Recently on the WSJT-X developer mail list a query was sent about why Q65 15 second mode does not work on EME. It started with this report from Dwayne, AB6A in which he asks about why Q65 15A does not work on EME. Joe goes on to explain that none of the 15 second period modes will work on EME and more importantly WHY they should not be used for EME.
I found this to be a very clear description of WHY things like FT8 and Q65-15 cannot be used on EME, I hope it helps others understand why the use of FT8 for EME is just a silly idea.
Dwayne Sinclair wrote:
There have been ongoing reports from various operators best summarized in the post attached that Q65-15A is “broken”. What is current consensus on this and if it is broken, is there any plans to address the issues?
Q65-15A Part II Testing and Sensitivity Bob Atkins - KA1GT
Q65-30A test results
Regards Dwayne AB6A
Joe Taylor wrote:
Hi Dwayne,
Thanks for the QSO with W2ZQ, last evening.
Who is saying Q65-15A is broken, and why?
Are you (or they?) talking about EME? If so, are you aware that the Q65-15x sub-modes are not supposed to work for EME. The 2.5 s EME delay is too large.
-- 73, Joe, K1JT
Dwayne Sinclair wrote:
Thanks Joe for QSO and the confirmation that Q65-15A is not to be used for EME. Our “bad” in that in our circles we may have been referring to it as broken.
The attached article details some of the testing that has occurred to date in an attempt to work with what we have.
I guess the EME community (including me) are looking for a faster QSO where conditions and equipment support it?
Regards Dwayne AB6A
Joe Taylor wrote:
Hi Dwayne,
I discussed these issues at some length with Bob, KA1GT, some 1.5 years ago. Here are the relevant details.
Sure, we could make a version of Q65-15A that can decode after normal EME delays. It would not be a very good fit for the channel, however.
This is because the duration of a Q65 transmission is 12.8 s, and the transmission starts at t = 0.5 s into the 15-second sequence. Add a 2.7s EME delay, and you're already 1 second into the next sequence -- even without any clock errors.
You could make the Tx duration smaller, say 10 s. Then, if the transmission starts at t=0.5 s and there's a 2.7 s EME delay, reception would be complete at t=10.5 + 2.7 = 13.2 s, and (if clocks were spot on) the decoder would have 1.8 s to do its thing.
Even this is not really enough, because clocks are not always "spot on". Moreover, you would lose an additional
10*log(12.8/10.0) = 1 dB of sensitivity. WSJT modes do a lot to gain every 0.1 dB of sensitivity, so we certainly don't want to to that!
Fifteen-second sequences are not an efficient choice for a path with delays as large as 2.7 s. If there were many times more EME activity than at present, so that very fast QSOs were desirable in contests, and if nearly everybody had many dB's to spare on the EME path, something like Q65-15A might be desirable.
As things stand, our advice is to use Q65-30B for EME on 1296 MHz.
-- 73, Joe, K1JT