useful to find bugs in both!

as another point, i was amused by your DST comments because i did the same for the uwatch some year or so back. world DST is not as simple as it sounds. i also had many bugs. i think this is why no real watchmaker provides this. also the rules keep changing :)

best,
-- hugh.
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #259 (Edited)
Re: Rc7

Hi, I've just learned of this development. i am a longtime developer of time and astro algorithms. right now im working on a world calendar clock for embedded microcontrollers like the PIC and AVR. i am interested to get the chronos and try myself.

to compare, my plan is for a clock that operates in different calendars for example, Islamic, Hebrew, Chinese etc. In addition, i am supporting the astronomical calendars such as Persian. for this i need to compute all the usuals; sunset/rise, equation of time, apparent solar position, sidereal time etc.

in addition, i need the accurate apparent solar longitude. this is to calculate the real start of the year (ie the first point of Aries). also i get the accurate times of seasons and solstices etc. the persian calendar for example starts at the real new year!

i have hammered the equation of time into simple form, but this one is a lot more complicated. right now im coding a super accurate version for reference. i have some code from a while back that is simplified which i plan to use, but i must check it first.

are your calculations for the chronos open source. if so, i will compare your values to mine. it will be useful to find bugs in both!

as another point, i was amused by your DST comments because i did the same for the uwatch some year or so back. world DST is not as simple as it sounds. i also had many bugs. i think this is why no real watchmaker provides this. also the rules keep changing :)

best,
-- hugh.


The firmware is not open-source since I am still keeping an eye on eventually getting a patent on some of the thermo-compensation (TC) stuff from inside.

The full astronomical algorithms from the ASTRO module are nothing very special - I was just playing and comparing the results from 3-4 sources/descriptions, but generally some of the books from Meeus are the most relevant source. For the actual implementation - again nothing huge was needed - just keeping in mind that the standard floating-point (software-emulation) library is just single-precision and the code must be made as small as possible in order to fit in the very limited amount of space available - so certain correction terms have been omitted in the current implemetation (also since the single-precision library was anyway kind of a little limited).

While in the simplified SIDEREAL module I was able to implement two operations strictly as 'fixed-point integer' (the sidereal mean time - which guarantees 2s of precision anyway, and a mean lunar time which is more accurate than whatever was claimed as the previous record in a wristwatch from Ulysse Nardin), the actual reasonably-precise approximations for sunrise/sunset and new/full moon can not be done without the use of floating point.

One other major limitation that might become quickly apparent in the ezChronos is the display - still very generous when compared to older models having just 4-6 digits, but definitely a very simple 7-segment one which I do not believe will be so usable for the very graphic arabic writing :roll: - but other than that the ezChronos is definitely a great way to start testing things!
 

·
Registered
Joined
·
3,107 Posts
Discussion Starter #260 (Edited)
Version 2.0 final

Since my next weeks are hard to predict I decided to make the most of this weekend, so here we have it - the final V2.0 firmware family - you can download the watch firmware from:

433MHz_v2_0.zip

868MHz_v2_0.zip

915MHz_v2_0.zip

and the 'companion PC programs' from:

ezc_pc_v2_0.zip

(all are dated on August 28th 2011).

Legal (but short-version) disclaimer - while I did test things the best I could, of course that any tests that you do are 'on your own risks' and no 'warranty' of any kind can be provided
The files are free for personal use only (I retain full copyright plus intelectual properties / patent rights) and downloading/using any of those programs should be done only if you agree to this terms!


Since the previous release there have been only two very minor and very specific bugfixes - one was in the BlueRobin Stopwatch (which since rc7 had a way to display seconds on the top line and HH:MM on the bottom line) and the other was in the ASTRO firmware (and was related to the 'long backlight').

In the PC programs the only change was a warning in case you are using EZC2PARAM_0.EXE with the ASTRO firmware (you can write the new ThermoCompensation parameters in RAM but the code to write those to FLASH and reset the watch had to be removed to make space for the astronomical calculations - in case you want to write those values to FLASH you will need to program one of the other 4 branches of the firmware, write the values and then eventually program back the ASTRO firmware - but of course all 5 branches read/use the values automatically).
 
241 - 260 of 361 Posts
Top