locked Consequences of mistimed computer clock #TechnicalHelpQuestion


lmeeny
 

Hello,

I've noticed that even with large, 3 to 4 second, offsets between the sending and receiving station messages can often be decoded. When I begin a reply late I assume transmission begins at the normal sequence start and is simply truncated at the end. Does the redundant error correction cover that large a time span?

As the time snyc offset between two stations increases does decoding continue just as well as say differences between 0.1 and 2 seconds and then just collapse? Can a weak signal be decoded just as reliably at a 0.1 second and 2 second offset. Are there diminishing returns as sync offset between two stations decreases?

73,
Ed Machak


Reino Talarmo
 

I've noticed that even with large, 3 to 4 second, offsets between the sending and receiving station messages can often be decoded. When I begin a reply late I assume transmission begins at the normal sequence start and is simply truncated at the end. Does the redundant error correction cover that large a time span?
As the time snyc offset between two stations increases does decoding continue just as well as say differences between 0.1 and 2 seconds and then just collapse? Can a weak signal be decoded just as reliably at a 0.1 second and 2 second offset. Are there diminishing returns as sync offset between two stations decreases?
Hi Ed,

There are two issues that affect to the decoding probability of FT8 signals. The first one is time span used for synchronization search. It is about -2.5 to +2.5 seconds of the nominal start time of the reception. The start time is defined by local PC time. At a late TX start the transmission is simply truncated at the beginning. I.e. equivalent to a very deep QSB at the start, hi. That minimizes potential synchronization time difference. If the time difference is too large synchronization is not found and decoding is not possible at all.

The second is error correction as you assumed. It can cover more than about 4 seconds depending on the signal to noise ratio.

My knowledge is limited about the timespan receiver is actually sampling received signal after a transmission. There can be a loss of the start of the received message for that reason as well. In any case there is a range of the sync difference where decoding capability is not affected. With a larger sync difference decoding probability goes faster and faster down depending on the signal to noise ratio. I have not seen any simulation of it to give more meaningful values. It is quite safe to keep own time sync within one second.

73, Reino OH3mA


lmeeny
 

Reino,

Thank you for the informative reply. That the beginning of the transmitted message is truncated from the front and error correction time window is new info for me. The various "FT" modes are impressively robust!

73,

Ed


neil_zampella <neilz@...>
 

FWIW .... time sync is actually part of the 'System Requirements' as
listed in the WSJT-X User Guide :
https://www.physics.princeton.edu/pulsar/k1jt/wsjtx-doc/wsjtx-main-2.5.4.html#SYSREQ

The guide is a good read and has a lot of good info such as this.

Neil, KN3ILZ

On 1/26/2022 8:28 AM, lmeeny wrote:
Reino,

Thank you for the informative reply. That the beginning of the transmitted message is truncated from the front and error correction time window is new info for me. The various "FT" modes are impressively robust!

73,

Ed




Gary - AG0N
 

On Jan 26, 2022, at 7:48 AM, neil_zampella <neilz@...> wrote:

FWIW .... time sync is actually part of the 'System Requirements' as
listed in the WSJT-X User Guide :
https://www.physics.princeton.edu/pulsar/k1jt/wsjtx-doc/wsjtx-main-2.5.4.html#SYSREQ

On that subject, a little expansion may be needed for newcomers and even as a reminder for some older ones who have forgotten.

I leave my rig on, monitoring 24/7/365. If I’m not using it myself, it is parked on an interesting band where I expect activity, especially overnight when the upper bands aren’t open. Overnight, I’m usually on 80 meters. During the day, I’m on 17 or as high as propagation will allow for a reasonable length of time.

When I get up in the morning, I look at the incoming list of stations to see how DX is doing. I am really surprised how many people aren’t paying attention to their clocks. It’s irritating to see so many that are a second or more off. Sometimes, I’ll try to let a station know about it with a free text, but when there are three or four of them that are way off, it just seems strange. The tools are right in front of each operator. Why aren’t they fixing it? Almost every morning I see KD1XXX (prefix could be wrong) sitting nearly 2 second off.

Some think that as long as they are within a second or two, they are okay. True - TO A POINT! Think about it a little more. If you are off +1.5 seconds, anyone who is off the other way more than 1 second (-1) will not decode you and you won’t decode them. That’s because the combined error is your signals is over 2.5 seconds total. You’re burning up the air, wasting some of your time.

Please strive to say well within 1 second. There are plenty of automatic ways to have your clock set properly to the right time. EVEN MANUALLY, your can get within 1 second very easily. Just look at the incoming signals. Not all of them are correctly time. BUT THE MAJORITY OF THEM ARE WELL WITHIN A HALF SECOND (.5).

73, and keep track of time (don’t waste it). 😉

Gary - AG0N