Interesting that this went away for you at 0126z after having been so strong, the network did not go off air until 0141z. I caught this network last night starting at about 0043z and recorded it until the end, at 0141z.
This is what I call the Pips Network in my logs (I often name signals I see more than once in my logs to assist searches in my database). A search of these forums for that term will find past posts about the network. This mystery network of ditters has appeared periodically for several years now (my earliest log is April 2012). It shows up a few times a year. I have no real idea what it is for, but I suspect some kind of ionospheric measurement tool.
You heard it on 6820 kHz, but if you had looked around a bit you would have found it on many other freqs as well. The pulse is (in this case, it uses several different formats) 62.5 msec long with a period of 3.0 seconds. It cycles up through its frequencies, freq 1, then freq 2, then freq 3, so on until it reaches its last freq and then back to freq 1 to start it all over again. With the 0.0625 sec x 3.0 sec format this allows up to 48 frequencies. I only found 38 of the possible 48 in use last night (from 6820 kHz to 17440 kHz), it is possible, even probable, that all 48 were in use but propagation conditions prevented me from finding the remaining 10. The upper freq Pips were very weak. With the lower freq Pips as strong as they were I suspect 6820 kHz was the lowest freq in use.
There sometimes are two sources active at the same time, they often share one or two frequencies and on those frequencies you may hear a double Pip with odd timing. Two frequencies last night (10990 kHz and 12010 kHz) had double Pips however I only found one network active for sure. It appears the second network may have been only using those two frequencies and no other.
The freqs I had active last night, from 0043z until 0141z, when it turned off:
6820
7570
8400
9050
9900
-*
10200
10350
10510
10670
10830
10990 (double Pip, 2nd Pip is with the other freqs, 1st Pip with 12010 kHz)
11150
11490
11660
11830
12010 (double Pip, 2nd Pip is with the other freqs, 1st Pip with 10990 kHz)
12190
12560
12940
13130
13530
13730
13940
14150
14360
14580
14800
15020
15250
15480
15710
15950
16190
16430
16680
16930
17180
17440
A couple of things to note:
The "-*" entry I made indicates that based on pulse timing there should be a frequency there, something between 9900 kHz and 10200 kHz. I found nothing between them, but the Pips may have been covered by another signal in that range, several other freqs were under other signals and I had to look very close to find them.
This signal often uses frequency sets, or combinations of frequency sets, it has shown before, but I have not logged this freq set in use before. Some of the freqs are in my logs, for example 9050 kHz and 11150 kHz are common freqs, others I have not seen before. I have not been as active in the past year or so as in years past though, so they may have been using this for some time and I just missed it.
10990 kHz and 12010 kHz showed double Pips. The second Pip on each of those freqs was in sequence with the other freqs found, i.e. the timing showed there was a probable relationship with the other freqs. However the first Pip on each of those freqs was not in sequence with the other freqs, but was in sequence with each other. SO while it looked like there should be two networks (sets of freqs) active (as there often is) the only freqs I could find with that second network signal were 10990 and 12010 kHz.
T!