Pages: [1] 2   Go Down

Author Topic: Why Does WSJT-X FT-8 "Wide Graph" Use Audio Frequency Versus RF Frequency?  (Read 317 times)

K4SSS

  • Member
  • Posts: 9

Hi All.  Bob, K4SSS here. 

After some 55+ years in the hobby I have decided to get my feet wet in the digital mode world.  I've downloaded and installed all the necessary WSJT-X software, and at least according to the WSJT software and my dummy load, it appears that I have it up and running with my Yaesu FT-911A.   I intend on trying FT-8 first, but before I go live I need some help in getting a few technical questions answered.  I've read the WSJT Manual cover-to-cover and looked at the WSJT working group threads, but I can't locate the information that I'm looking for and hope one of you much more experienced digital operators out there can offer me some assistance and set me straight.

Specifically:  With regard to the WSJT "Wide Graph," which is the upper graph containing the "waterfall."  The manual states (Version 2.6.0, Pg 28) that this "graph window displays the frequency spectrum of the received audio."  The user inputs the lower and upper audio frequency limit, which "define the FT-8 decoders search window."

I understand the purpose, well, at least I think I do, of the lower and upper audio frequency limits from a technical aspect,
but fail to understand why WSJT is using audio frequency in the Wide Graph window versus using the RF Frequency, similar to that displayed in any "standard waterfall," like that displayed in my Yaesu FT-991A or Icom IC-7300.  Hence, I feel that I must be overlooking some technical aspect as to why audio frequency is being used versus RF frequency.  Folks, what am I missing here?

If RF frequency were displayed (versus audio frequency), you could still time separate the decoding sequences and it would be unnecessary, as it is now,
for the user to need to click on the left side of a FT-8 signal (or other mode) to discover the signals actual RF frequency, because its RF frequency would now be displayed overhead where the audio frequency (in HZ) is presently being displayed.  As it is now, the user must click on the left side of a FT-8 emission symbol to actually discover the FT-8's RF frequency, which will then be displayed in the bottom left "Band Activity" window.  And if RF frequency were displayed in the Wide Graph (versus audio frequency), you could possibly have the the FT-8 (or other digital mode) standard band "calling frequency" centered in the Wide Graph window and be better able to determine at a glance if your signal of interest is within your band boundary limits without having to click on it to determine its RF frequency.  Am I making any sense here or am I missing something basic to my understanding of FT-8 or other modes?

The bottom line is that I'm at loss to understand why WSJT is using audio frequency versus RF frequency in its "Wide Graph" window and won't really feel comfortable "going live" with FT-8 and WSJT until I understand the technical aspects as to why audio frequency is being used.

I very much appreciate any assistance you can offer me in this matter.

Best regards -- Bob  K4SSS
« Last Edit: February 17, 2023, 03:05:41 PM by K4SSS »
Logged

W7XTV

  • Member
  • Posts: 1269

Digital modes are encoded and decoded at audio.  The waterfall display and decoding is processed at audio, not like a spectrum analyzer which does its work at RF.

You set the rig's carrier frequency to where the suppressed carrier would be, just like any other mode that uses the SSB hardware.  WSJT-X does that for you if your rig can be controlled externally, but there's no reason you can't do it yourself.  For example, for 20 meters, the rig would be set to 14074 kHz USB (never LSB for any digital mode, even on 40, 80, or 160). 

The audio frequency that's displayed is the actual tone being sent and received, and is the offset from the suppressed carrier.  That tone is modulated with FT8 info, and will be between 300 and 2400 Hz, although some rigs will go higher in USB DATA mode (mine won't).  No different than sending CW on an SSB rig, really, where the 700 Hz sine wave tone is the offset from the carrier frequency displayed on your rig.  The FT8-modulated tone that you're sending is user-selectable.  Just don't plop your carrier on top of someone else.  The waterfall display is good for finding open places to transmit.

If you're sending a CQ in FT8 on 14074 using a 1600 Hz tone, the actual frequency transmitted is 14075.6, although you don't need to worry about that.  The receive decoder in the software can decode everything in the 300-2400 Hz passband, and it will differentiate between the stations it hears.
Logged
He speaks fluent PSK31, in FT8...  One QSO with him earns you 5BDXCC...  His Wouff Hong has two Wouffs... Hiram Percy Maxim called HIM "The Old Man..."  He is... The Most Interesting Ham In The World!

