locked Optimum RECEIVE Level dB setting. WSJT-X for FT8 and Q65 #AudioIssues #FT8 #Q65


invl160 Davis
 

I am using a TS590SG with version 2.4.0 rc4.  I have been playing with Q65.  Upon not decoding his transmission a friend asked me recently what my receive level setting was on the left hand "level meter".  I said 60.  He said "thats way too high".

There are a number of settings that can "vary " the receive level meter on a quiet band.
1.) The USB Port Output Level in TS590SG menu #72
2.) The RF Gain setting - with AGC OFF
3.) The dB slider position in Win 10 Sound Control panel - Recording tab / Level Slider

I typically run the RECEIVE level 9on FT8) as follows between 50-60 dB
1.) Menu #72 at 1 or 2
2)  RF Gain fully CW with AGC OFF
3)  Win 10 Sound Control/Recording/Level Slider at 0dB or close to it.

The User Guide Section 5 Transceiver Setup says "Use the receiver gain controls and/or the computer’s audio mixer controls to set the background noise level  to around 30 dB when no signals are present. It is usually best to turn AGC off or reduce the RF gain control to minimize AGC action."

In order to set the level at 30db in my case the Menu# 72 setting must be at 1 (the minimum setting) and the Sound Control slider must be at -19dB.

1.- Is there a common practice out there regarding the level control? 
2.- Does the magnitude of the RECEIVE level control impact on the DECODE reliability of the software.? 
3. - Is the decode reliability negatively impacted by setting the level control to 50-60dB?
4 -  Is the setting for Receive level specific to mode  i.e. 30dB  for Q65  ;  50dB for FT8 etc.

73 Frank VO1HP


Martin G0HDB
 

On Tue, May 11, 2021 at 01:07 AM, invl160 Davis wrote:
I am using a TS590SG with version 2.4.0 rc4.  I have been playing with Q65.  Upon not decoding his transmission a friend asked me recently what my receive level setting was on the left hand "level meter".  I said 60.  He said "thats way too high".

Hi Frank, I don't know how the TS590SG behaves but with some rigs, eg. my Icom IC-7610 (and my previous IC-7600), when the AGC is off the audio output from the rig becomes distorted when even modest-strength signals (eg. -85dBm, S3-4) are present at the antenna input.  The presence of any distortion artefacts on the audio being fed from the rig into WSJT-X can really screw up the decoding.

I've found, as has Rob NC0B, that it's best to operate our Icom rigs with the AGC on (with a fast time constant - I use a setting of 0.8secs) and set the various rig and Windows audio level controls so that the WSJT-X audio level 'thermometer' sits at approx the 60dB level when receiving only band noise.  The rig's AGC will ensure that the level of the audio coming out of the rig and going into WSJT-X nevers gets near the clipping threshold - on my 7610 I don't recall ever seeing a reading on the WSJT-X 'thermometer' of more than about 74dB.

You could try operating with your TS590SG's AGC on and setting the audio level controls so that you have a reading of around the 60dB mark, as you have at present, on the WSJT-X 'thermometer'.  At least then you can be confident that you're (almost!) never likely to suffer from the presence of any distortion artefacts on the audio coming out of the rig and going into WSJT-X.

73
--
Martin G0HDB


Bill Somerville
 

On 11/05/2021 19:27, Martin G0HDB wrote:
On Tue, May 11, 2021 at 01:07 AM, invl160 Davis wrote:
I am using a TS590SG with version 2.4.0 rc4.  I have been playing with Q65.  Upon not decoding his transmission a friend asked me recently what my receive level setting was on the left hand "level meter".  I said 60.  He said "thats way too high".

Hi Frank, I don't know how the TS590SG behaves but with some rigs, eg. my Icom IC-7610 (and my previous IC-7600), when the AGC is off the audio output from the rig becomes distorted when even modest-strength signals (eg. -85dBm, S3-4) are present at the antenna input.  The presence of any distortion artefacts on the audio being fed from the rig into WSJT-X can really screw up the decoding.

I've found, as has Rob NC0B, that it's best to operate our Icom rigs with the AGC on (with a fast time constant - I use a setting of 0.8secs) and set the various rig and Windows audio level controls so that the WSJT-X audio level 'thermometer' sits at approx the 60dB level when receiving only band noise.  The rig's AGC will ensure that the level of the audio coming out of the rig and going into WSJT-X nevers gets near the clipping threshold - on my 7610 I don't recall ever seeing a reading on the WSJT-X 'thermometer' of more than about 74dB.

You could try operating with your TS590SG's AGC on and setting the audio level controls so that you have a reading of around the 60dB mark, as you have at present, on the WSJT-X 'thermometer'.  At least then you can be confident that you're (almost!) never likely to suffer from the presence of any distortion artefacts on the audio coming out of the rig and going into WSJT-X.

73
--
Martin G0HDB

Martin,

it should be no surprise that with no AGC strong signals will overload your receiver, it is your responsibility to adjust the RF GAIN to compensate. The advantage of using the RF GAIN instead of the rig's AGC is that you can consider the weakest signals whereas the AGC is only set by the strongest signal.

73
Bill
G4WJS.


Hasan Schiers N0AN
 

Frank,
I'm using a TS-590sg, and have been for several years. In terms of the USB output, I also set it at 1, or it has too much output. 

With the audio out (USB) at 1 (from  the 590sg menu), and windows set to 0 dB as Bill suggests, I get the following readings on the WSJT-X rx audio thermometer:

1. AGC OFF: 54 dB
2. AGC ON:  45 dB

This is on 6m with no signal, but the antenna noise and an ARR GAsFET preamp (external) and the internal preamp of the 590sg turned off.

