Topics

K3S NR

Jim Denneny
 

I run WSJT on TB.  I use 2.8 khz bandwidth. QRN can run S7 to S9 making it difficult to decode weak signals.  I prefer to use TXDATA mode but NR cannot be engaged in this mode.  If I want to use NR, I must use USB mode. Why can't I use NR in TXDATA mode?

Jim K7EG

Bob McGraw - K4TAX
 

The WSJT software suggests one should not run NR on data modes due to the fact it typically modifies the data stream.   Decoding difficulty then occurs.   The software is designed to work in an environment where QRN exists.   At the same time NB is available in TX DATA mode.   However the action of the NB is typically for repetitive impulse noise whereas QRN is usually random and of longer duration.

73

Bob, K4TAX

On 8/25/2019 4:27 PM, Jim Denneny wrote:
I run WSJT on TB.  I use 2.8 khz bandwidth. QRN can run S7 to S9 making it difficult to decode weak signals.  I prefer to use TXDATA mode but NR cannot be engaged in this mode.  If I want to use NR, I must use USB mode. Why can't I use NR in TXDATA mode?

Jim K7EG

Bob McGraw - K4TAX
 


On 8/25/2019 4:43 PM, Bob McGraw - K4TAX wrote:

The WSJT software suggests one should not run NR on data modes due to the fact it typically modifies the data stream.   Decoding difficulty then occurs.   The software is designed to work in an environment where QRN exists.   At the same time NB is available in TX DATA mode.   However the action of the NB is typically for repetitive impulse noise whereas QRN is usually random and of longer duration. 

Perhaps better use is the implementation of the Attenuator function and RF Gain reduction to improve S/N ratio performance.

73

Bob, K4TAX

On 8/25/2019 4:27 PM, Jim Denneny wrote:
I run WSJT on TB.  I use 2.8 khz bandwidth. QRN can run S7 to S9 making it difficult to decode weak signals.  I prefer to use TXDATA mode but NR cannot be engaged in this mode.  If I want to use NR, I must use USB mode. Why can't I use NR in TXDATA mode?

Jim K7EG

Rick Bates, NK7I
 

As one that made DXCC last (northern hemisphere) winter on TB, hearing flea gas (near ESP) in a wind storm (QRN) signals usually means running the rx wide open and letting the FT8 software sort it. 

NB can help. 

Rick WA6NHC. (109 worked on TB, awaiting more QSL)

Smell Czech correction happen

On Aug 25, 2019, at 2:46 PM, Bob McGraw - K4TAX <rmcgraw@...> wrote:


On 8/25/2019 4:43 PM, Bob McGraw - K4TAX wrote:

The WSJT software suggests one should not run NR on data modes due to the fact it typically modifies the data stream.   Decoding difficulty then occurs.   The software is designed to work in an environment where QRN exists.   At the same time NB is available in TX DATA mode.   However the action of the NB is typically for repetitive impulse noise whereas QRN is usually random and of longer duration. 

Perhaps better use is the implementation of the Attenuator function and RF Gain reduction to improve S/N ratio performance.

73

Bob, K4TAX

On 8/25/2019 4:27 PM, Jim Denneny wrote:
I run WSJT on TB.  I use 2.8 khz bandwidth. QRN can run S7 to S9 making it difficult to decode weak signals.  I prefer to use TXDATA mode but NR cannot be engaged in this mode.  If I want to use NR, I must use USB mode. Why can't I use NR in TXDATA mode?

Jim K7EG

Jim Denneny
 

Thanks Bob.  I will stick with TXDATA & NB
Jim K7EG

Bob McGraw - K4TAX
 

Jim:

You might give the ATTN and RF Gain some work to do.   With information provided by a highly respected ham and one that evaluates a lot of radios, in a conversation and discussion we had at the Huntsville hamfest a couple of years ago,  he suggested that most hams use too much RF Gain for typical band conditions.   In his words, for legacy receivers such as the K3 series and other like radios the optimum receiver performance occurs when the no signal band noise is some 10 dB above the receiver noise floor.

How does one accomplish this in as much as we have no control over receiver noise floor?   The solution is then to attenuate the signal and noise arriving at the antenna input.  The attenuator and RF gain is used as a step in this direction.  

To do a bit of math;  the noise floor of the K3 series radio is typically about -127 dBm.  This is equivalent of S-0.   Now at our QTH with the antenna of choice the no signal band noise is S-4 or -103 dBm.  The difference between the two values is 24 dB.  Now to have the noise appear 10 dB above the noise floor is should be at a level of -117 dBm.   The optimum level is then a difference of -117 dBm {RX noise floor + 10 dB}  minus -103 dBm or 14 dB.  Thus 15 dB of attenuation brings the configuration to within 1 dB of the optimum receiver noise floor to band noise. 