K0UA

  • Member
  • Posts: 9583

I guess I don't understand why you would want to always know the RF frequency. And you want the audio decoding as wide as the receiver can make it. Hopefully from 200 hz to well past 3000 hz. That is the whole idea of FT8. Panoramic decoding. Decode everything all at once. You sure don't want to screw that down with filters. You WILL miss many calls that way. I suspect some people actually do that, as evidenced by calling and calling and calling a station with no response, yet if it call them on their frequency (which I rarely do for several reasons) then they answer right up. Reason, is now I am in their receiver passband and they actually hear me. To my way of thinking only rookies work a station on their transmit frequency. Where other rookies may be calling if it is a "juicy" piece of DX. Always avoid others. Always look for a "black hole" to transmit in. That can be difficult on a crowded band. Always hold your transmit frequency whenever possible. Sometimes there just isn't a place to get a signal in edgewise in the clear. Sometimes you can find someone with a good signal using the opposite timeslot from what you want to call the DX and that is fine. You will not interfere with each other hopefully. BUT you can never know for sure what is occupied at one end of the world where you are not.
As pointed out above if for some reason you want to know the RF frequency, some simple addition that you can do in your head will tell you that, and yes the program will log the transmit frequency in RF terms, but who cares?
Logged
73  James K0UA

K6BRN

  • Member
  • Posts: 2229

Oddly enough, I do have an independent frequency counter sampling TX.  It came in handy when discovering the FTDX-3000 DTATA offset and also to make sure I'm on-channel on 60M.  But it's more of a "nice to have", whereas the FT8 waterfall is a "Must Have" and reflects what's in-band to my receiver's audio chain. (see above posts).

Brian - K6BRN
Logged

K4SSS

  • Member
  • Posts: 9

Keith and James, thank you both very much for your prompt replies.  I didn't want to make my post lengthy so I tried my best to get to the point in a manner I hoped would be understood.  I do understand that the modes are decoded by audio and that we use USB.  Keith, your mentioning that  the waterfall display is good for finding openings to transmit is in line with what I was trying to solicitate for an answer, and I guess I could have done a better job by just asking what the benefits of the audio window are.  That's my bad and yet you came through anyways.  Tx.  And though I haven't gone hot yet, I've spent a lot of late evening time playing SWL trying to learn the why's of the WSJT methodology, and yes, one of my first observations was, like you said, how the Wide Graph provides excellent info for deconflicting QSO's.  James, your reply was extremely helpful and spot on too, and between both your replies I now have a much better feel and understanding of the important information that the "audio window" provides. Tx.   While I have your attention, can you two recommend some good digital mode books for me to read?  Again, thanks for your assistance guys.  I really appreciate it!  Best regards -- Bob K4SSS   
Logged

VK6HP

  • Member
  • Posts: 1222

Bob,

I think your question demands a bit of higher-level (in the design sense) thinking, if only because there are times when seeing the RF frequency on the WSJT-X display would be useful.  In some of my activity, just to pick an example from some recent FST4W operations, it can be surprisingly hard at 2 am in a cow paddock to add (say) 1433 Hz to 474.2 kHz in order to figure out where to tune the variometer in the MF antenna coupler!  (It's much easier at 2 pm after sufficient coffee).

I think the best justification for the way it's currently done is that the computer+WSJT-X is the real instrument, and it only deals with baseband (audio) data.  So, in the strict sense, that's all it's competent to report on.  Transceivers or other exciters/receivers are of course interfaced to the (say) PC and there is an exchange of information (dial frequency, etc.).  But the PC is far from being any sort of comprehensive embedded controller for the whole system: there are many ways that data can be modified/corrupted and the PC is none the wiser.  A really simple example, again from my shack, is forgetting to turn off the TS-890S RIT and bumping the RIT control:  the audio frequency changes and WSJT-X remains ignorant of the control change.  All it can honestly do is report the audio shift via the wide graph and baseband decode frequency.

I could imagine a more integrated signalling system, maybe an all-digital transceiver with fully embedded controller dispensing with analog audio entirely, in which it would be reasonable to toggle between RF and baseband spectra.

73, Peter.

Logged

W1VT

  • Member
  • Posts: 6067

I was going to say the WSJT-X is open source but it may be more accurate to post links
https://sourceforge.net/projects/wsjt/

License
GNU General Public License version 2.0 (GPLv2), GNU General Public License version 3.0 (GPLv3)

I think it is wonderful that they provide that thus allowing an advanced radio designer to incorporate digital detection without going to audio.
« Last Edit: February 19, 2023, 05:56:29 AM by W1VT »
Logged

K4SSS

  • Member
  • Posts: 9

Hi Peter (VK6HP). 

