April 14, 2014 at 3:29 pm #20272
I completed my Alpha 5 this weekend. I have v2 and loaded the GPS program. All works perfectly except the time stability with the GPS interface. When the GPS is enabled the Alpha 5 drifts about 2 seconds an hour when compared to WWV. The GPS has a good lock and updates the Alpha 5 initially but drifts rapidly after that. I have tried it with the GPS disabled and it is much more stable with the Chrono Dot. The GPS locks quickly and never loses Lock.
April 14, 2014 at 3:37 pm #21512
- This topic was modified 4 years, 2 months ago by Windell Oskay.
To the extent that you do have a lock, it is likely not actually drifting. That means that you are apparently having periods of a sizeable fraction of an hour when you do not have lock. We haven’t had this problem with our example running the code, so it’s likely not a firmware bug.The basic problem– and the reason that Alpha Clock Five does not come with the GPS module as a default option — is that it’s hard to guarantee a GPS lock, especially indoors. You might consider moving it elsewhere or adding an external antenna to the GPS module.April 14, 2014 at 4:28 pm #21513
Thanks for the response Windell.
The GPS module has a red LED that flashes @ a 2 second rate when acquiring and at a 12 second rate when locked. It is always locked. I have about 25 GPS receivers in my home (some using the exact same GPS module) and none of them ever drop out. They are associated with my 25 Nixie clocks. So I don’t believe its a GPS signal strength issue.
What is strange is the drift rate is much higher with the GPS then the basic clock. Would it make sense even if I was losing lock with the GPS that the drift rate would be higher?
I suppose it would be a function of how the code is implemented…April 14, 2014 at 4:45 pm #21514
The author thought that using the Chronodot was not important when the GPS was used… and that is part of the explanation.Does the Alpha Clock Five indicate that it is staying locked? It indicates unlocked condition with the alternating colon separators.April 14, 2014 at 6:35 pm #21515
I didn’t know that. I will look closely tonight. I wonder if the Chronodot should be removed when the GPS is enabled? I looked closely at the GPS illustrations and could see a Chronodot in place.April 14, 2014 at 11:45 pm #21516
The colons are steady with the GPS enabled.
Please tell me your settings so I can duplicate them. Maybe it’s an issue with the settings…
JimApril 16, 2014 at 1:20 pm #21517
We have GPS on and DST on… I’m not sure that there are other settings that are relevant.April 16, 2014 at 2:13 pm #21518
I have been in contact with the author. I was able to resolve the issue. I’m not quite sure what got it working. He gave me alot more details about the display indications when it is updating or not. At any rate all is well!
JimApril 18, 2014 at 1:44 pm #21519
Okay great– glad to hear it. If you do figure out what made the difference, it would be nice to know for people who have the same problem in the future. :)July 30, 2022 at 8:37 pm #29987NeusseParticipant
IO know this is old but I thought I would add my two cents. I found that you do not need a GPS location fix that is communicating with 4 or more satellites. You only need one unless you are looking for millisecond accuracy. Turns out that even without a fix you are receiving a good time from at least one satellite. That is all most of us really need. I modified my the GPS code to ignore the fix and simplify the whole acquisition of date and time. Here is a quote from an article that seems to validate this.
Explination of not waiting for a GPS fix to set time:
Each satellite broadcasts time. Only one is needed to determine the time with accuracy to distance to the satellite in light seconds – under 0.088s of error. But you need at least three, preferably more satellites to get the location fix and the high-precision time (adjusted for distance from the satellite = time it takes for the signal to reach you.) As your GPS obtains more signals, it calculates its position (fix) and adjusts the clock, but the initial first broadcast is enough to establish timestamps.
- You must be logged in to reply to this topic.