In some cases, additional signal / noise reduction is needed in the case where the no signal band noise is higher.   Again an example:   Noise floor is -127 dBm,  No signal band noise is S-7 or -85 dBm.  Thus the different is 32 dB minus our desired 10 dB.   Here we need RF gain reduction of 17 dB or about 3 S units, plus 15 dB of attenuation.

Once the method is applied, we then find that signals pop out of the noise.  And in this case the actual receiver noise only adds about 0.5 dB to the total noise value.  I've tried it on several receivers including my K3S, my Tentec Eagle, my Tentec Omni VII and my KW TS-590S.   It really does work.   It is just difficult to get past the mental notion that more gain is needed to hear weak signals.   Where in practice, less noise is required to hear weak signals.

73

Bob, K4TAX


On 8/25/2019 8:06 PM, Jim Denneny wrote:
Thanks Bob.  I will stick with TXDATA & NB
Jim K7EG

Carl Moreschi
 

For best results on FT8, you should set AGC to OFF, and adjust RF gain for a 30 db signal in WSJT with no signal being received. If some signals push the signal above 80 db, you may need to lower the RF gain some.

WSJT does all the work here so let it do its job.

Also, NB can help but certainly not NR.

Carl Moreschi N4PY
58 Hogwood Rd
Louisburg, NC 27549
www.n4py.com

On 8/25/2019 10:06 PM, Bob McGraw - K4TAX wrote:
Jim:

You might give the ATTN and RF Gain some work to do. With information
provided by a highly respected ham and one that evaluates a lot of
radios, in a conversation and discussion we had at the Huntsville
hamfest a couple of years ago, he suggested that most hams use too much
RF Gain for typical band conditions. In his words, for legacy
receivers such as the K3 series and other like radios the optimum
receiver performance occurs when the no signal band noise is some 10 dB
above the receiver noise floor.

How does one accomplish this in as much as we have no control over
receiver noise floor? The solution is then to attenuate the signal and
noise arriving at the antenna input. The attenuator and RF gain is used
as a step in this direction.

To do a bit of math; the noise floor of the K3 series radio is
typically about -127 dBm. This is equivalent of S-0. Now at our QTH
with the antenna of choice the no signal band noise is S-4 or -103 dBm.
The difference between the two values is 24 dB. Now to have the noise
appear 10 dB above the noise floor is should be at a level of -117
dBm. The optimum level is then a difference of -117 dBm {RX noise
floor + 10 dB} minus -103 dBm or 14 dB. Thus 15 dB of attenuation
brings the configuration to within 1 dB of the optimum receiver noise
floor to band noise.

In some cases, additional signal / noise reduction is needed in the case
where the no signal band noise is higher. Again an example: Noise
floor is -127 dBm, No signal band noise is S-7 or -85 dBm. Thus the
different is 32 dB minus our desired 10 dB. Here we need RF gain
reduction of 17 dB or about 3 S units, plus 15 dB of attenuation.

Once the method is applied, we then find that signals pop out of the
noise. And in this case the actual receiver noise only adds about 0.5
dB to the total noise value. I've tried it on several receivers
including my K3S, my Tentec Eagle, my Tentec Omni VII and my KW
TS-590S. It really does work. It is just difficult to get past the
mental notion that more gain is needed to hear weak signals. Where in
practice, less noise is required to hear weak signals.

73

Bob, K4TAX


On 8/25/2019 8:06 PM, Jim Denneny wrote:
Thanks Bob. I will stick with TXDATA & NB
Jim K7EG

Bob McGraw - K4TAX
 

Why not just reduce the audio level via the computer CODEC level and be done with it?   That doesn't affect the receiver performance.

Agreed on the NB works well.

73

Bob, K4TAX

On 8/25/2019 9:12 PM, Carl Moreschi wrote:
For best results on FT8, you should set AGC to OFF, and adjust RF gain for a 30 db signal in WSJT with no signal being received.  If some signals push the signal above 80 db, you may need to lower the RF gain some.

WSJT does all the work here so let it do its job.

Also, NB can help but certainly not NR.