Thanks for your informative reply.  Since the statue-of-limitations is long behind me now, I will, for the very first time, own up to the fact that I never used a Collins crystal "novice adapter" when using my Dad's (K1RAW) Collins S-Line while he was in Vietnam in 1966-67.  There, I feel better now, hi.  I'll will also never forget my very first CW contact as a Novice, with a ham in Alabama (my QTH was in Massachusetts) and his futilely trying to explain to me to use "DE" instead of "this is" numerous times. I was 11 years old back then and it took awhile for DE to sink in, but I swore after that I was going to learn the ins-and-outs about ham radio before doing something foolish like that again.  This is why I'm taking a cautious approach before jumping headfirst into digital, to hopefully avoid as K0UA has correctly said, "rookie" mistakes. 
One of my concerns was transmitting out-of-band, so RF frequency is important to me.  Many rigs, such as my IC-7300, has the ability to prevent out-of-band transmission.  My Yaesu FT-911A, however, apparently does not, and will only provide a flashing TX led when xmitting out-of-band. Will this ever be a problem?  I don't know. But the established 80M CW-RTTY-DATA Band Plan extends down to the 3.5 MHZ band edge and the possibility exists.  Hence, my thinking that having some RF Transmitting Frequency reference may be of value. Especially if I'm up way beyond my bedtime at 2am and imbibing in a great Islay Single Malt Scotch Whisky.  Hmmm, now that I think about, these automated computer digital modes maybe just the ticket and go hand-in-hand with my other favorite pastime, hi...!  See Peter, you'll never know what you'll learn from Elmers!  It just troubles me that it took me so long to catch on...

On a more technical side, I have noted a few things that seem odd to me when using WSJT with my FT-991A.
A. After transmitting into a dummy load something is changing my selected USB-DATA mode to the RTTY-DATA mode.
B. After transmitting into a dummy load something is changing my selected 3.573 frequency to another frequency.

Does anyone have a clue what may be causing this to happen?  Perhaps I should be using the USB mode?
And please advise me if these issues are better placed in either a new or different thread.

For W1VT:  I'm confused by your last sentence.  Do you mean using "digital detection" to digitally detect, analyze, and translate FT-8 audio, etc., within our PC's?  Isn't that what we are doing now?  Or do you mean something different?  I'm not trying to be some smart ass here but sincerely trying to learn.  Tx

Thanks again to all who have replied. 
Bob  K4SSS
Logged

W1VT

  • Member
  • Posts: 6067

A software defined radio normally produces an audio output, but it could take over the PC functions and act as a WSJT-X decoder with the proper programming.

There used to be a patent that prevented the widespread use of inexpensive PC keyboards that plugged into radios, but it has expired.

Thirty years ago I used to design and build my own SSB/CW radios.  The technology is now obsolete but it was a lot of fun!
« Last Edit: February 19, 2023, 02:58:50 PM by W1VT »
Logged

K6JH

  • Member
  • Posts: 701

Is wsjt audio really that different than the concept of dial or carrier frequency indicated on a SSB transceiver? And when on USB the audio extends higher than the carrier?

I recall once at field day a friend was letting his niece get her feet wet in a contest and she accidentally was transmitting OOB since she didn’t quite get the concept. It earned him an “official observer” notice against his call, and a teaching opportunity for him and his niece!
Logged
73
Jim K6JH

AJ8MH

  • Member
  • Posts: 308
    • homeURL

Oddly enough, I do have an independent frequency counter sampling TX.

Yup, me too.  I pulled an old DD-1 from Oak Hills Research out of the junk pile and put it to good use.  It's tied into my K9AY loops and noise canceller. At times, it will lock on the local AM station about a mile away, also!
Logged

VK6HP

  • Member
  • Posts: 1222


On a more technical side, I have noted a few things that seem odd to me when using WSJT with my FT-991A.


Bob, I don't own an FT-991A and I haven't thought closely about the symptoms you describe but I would encourage you to upgrade your WSJT-X software to the very latest version.  We've been pulling our hair out in our LF/MF group trying to identify a receive frequency shift evident on two FT-991A's but ONLY on 2200 m (136 kHz).  To cut a long story short, a software upgrade fixed the issue.


Is wsjt audio really that different than the concept of dial or carrier frequency indicated on a SSB transceiver?


Jim, it's not much different to SSB in the way it's currently implemented but the path forward is to see the modulation process as redundant.  With versatile and high-quality digital frequency synthesis now ubiquitous, there's no reason why something like WSJT-X (running on the radio or external computer) couldn't directly send frequency shift instructions to the synthesizer. No modulation required, but simply some envelope shaping implemented in some modes.

