Theoretical maximum for one FT8 band channel? #FT8


Bruce KX4AZ
 

With the ever-increasing popularity of the FT8 mode, there are times when I am decoding in excess of 36 stations on a single channel (JTAlert display slot maximum).  And I am practically forced to transmit on a slot that straddles two adjacent stations, hoping that increases the chances of being decoded, without impairing the neighbor's decodes.

This has me wondering if anyone has ever tried to estimate the theoretical maximum station capacity of a 0-2500 Hz FT8 channel, beyond a simple non-overlapping calculation (using the individual FT8 signal bandwidths).  Given all of the variables (signal strength, timing, degree of overlap, etc) this would be a challenging calculation, but maybe an idealized case with all the same strengths & 50% overlap might be possible.  And the way the software uses the subtraction & additional decoding passes you can even have stations sharing the same audio channel, which I find amazing.

Lacking a theoretical calculation, it would be interesting to hear what maximum station counts people have observed when a FT8 band is packed full.


Bill Somerville
 

On 21/01/2021 14:11, Bruce KX4AZ wrote:
With the ever-increasing popularity of the FT8 mode, there are times when I am decoding in excess of 36 stations on a single channel (JTAlert display slot maximum).  And I am practically forced to transmit on a slot that straddles two adjacent stations, hoping that increases the chances of being decoded, without impairing the neighbor's decodes.

This has me wondering if anyone has ever tried to estimate the theoretical maximum station capacity of a 0-2500 Hz FT8 channel, beyond a simple non-overlapping calculation (using the individual FT8 signal bandwidths).  Given all of the variables (signal strength, timing, degree of overlap, etc) this would be a challenging calculation, but maybe an idealized case with all the same strengths & 50% overlap might be possible.  And the way the software uses the subtraction & additional decoding passes you can even have stations sharing the same audio channel, which I find amazing.

Lacking a theoretical calculation, it would be interesting to hear what maximum station counts people have observed when a FT8 band is packed full.
Hi Bruce,

we have example .WAV files that decode as many as 60 signals. Of course propagation variations and geographical distribution will mean that many more stations can be actively engaged in QSOs on one SSB width channel World wide. Those with SDRs and other receivers with wider bandwidths can extend their operation to more than a typical 2400 Hz width, others less fortunate can always tune up of down a bit to capture signals of interest, even though they may not be able to decode all signals present at once.

One thing is certain, if everyone obeyed there licence conditions and used the minimum power necessary to complete QSOs there would be even higher occupancy.

73
Bill
G4WJS.


Jon Hall
 

I was interested (horrified) when I read that a an FT8 user was now on his third HF linear....

I really hope that he was pulling my leg.!!!


groups@...
 

On 22/01/2021 09:59, Jon Hall wrote:
I was interested (horrified) when I read that a an FT8 user was now on his third HF linear....
I really hope that he was pulling my leg.!!!
Judging by the disparity of some reports I receive I can quite believe it.

73
Roger
G#4HZA

Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#20924): https://WSJTX.groups.io/g/main/message/20924
Mute This Topic: https://groups.io/mt/80004223/288299
Mute #ft8:https://WSJTX.groups.io/g/main/mutehashtag/ft8
-=-=-
Approved Subject Tags = #AudioIssues #BugReport #Cat_RigControl #ContestMode #FT4 #FT8 #GeneralGroupInfo #install #linux #logging #mac #N1MMlogger #NewUser #QSO_practices #raspberryPi #Timesync #txaudio #udp #WSJTX_config #wsjt-x-crashing Email without a tag is automatically subject to a review by a moderstor.
-=-=-
Group Owner: main+owner@WSJTX.groups.io
Unsubscribe: https://WSJTX.groups.io/g/main/leave/defanged [groups@thebluecastle.org.uk]
-=-=-=-=-=-=-=-=-=-=-=-


Williams, G (af8c) <af8c@...>
 

This same question asked over two months ago without any replies:

11/16/2020, 2:26 PM

Subject: [WSJTX] Upper limit

How many simultaneous QSO transmissions can technically fit in the typical bandwidth (at the same time - any number stacked up two or three layers deep, overlapping others) for FT8 and FT4?

--73, Glenn, AF8C

View/Reply Online (#18392): https://WSJTX.groups.io/g/main/message/18392


On 1/22/2021 5:13 AM, groups via groups.io wrote:
On 22/01/2021 09:59, Jon Hall wrote:
I was interested (horrified) when I read that a an FT8 user was now on his third HF linear....

I really hope that he was pulling my leg.!!!



Judging by the disparity of some reports I receive I can quite believe it.