Carl Moreschi N4PY
58 Hogwood Rd
Louisburg, NC 27549
www.n4py.com

On 8/25/2019 10:06 PM, Bob McGraw - K4TAX wrote:
Jim:

You might give the ATTN and RF Gain some work to do.   With information
provided by a highly respected ham and one that evaluates a lot of
radios, in a conversation and discussion we had at the Huntsville
hamfest a couple of years ago,  he suggested that most hams use too much
RF Gain for typical band conditions.   In his words, for legacy
receivers such as the K3 series and other like radios the optimum
receiver performance occurs when the no signal band noise is some 10 dB
above the receiver noise floor.

How does one accomplish this in as much as we have no control over
receiver noise floor?   The solution is then to attenuate the signal and
noise arriving at the antenna input.  The attenuator and RF gain is used
as a step in this direction.

To do a bit of math;  the noise floor of the K3 series radio is
typically about -127 dBm.  This is equivalent of S-0.   Now at our QTH
with the antenna of choice the no signal band noise is S-4 or -103 dBm.
The difference between the two values is 24 dB.  Now to have the noise
appear 10 dB above the noise floor is should be at a level of -117
dBm.   The optimum level is then a difference of -117 dBm {RX noise
floor + 10 dB}  minus -103 dBm or 14 dB.  Thus 15 dB of attenuation
brings the configuration to within 1 dB of the optimum receiver noise
floor to band noise.

In some cases, additional signal / noise reduction is needed in the case
where the no signal band noise is higher.   Again an example: Noise
floor is -127 dBm,  No signal band noise is S-7 or -85 dBm. Thus the
different is 32 dB minus our desired 10 dB.   Here we need RF gain
reduction of 17 dB or about 3 S units, plus 15 dB of attenuation.

Once the method is applied, we then find that signals pop out of the
noise.  And in this case the actual receiver noise only adds about 0.5
dB to the total noise value.  I've tried it on several receivers
including my K3S, my Tentec Eagle, my Tentec Omni VII and my KW
TS-590S.   It really does work.   It is just difficult to get past the
mental notion that more gain is needed to hear weak signals. Where in
practice, less noise is required to hear weak signals.

73

Bob, K4TAX


On 8/25/2019 8:06 PM, Jim Denneny wrote:
Thanks Bob.  I will stick with TXDATA & NB
Jim K7EG

Jim Denneny
 

Bob

Thanks again for this procedure.  I really appreciate this.  I will practice with this.

73, Jim K7EG

Carl Moreschi
 

Another big thing when using FT8 is to call the station on a clear frequency. Look at the wide graph and press and hold the shift key. Then mouse click on a clear frequency. This greatly increases your chances of the station hearing you. The shift key held while you click just moves your transmit frequency.

Carl Moreschi N4PY
58 Hogwood Rd
Louisburg, NC 27549
www.n4py.com

On 8/26/2019 11:46 AM, Jim Denneny wrote:
Bob

Thanks again for this procedure. I really appreciate this. I will
practice with this.

73, Jim K7EG

Bob McGraw - K4TAX
 

I'd like a bit of feedback on your experiences. 
Tnx in advance. 
73
Bob, K4TAX



On Aug 26, 2019, at 10:46 AM, Jim Denneny <57JNDenneny@...> wrote:

Bob

Thanks again for this procedure.  I really appreciate this.  I will practice with this.

73, Jim K7EG

Rick Bates, NK7I
 

True.  But that premise is based on what YOU hear, not the station you're calling might be hearing (a typical issue in ham radio).

So if after a (half) dozen calls with no response, change to another 'blank' tone frequency (that means you wait one cycle to ensure you're not competing) and try again.  They're listening (as you are) on a 2.7 kHz bandwidth (at least) so if you move as they call you, it doesn't really affect things.

Rick nhc

On 8/26/2019 8:54 AM, Carl Moreschi wrote:
Another big thing when using FT8 is to call the station on a clear frequency.  Look at the wide graph and press and hold the shift key. Then mouse click on a clear frequency.  This greatly increases your chances of the station hearing you.  The shift key held while you click just moves your transmit frequency.

Carl Moreschi N4PY
58 Hogwood Rd
Louisburg, NC 27549
www.n4py.com

Steve Sacco
 

I very typically run NR on my FTdx5000 when I'm running WSJT-X.  I can have a profound effect - I've seen 6-8 dB of S/N improvement in some circumstances.

73,
Steve
NN4X

Bob McGraw - K4TAX
 

So what would be the lowest DB report decoded using this method?  
I normally see lowest values of -24 dB. 
With 6dB improvement do you decide to -30 or so? 

