Well it’s not packet. There’s your problem. You should select normal USB and the recommended dial frequency as conveniently listed in the program. You need no more than 2 kHz bandwidth as the JT65 “band” starts 2 kHz higher than FT8.
The computer/tablet/smart phone will produce the tones with a base frequency of whatever you select. in the range of say 300 Hz to 1975 Hz. It’s probably best not to go below 300 Hz as the IF band-pass of the other guy’s Rx will likely reduce the signal (but not the S/N) below that.
You need a keying signal, most of us use an audio derived switching signal to operate a FET or BT. I suppose you could use the transceiver’s VOX but I haven’t tried that.
The main difference between different A-Ds is in the number of bits. The more bits the greater the dynamic range and the better the sensitivity. There is an optimum sampling rate too. This are explained in the Guides.
FT8 is intended to be quick(relatively) and easy and very sensitive. That is no effort to learn a code but able to work stuff you can’t hear. Easy doesn’t equate with bad. The majority of DX stations have voted in favor of it. It will gain acceptance in SOTA too.
I think the fish are safe in most barrels, it sounds like a no brainer, until you try it. I shall say no more lest I get into trouble.
I saw someone using it and thought that’s neat. Also you could initially call the other station with message 2 instead of message 1 if you didn’t want to give the grid square. Then run the modified message 4 after receiving a R and report as per message 3. Done when you receive either RRR or 73…
For SOTA a modified message 5 with the SOTA peak reference would need to be used often enough for the chasers to be fully informed without a spotter. IMO.
I just thought the suggestion might be helpful - Packet is used because all the interfacing for rig control is there but the standard setting in the FT5000 (and other rigs of the same family) is with a 1500 Hz offset which is easily adjusted to whatever you like (+/- 3000 Hz)
You can set that up as a standard (alternative) response in WSJT-X
It might be the Signalink. That is not a great USB audio interface. They didn’t bypass the 5 V from the USB. That 5 V is pretty “dirty” with noise. The noise gets into the audio. 10 dB seems similar to the poor decodes other people have seen with the Signalink.
This page has an extensive set of mods to improve the design of the Signalink USB.
But you are probably better off spending a few dollars on a better audio interface. This page has a great set of tests on USB audio interfaces. The SYBA is quite good and usually available for $12 (US).
Suggestions are good. If operating Packet with a 1.5 kHz dial offset gives you a better operating experience, go for it I prefer to keep the operation with computers to a minimum so don’t use rig control programs. My concern with a Packet selection would be whether you get a restricted band-pass. I’m showing my ignorance - I’ve never used Packet. I’m also vaguely aware that you need to set baud rates properly too.
It just seemed to be making it more complex than necessary.
Modifying the RRR in the preformed message 4 to RR73 is much simpler than fiddling about putting his call into message 5.
Thanks for the Info on the Signalink. I’ve been very happy with mine and prefer it to a more expensive US built one and my own old home brew unit. I’ve not seen a problem with the USB power noise, perhaps mine has the intended value of bypass, I will however check more carefully and if noise shows up see if I can find a good small choke that could be put inside. It may be computer dependent.
I’ve not checked the frequency response carefully - it’s the S/N ratio that is important so a 6 dB or so of inequality in the pass band is probably tolerable. Of course I’d like it within 1 dB and if mine is much worse than that if I see some small 600:600 ohm transformers I’ll grab them.
I have two issues with FT8 that overshadow any interface imperfections.
1 Power. My reports are averaging at least 8 dB lower than I’m handing out. It’s clearly not being treated as a QRP mode. I see many signals in the positive dB region and we should in theory be able to work QRP. One station gave me a report of -17 when he was 0 dB. I was running 50 W. Work that out. Yep 2.5 kW! It may be legal where he lives however I have two stations within 4 km of me and if we all operate in the same 2 kHz of the band running our maximum legal 120 W it will be impossible. At present I transmit in the same time slot as the loudest and that works but it can’t be done all the time. I can use the RX notch to knock the signal down into the negative dB region but that’s got to be adjusted every time he changes frequency and I can’t deal with more than 1 signal in practice even though i have 2 notches.
2 QRM. There is difficulty finding a clear channel in the 2 kHz band. That’s with the propagation being crook and the mode being new. It’s going to be awful when conditions pick up and more people get on the mode. We will need more frequencies. Band plan administrators take note please.
You can get better signal to noise with a $12 dongle. Read the links. And nearly all rigs already have VOX, so you don’t need the external VOX implemented in the Signalink USB.
However, if someone really wants a Signalink USB, mine is still for sale. Make an offer.