73
Roger
G#4HZA

Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#20924): https://WSJTX.groups.io/g/main/message/20924
Mute This Topic: https://groups.io/mt/80004223/288299
Mute #ft8:https://WSJTX.groups.io/g/main/mutehashtag/ft8
-=-=-
Approved Subject Tags =  #AudioIssues   #BugReport   #Cat_RigControl  #ContestMode  #FT4  #FT8   #GeneralGroupInfo   #install   #linux  #logging   #mac  #N1MMlogger  #NewUser #QSO_practices  #raspberryPi  #Timesync  #txaudio  #udp   #WSJTX_config   #wsjt-x-crashing Email without a tag is automatically subject to a review by a moderstor.
-=-=-
Group Owner: main+owner@WSJTX.groups.io
Unsubscribe: https://WSJTX.groups.io/g/main/leave/defanged [groups@...]

Virus-free. www.avast.com


K9RX - Gary
 

I routinely see 50+ decodes on busy bands. The most I've seen to date without paying too much attention to it is 58. I'm using a 200 - 3000Hz passband. 

Gary 
K9RX


neil_zampella
 

Unknown ... are you talking those that you can hear, or anywhere?

Figure a 3k bandwidth, then divide that by the width of an FT8 or FT4 signal.       But then you have to account for the fact that the decoders can pull out signals even with some overlap.

But eliminate that, do the math.

Neil, KN3ILZ

On 1/22/2021 7:55 AM, Williams, G (af8c) via groups.io wrote:

This same question asked over two months ago without any replies:

11/16/2020, 2:26 PM

Subject: [WSJTX] Upper limit

How many simultaneous QSO transmissions can technically fit in the typical bandwidth (at the same time - any number stacked up two or three layers deep, overlapping others) for FT8 and FT4?

--73, Glenn, AF8C

View/Reply Online (#18392): https://WSJTX.groups.io/g/main/message/18392


On 1/22/2021 5:13 AM, groups via groups.io wrote:
On 22/01/2021 09:59, Jon Hall wrote:
I was interested (horrified) when I read that a an FT8 user was now on his third HF linear....

I really hope that he was pulling my leg.!!!



Judging by the disparity of some reports I receive I can quite believe it.

73
Roger
G#4HZA

Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#20924): https://WSJTX.groups.io/g/main/message/20924
Mute This Topic: https://groups.io/mt/80004223/288299
Mute #ft8:https://WSJTX.groups.io/g/main/mutehashtag/ft8
-=-=-
Approved Subject Tags =  #AudioIssues   #BugReport   #Cat_RigControl  #ContestMode  #FT4  #FT8   #GeneralGroupInfo   #install   #linux  #logging   #mac  #N1MMlogger  #NewUser #QSO_practices  #raspberryPi  #Timesync  #txaudio  #udp   #WSJTX_config   #wsjt-x-crashing Email without a tag is automatically subject to a review by a moderstor.
-=-=-
Group Owner: main+owner@WSJTX.groups.io
Unsubscribe: https://WSJTX.groups.io/g/main/leave/defanged [groups@...]

Virus-free. www.avast.com




Jon Ermels
 

58 is the top here also.

73 de NØIGU Jon


On Friday, January 22, 2021, 08:05:18 AM CST, K9RX - Gary <amateurk9rx@...> wrote:


I routinely see 50+ decodes on busy bands. The most I've seen to date without paying too much attention to it is 58. I'm using a 200 - 3000Hz passband. 

Gary 
K9RX




Martin Lafferty
 

I routinely get 50+ here, but did get 65 one time.

73
Martin W4PNY

On Friday, January 22, 2021, 10:51:36 AM EST, Jon Ermels via groups.io <n0igu@...> wrote:


58 is the top here also.

73 de NØIGU Jon


On Friday, January 22, 2021, 08:05:18 AM CST, K9RX - Gary <amateurk9rx@...> wrote:


I routinely see 50+ decodes on busy bands. The most I've seen to date without paying too much attention to it is 58. I'm using a 200 - 3000Hz passband. 

Gary 
K9RX







Reino Talarmo
 

One 72 h sample with two antennas and counting only stations received by both antenna at the same time I got on 20 m band with bandwidth

Typical number of spots 37 was 358 times

Spots instances more than 180 times were at range 30 to 48

Maximum number of spots 67 and  60 or more were 20 times

Total spots were 290 000.

 

Antennas were turnstiles at low elevation.

 

A potential maximum could be 2500/50 = 50 “slots” and each containing 2 stations would be 100. Well, Probability of that special instance is very small. My guess is that probable limit is about 70 to 73.