Bob, K4TAX



On Aug 26, 2019, at 3:08 PM, Steve Sacco via Groups.Io <nn4x@...> wrote:

I very typically run NR on my FTdx5000 when I'm running WSJT-X.  I can have a profound effect - I've seen 6-8 dB of S/N improvement in some circumstances.

73,
Steve
NN4X

Rob Sherwood
 

FT8 won’t decide to -30 dB.  Rob, NC0B


On Aug 26, 2019, at 2:21 PM, Bob McGraw - K4TAX <rmcgraw@...> wrote:

So what would be the lowest DB report decoded using this method?  
I normally see lowest values of -24 dB. 
With 6dB improvement do you decide to -30 or so? 

Bob, K4TAX



On Aug 26, 2019, at 3:08 PM, Steve Sacco via Groups.Io <nn4x@...> wrote:

I very typically run NR on my FTdx5000 when I'm running WSJT-X.  I can have a profound effect - I've seen 6-8 dB of S/N improvement in some circumstances.

73,
Steve
NN4X

Bob McGraw - K4TAX
 

I didn't think so. 

Bob, K4TAX



On Aug 26, 2019, at 4:18 PM, Rob Sherwood <rob@...> wrote:

FT8 won’t decide to -30 dB.  Rob, NC0B


On Aug 26, 2019, at 2:21 PM, Bob McGraw - K4TAX <rmcgraw@...> wrote:

So what would be the lowest DB report decoded using this method?  
I normally see lowest values of -24 dB. 
With 6dB improvement do you decide to -30 or so? 

Bob, K4TAX



On Aug 26, 2019, at 3:08 PM, Steve Sacco via Groups.Io <nn4x@...> wrote:

I very typically run NR on my FTdx5000 when I'm running WSJT-X.  I can have a profound effect - I've seen 6-8 dB of S/N improvement in some circumstances.

73,
Steve
NN4X

Rob Sherwood
 

JT9 that I use on 630 meters decodes to -28, and WSPR to -30.  Rob, NC0B

 

From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Bob McGraw - K4TAX
Sent: Monday, August 26, 2019 3:35 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] K3S NR

 

I didn't think so. 

Bob, K4TAX

 

 


On Aug 26, 2019, at 4:18 PM, Rob Sherwood <rob@...> wrote:

FT8 won’t decide to -30 dB.  Rob, NC0B


On Aug 26, 2019, at 2:21 PM, Bob McGraw - K4TAX <rmcgraw@...> wrote:

So what would be the lowest DB report decoded using this method?  

I normally see lowest values of -24 dB. 
With 6dB improvement do you decide to -30 or so? 

 

Bob, K4TAX

 

 


On Aug 26, 2019, at 3:08 PM, Steve Sacco via Groups.Io <nn4x@...> wrote:

I very typically run NR on my FTdx5000 when I'm running WSJT-X.  I can have a profound effect - I've seen 6-8 dB of S/N improvement in some circumstances.

 

73,

Steve

NN4X

 

Doug AA3S
 

or transmit on the frequency of the station you're want, presumably he picked a clear frequency in his receiver (not my brainstorm, heard that first from W3LPL).

73, Doug AA3S

On 8/26/2019 3:09 PM, Rick WA6NHC wrote:
True.?? But that premise is based on what YOU hear, not the station you're calling might be hearing (a typical issue in ham radio).

So if after a (half) dozen calls with no response, change to another 'blank' tone frequency (that means you wait one cycle to ensure you're not competing) and try again.?? They're listening (as you are) on a 2.7 kHz bandwidth (at least) so if you move as they call you, it doesn't really affect things.

Rick nhc


On 8/26/2019 8:54 AM, Carl Moreschi wrote:
Another big thing when using FT8 is to call the station on a clear frequency.?? Look at the wide graph and press and hold the shift key. Then mouse click on a clear frequency.?? This greatly increases your chances of the station hearing you.?? The shift key held while you click just moves your transmit frequency.

Carl Moreschi N4PY
58 Hogwood Rd
Louisburg, NC 27549
www.n4py.com





Rick Bates, NK7I
 

I’ve always felt that is a poor choice; FT8 needs (from my observation) at least a 5 cycle variance from other stations attempting to use the caller frequency, for decoding; else neither decode. 

Rick WA6NHC

Smell Czech correction happens

On Aug 26, 2019, at 1:35 PM, Doug via Groups.Io <theharts04@...> wrote:

or transmit on the frequency of the station you're want, presumably he picked a clear frequency in his receiver (not my brainstorm, heard that first from W3LPL).

73, Doug AA3S

Jim Denneny
 

Thanks for the shortcut tip Carl