As one designer noted, audio is the lowest common denominator but my view is that it's getting to the point where the clunkiness is architecturally offensive!  We generate often relatively poor quality JT-mode (and some other) digital signaling in an over-complex box, then amplify it in an inefficient linear amplifier, while desperately trying to dump the heat and not melt power devices - all the time aware that a Class C, D or E implementation would do a better job. Perhaps the versatility of using an SSB signal chain was understandable when digital modes were a relative novelty, but that's hardly the case in 2023.

73, Peter.

Logged

KF5KWO

  • Member
  • Posts: 113

@Bob/K4SSS,

You mention deconfliction of QSOs. I too found it difficult to conscientiously not transmit on top of someone because of the frequently very crowded, almost wall to wall, conditions on FT8. I found that moving to FT4 made it easier to deconflict. Generally, there are not as many people on the FT4 subbands, although that just may have been a reflection of when I was operating at that time. But I often use the graph to select my transmit frequency in order to not transmit on top of anybody and to make sure that my signal can be heard and decoded well on the other end. Granted, the graph is only a reflection of what my antenna hears at my QTH and not necessarily what other hams hear at their respective locations, but of course that issue also exists on any mode because we can't hear everything around the world. Plus, the FT4 transmit times are half of what they are for FT8, so that makes for a quicker and therefore seemingly faster paced operation. For me, FT8 is much slower.

So the time factor, plus the ease of deconfliction that I mentioned make FT4 the preferred mode for me. I also do other digital modes such as HELL, RTTY, PSK, and others in the FLDigi software, but I'll move to FT4 if there's no activity on those modes.

Welcome, good luck, and 73,

Jeff, KF5KWO
Logged

K0UA

  • Member
  • Posts: 9583

If you notice your transmit frequency "shifting" around. And lets say you set it on 3.573 the standard FT8 frequency for 80 meters, that could be because you have your Shift in WSJT-x set to "fake it". AS it should be. Fake it, is a misnomer, it should be labeled "set transmit frequency to keep the audio in the sweet spot of the passband of the modulator of the rig".  But I guess fake it was shorter. The dial frequency will jump around once you leave the 500khz wide segment of audio frequencies around 1500 hz. The reason for that is as stated above to keep the audio centered in the passband. Otherwise roll-off (attenuation) will occur in many/most rigs.  I can't imagine your worry about transmitting out of band when using wsjt-x. I don't know how you would be able to do that.  In other words: "huh?" "what?

I cannot explain the going into USB-RTTY mode either. It should stay in USB-DATA.  The reason to NOT operate in USB is so that the microphone in the rig will be electrically disconnected during transmit. You often hear people with "hot" microphones transmitting FT8 and saying unmentionable things into their microphones that they are unwittingly leaving hot by transmitting on USB. Don't do that.
Logged
73  James K0UA

K6JH

  • Member
  • Posts: 701



Is wsjt audio really that different than the concept of dial or carrier frequency indicated on a SSB transceiver?


Jim, it's not much different to SSB in the way it's currently implemented but the path forward is to see the modulation process as redundant.  With versatile and high-quality digital frequency synthesis now ubiquitous, there's no reason why something like WSJT-X (running on the radio or external computer) couldn't directly send frequency shift instructions to the synthesizer. No modulation required, but simply some envelope shaping implemented in some modes.

As one designer noted, audio is the lowest common denominator but my view is that it's getting to the point where the clunkiness is architecturally offensive!  We generate often relatively poor quality JT-mode (and some other) digital signaling in an over-complex box, then amplify it in an inefficient linear amplifier, while desperately trying to dump the heat and not melt power devices - all the time aware that a Class C, D or E implementation would do a better job. Perhaps the versatility of using an SSB signal chain was understandable when digital modes were a relative novelty, but that's hardly the case in 2023.

73, Peter.



I kind of see where you’re going with this. It might be interesting to you to take a look at the QRP Labs QDX digital transceiver design presentations. It takes the nominal carrier frequency, and measures the sound card frequency, and computes the frequency to program the synthesizer to all on the fly. So in a sense it’s doing what you’re talking about. The disadvantage for the QDX is it only works for FSK modes like WSJT - it won’t work for phase shift modes or amplitude variations. Can’t program rise and fall times like for CW.

But with more sophisticated HW who knows how things will be done in the future. Modern SDR architectures could implement things directly, it’s just that there is so much inertia with the concept of doing things at baseband it will take many years to catch on. Easier to migrate baseband DSP code etc.
Logged
73
Jim K6JH
Pages: [1] 2   Go Up