 

73, Reino OH3mA

 

From: main@WSJTX.groups.io [mailto:main@WSJTX.groups.io] On Behalf Of Jon Ermels via groups.io
Sent: 22. tammikuuta 2021 17:51
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] Theoretical maximum for one FT8 band channel? #FT8

 

58 is the top here also.

 

73 de NØIGU Jon

 

 

On Friday, January 22, 2021, 08:05:18 AM CST, K9RX - Gary <amateurk9rx@...> wrote:

 

 

I routinely see 50+ decodes on busy bands. The most I've seen to date without paying too much attention to it is 58. I'm using a 200 - 3000Hz passband. 

Gary 
K9RX




Reino Talarmo
 

[Updated message]

My numbers below were for a single antenna case without any comparison with another antenna,

For two antenna case I had once 61 in about 24 h measurement time.

 

73, Reino OH3mA

 

From: main@WSJTX.groups.io [mailto:main@WSJTX.groups.io] On Behalf Of Reino Talarmo
Sent: 22. tammikuuta 2021 19:39
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] Theoretical maximum for one FT8 band channel? #FT8

 

One 72 h sample with two antennas and counting only stations received by both antenna at the same time I got on 20 m band with bandwidth

Typical number of spots 37 was 358 times

Spots instances more than 180 times were at range 30 to 48

Maximum number of spots 67 and  60 or more were 20 times

Total spots were 290 000.

 

Antennas were turnstiles at low elevation.

 

A potential maximum could be 2500/50 = 50 “slots” and each containing 2 stations would be 100. Well, Probability of that special instance is very small. My guess is that probable limit is about 70 to 73.

 

73, Reino OH3mA

 

From: main@WSJTX.groups.io [mailto:main@WSJTX.groups.io] On Behalf Of Jon Ermels via groups.io
Sent: 22. tammikuuta 2021 17:51
To: main@WSJTX.groups.io
Subject: Re: [WSJTX] Theoretical maximum for one FT8 band channel? #FT8

 

58 is the top here also.

 

73 de NØIGU Jon

 

 

On Friday, January 22, 2021, 08:05:18 AM CST, K9RX - Gary <amateurk9rx@...> wrote:

 

 

I routinely see 50+ decodes on busy bands. The most I've seen to date without paying too much attention to it is 58. I'm using a 200 - 3000Hz passband. 

Gary 
K9RX

 


Ken WB8UFC
 

On Thu, Jan 21, 2021 at 09:21 AM, Bill Somerville wrote:
One thing is certain, if everyone obeyed there licence conditions and used the minimum power necessary to complete QSOs there would be even higher occupancy.
Perhaps an option to reduce or increase the Pwr setting based on reported dB by the station being worked?
I would definitely check that option!

Doing this on a per QSO basis manually due to the short duration of the QSOs is... unlikely to occur given the nature of human operators.
I make a good faith effort by starting at low power on a band and increasing power as necessary, but even this is not ideal. 


JP Tucson, AZ <samcat88az@...>
 

Hi Ken,

In a perfect world, that would be nice...

However, in reality, we are dealing with physics!  And the physics of radio propagation means that rarely are 2 signals likely to be seeing the same conditions over even a short time period. Meaning...  fading. 

I, no... WE have all seen signals that go from +2dB to -18dB in a 30 second time period.

With that in mind, if you automatically turn down the power, you will almost always lose a QSO.  Because the other factor is that you have really no idea how much to turn down the power!  Remember, it's a logarithmic scale too.

 

73 - John - N7GHZ


On Fri, Jan 22, 2021, 12:50 PM Ken WB8UFC <wb8ufc@...> wrote:
On Thu, Jan 21, 2021 at 09:21 AM, Bill Somerville wrote:
One thing is certain, if everyone obeyed there licence conditions and used the minimum power necessary to complete QSOs there would be even higher occupancy.
Perhaps an option to reduce or increase the Pwr setting based on reported dB by the station being worked?
I would definitely check that option!

Doing this on a per QSO basis manually due to the short duration of the QSOs is... unlikely to occur given the nature of human operators.
I make a good faith effort by starting at low power on a band and increasing power as necessary, but even this is not ideal. 



kingep@...
 

But surely we will not be forever limited to a single 3 kHz FT8 band, given the popularity?
Peter, ve6epk


From: "neil_zampella" <neilz@...>
To: main@WSJTX.groups.io
Sent: Friday, January 22, 2021 7:06:02 AM
Subject: Re: [WSJTX] Theoretical maximum for one FT8 band channel? #FT8

Unknown ... are you talking those that you can hear, or anywhere?