The reports in FT8 do not tell you how strong the signals are and they can’t be compared. They just tell you how strong an audio signal is compared to ~2kHz of audio passband at the receiver. All the other signals in the passband, receiver AGC slope and audio amplifier frequency response all affect FT8 signal reports. To put it more clearly:
if you are the only station received you get one report, but if there are other stations received at the same time as you you will get a different report.
if they have 3 different receivers at their site (on the same antenna) they will give you 3 different reports.
if you transmit at +800Hz you get one report, if you transmit at +1800Hz you get a different report.
if you increase power with 6dB, your report won’t go up by 6dB.
Thankfully that is the last thing I would be concerned about when on a summit. Making the contact is the prime task.
I have started to use the mode for a few contacts from home running low power. The program works well down to -20dB which is in theory is 6 dB worse than JT-65, but as you say Razvan, there is no common datum point for these figures. The two main things that I like about the program are the speed of the contact and the fact that it takes over the running once the QSO is underway. This provides relief from the stress that could otherwise result due to the short “thinking period” between overs and that is the last thing I want when activating.
Actually, the biggest challenge I found while activating summits in FT8 was actually adjusting the time correctly. Every milisecond you are off cuts in the decoding time window, which is just 1.5s.
The Windows 10 tablets going in and out of sleep mode loose time sync pretty fast. You need pretty good mobile data coverage to get a good time sync (I use time.is to check).
I use a freeware program called Atomic Clock Sync and log on before I leave for the summits. No issues so far on JT-65, but I agree that accuracy is essential on FT8. Hopefully catch you on a summit sometime.
This $10 audio interface gets good reviews. Use the VOX in your rig to key it. The Signalink is an outboard VOX, but nearly all rigs have that built-in.
For those with only an old bag of scrap for a laptop, be encouraged. This evening I hooked up the portable interface, set the FT-857D to 10 watts and VOX and tried FT8 and it worked. This is with a Ssmsung N310 which has a 1.6GHz Atom processor and 2GB of RAM. The auto-response system worked brilliantly even though it was taking several seconds to decode in monitor mode when it looks at all the signals in the passband. Obviously once in QSO the software has less to concentrate on. Overall I am impressed big time.
Edit 1: further QSOs have revealed that the processor does have difficulty decoding when there are a lot of stations active. I was fortunate to have good copy the first time around, but other contacts have required some repeat overs. I might experiment with a narrower filter than the stock SSB one.
Edit 2: following more tests I have come to the conclusion that the netbook is insufficient to run the mode satisfactorily and I will stick to PSK31 / JT65 on summits. The issue relates to both the number of signals received and their levels relative to the noise background. If there are a few strong stations, the netbook works after a fashion, but with more stations at lower signal levels then the processor struggles. Using a 500Hz filter did not appear to improve the situation, rather the reverse.
The FT8 and other JT programs are designed for CD quality A/D D/A, 16 bits and 48 kHz sample rate. The SignaLink meets that spec. To compare its audio capability with other A/D systems isn’t necessarily meaningful. The one area of possible failing was the power supply line filtering being pretty average and they claim that was fixed.
I’m happy with my SignaLink. It’s easy to set-up out of the box, fits the pack with little extra weight. Yes it may be possible to get a couple of dB improvement but given that my reports average 8 dB less than what I give out, what’s the point of heating up the soldering iron?
Dongles vary enormously in performance and require a bit of work to interface with the rig. . They would not have any better power filtering then the SignaLink and there is no galvanic isolation. Will I hear better? I doubt it. Even on UHF I can hear JT65 stations very well with my Signalink.
A home brew system using the guts of a cheap dongle might extract every last 0.1 dB out of what is possible, but I made the choice to buy a SignaLink, work stations and build other stuff.
I use Dimension 4 and a USB 4G telephone internet connection in the field, which of course doesn’t work on some summits. I am considering using a GPS time source I have gathering dust on the shelf - just need to find the disc with the drivers.
Re Reports. I think it’s the same process for all of the JT programs. I could be wrong WRT FT8 but I understood it measured noise in an unused channel and then adjusted that for a 2 kHz bandwidth in the calculation for the reference. Joe gives a warning that it will be inaccurate for strong signals. It was after all designed for EME operations. The test data I have seen from Rex VK7MO leads me to believe the reports are far better than any “S meter” or by ear reports. Admittedly Rex is an EME and weak troppo signal operator but at lower signal strengths 3 dB does seem to be 3 dB.
I will keep your comments in mind as I haven’t seen any test results for FT8. Until then I am happy to accept the reports as indicative rather than imaginative. I have read in the User Guide that for JT65 the receiver audio range should be limited to less than 30 dB and maybe for FT8 it should be less.
IMO FT8 should have been promoted as a QRP mode - some ops think that reducing their output to 1.5kW is OK. They become alligators. Their signal reports are probably meaningless as they are often in the + dB territory…
It can be distinguished from the older WSJT-X User Guide V 1.7.0 by the V 1.7.1 designation and it’s publishing date 3 August 2017. No doubt there will be more updates when FT8 is moved from it’s Beta status.