Figure a 3k bandwidth, then divide that by the width of an FT8 or FT4 signal.       But then you have to account for the fact that the decoders can pull out signals even with some overlap.

But eliminate that, do the math.

Neil, KN3ILZ

On 1/22/2021 7:55 AM, Williams, G (af8c) via groups.io wrote:

This same question asked over two months ago without any replies:

11/16/2020, 2:26 PM

Subject: [WSJTX] Upper limit

How many simultaneous QSO transmissions can technically fit in the typical bandwidth (at the same time - any number stacked up two or three layers deep, overlapping others) for FT8 and FT4?

--73, Glenn, AF8C

View/Reply Online (#18392): https://WSJTX.groups.io/g/main/message/18392


On 1/22/2021 5:13 AM, groups via groups.io wrote:
On 22/01/2021 09:59, Jon Hall wrote:
I was interested (horrified) when I read that a an FT8 user was now on his third HF linear....

I really hope that he was pulling my leg.!!!



Judging by the disparity of some reports I receive I can quite believe it.

73
Roger
G#4HZA

Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#20924): https://WSJTX.groups.io/g/main/message/20924
Mute This Topic: https://groups.io/mt/80004223/288299
Mute #ft8:https://WSJTX.groups.io/g/main/mutehashtag/ft8
-=-=-
Approved Subject Tags =  #AudioIssues   #BugReport   #Cat_RigControl  #ContestMode  #FT4  #FT8   #GeneralGroupInfo   #install   #linux  #logging   #mac  #N1MMlogger  #NewUser #QSO_practices  #raspberryPi  #Timesync  #txaudio  #udp   #WSJTX_config   #wsjt-x-crashing Email without a tag is automatically subject to a review by a moderstor.
-=-=-
Group Owner: main+owner@WSJTX.groups.io
Unsubscribe: https://WSJTX.groups.io/g/main/leave/defanged [groups@...]


Virus-free. www.avast.com



    




Ken WB8UFC
 

"With that in mind, if you automatically turn down the power, you will almost always lose a QSO.  Because the other factor is that you have really no idea how much to turn down the power!  Remember, it's a logarithmic scale too."

Upon what basis do you set your output power if not your reported signal strength?

When using JS8Call, I do exactly that to great effect and those QSOs are typically 20 - 30 times the duration of an FT8 QSO.

An obvious heuristic to address signal fade would be to retry the next transmission at say 3dB greater output if the last transmission were not confirmed.

I would expect that FT8's resilience to signal strength variation would make it an ideal candidate for this.

What am I overlooking?


Ken WB8UFC
 

"But surely we will not be forever limited to a single 3 kHz FT8 band, given the popularity?"

I'm fairly new to FT8, but I'm told that during contests and other busy times the alternate frequencies here are utilized:
https://www.sigidwiki.com/wiki/FT8


K9RX - Gary
 

Out of curiosity I checked out FT8 on 20 yesterday .... I was seeing the typical 50's decodes per sequence.... and I wondered about someone that said he had seen 67 - but that he was using a non-gain antenna.... I have 10 elements on 20 meters, 2 50' boom stacked yagis at 135' and 75' ... so the angle is pretty low and there is lots of signal rejection other than straight forward.... so I switched to the lower antenna - a higher AOR ... and BINGO 64 decodes! Then a bit later I saw 67 and then 74. 74 decodes in one sequence. That is using 200 - 3000Hz passband. Amazing. That stated there will need to be something figured out on how to double the available window, at least. 

Gary 
K9RX


David Ackrill
 

On 6M the solution has been another "DX" frequency.  So now there is 50.313MHz and 50.323MHz and, yes, the DX Police have been seen to be active on 323 as well...

It's like 50.110MHz all over again.  LOL.

On Tue, 26 Jan 2021 at 13:45, K9RX - Gary <amateurK9RX@...> wrote:
Out of curiosity I checked out FT8 on 20 yesterday .... I was seeing the typical 50's decodes per sequence.... and I wondered about someone that said he had seen 67 - but that he was using a non-gain antenna.... I have 10 elements on 20 meters, 2 50' boom stacked yagis at 135' and 75' ... so the angle is pretty low and there is lots of signal rejection other than straight forward.... so I switched to the lower antenna - a higher AOR ... and BINGO 64 decodes! Then a bit later I saw 67 and then 74. 74 decodes in one sequence. That is using 200 - 3000Hz passband. Amazing. That stated there will need to be something figured out on how to double the available window, at least. 

Gary 
K9RX




--
David Ackrill