For weak signal work the AGC should be OFF, unless you are overwhelmed by local competing signals, in which case you either have to ride the RF gain , or put AGC on FAST.

MSK144 the AGC can be off all the time and even distorted pings seem to decode (again, unless local and silly loud.

Q65, which we run for an hour every morning on 6m,  can  be left with AGC off nearly all the time, as it is not busy, but very big signals will distort to the point of killing decodes, and if that happens, simply ride the RF gain, or turn AGC ON.

For 6m, I run NB 1 and 2 both on and both set for maximum blanking. No problems with either MSK144 or Q65 in terms of NB causing issues. FT8, high Noise Blanking will produce images of the signal (multiple decodes).

Hope this  helps ya. 73, N0AN

Hasan


On Mon, May 10, 2021 at 7:07 PM invl160 Davis <invl160@...> wrote:
I am using a TS590SG with version 2.4.0 rc4.  I have been playing with Q65.  Upon not decoding his transmission a friend asked me recently what my receive level setting was on the left hand "level meter".  I said 60.  He said "thats way too high".

There are a number of settings that can "vary " the receive level meter on a quiet band.
1.) The USB Port Output Level in TS590SG menu #72
2.) The RF Gain setting - with AGC OFF
3.) The dB slider position in Win 10 Sound Control panel - Recording tab / Level Slider

I typically run the RECEIVE level 9on FT8) as follows between 50-60 dB
1.) Menu #72 at 1 or 2
2)  RF Gain fully CW with AGC OFF
3)  Win 10 Sound Control/Recording/Level Slider at 0dB or close to it.

The User Guide Section 5 Transceiver Setup says "Use the receiver gain controls and/or the computer’s audio mixer controls to set the background noise level  to around 30 dB when no signals are present. It is usually best to turn AGC off or reduce the RF gain control to minimize AGC action."

In order to set the level at 30db in my case the Menu# 72 setting must be at 1 (the minimum setting) and the Sound Control slider must be at -19dB.

1.- Is there a common practice out there regarding the level control? 
2.- Does the magnitude of the RECEIVE level control impact on the DECODE reliability of the software.? 
3. - Is the decode reliability negatively impacted by setting the level control to 50-60dB?
4 -  Is the setting for Receive level specific to mode  i.e. 30dB  for Q65  ;  50dB for FT8 etc.

73 Frank VO1HP




Martin G0HDB
 

On Wed, May 12, 2021 at 07:04 AM, Bill Somerville wrote:

Martin,

it should be no surprise that with no AGC strong signals will overload your receiver, it is your responsibility to adjust the RF GAIN to compensate. The advantage of using the RF GAIN instead of the rig's AGC is that you can consider the weakest signals whereas the AGC is only set by the strongest signal.

73
Bill
G4WJS.

Hi Bill, I would hardly call a signal at -85dBm, which equates roughly to S3-4, a particularly strong signal yet this is the signal level at which distortion begins to appear on the audio output from my IC-7610 when I operate the Rx with AGC off.  My previous IC-7600 showed almost exactly the same characteristic; however Rob NC0B has found that his IC-7300 doesn't begin to exhibit distortion on the audio output until the RF input signal is at a *very* much higher level.  I can't recall the exact figure that Rob has mentioned but I believe he found it needed an input signal of around -30dBm, which equates to approx S9+40dB, to cause the onset of audio distortion with AGC off.  Presumably the presence of distortion artefacts on the audio going into WSJT-X won't be particularly conducive to effective decoding.

As I expect you're aware, Rob NC0B did some tests a couple of years ago to investigate how WSJT-X's decoding of off-air WSPR signals on 630m was affected by the presence of a strong, locally-injected in-passband unwanted signal when he operated his IC-7300 with AGC on.  Rob found that the unwanted signal had to be at a level of approx -15dBm (S9+++!) before the decoding of the wanted WSPR signals, which were decoded at levels of -10 and -24dB, began to be affected.  From his measurements Rob surmised that even with AGC on, WSJT-X was capable of decoding wanted signals that are 75-80dB below the level of a high-level signal within the same passband.  I've briefly performed similar AGC-on tests to Rob, using off-air FT8 signals on 160m and injecting an interfering signal from a (calibrated) signal generator via a hybrid combiner, and found similar results, ie. the level of the in-passband interfering signal had to be quite a few 10s of dB (eg. 60-70dB) higher than the levels of the wanted signals in the passband before the decoding of the wanted signals began to be degraded.  I suspect that many people who operate their Rx's with AGC off as per the guidance might not be aware that even medium-level input signals could be overloading something along their Rx chain with the consequence that the audio going from their Rx into WSJT-X could be horribly distorted.

I would venture to suggest that, given the results of Rob's tests which I've partially verified by my own tests, there's a high degree of certainty that operating an Rx with AGC on will not result in any significant loss of the ability of WSJT-X to decode wanted signals that are several 10s of dB below the levels of the strongest signals within the passband even though those strong signals will have 'captured' the AGC.  Furthermore, operating the Rx with AGC on will undoubtedly eliminate the possibility of the audio output from an Rx being driven into audio distortion by even medium-strength input signals - although the occurrence of this may vary between different makes and models of Rx (vis the significant difference between my IC-7600/7610 and Rob's IC-7300), any tendency for audio distortion artefacts to begin to appear on the audio output can't be good for WSJT-X's decoding.

Finally, I believe that when Rob NC0B operates his Rx with AGC on he sets the audio levels to be at around the 68dB mark on the WSJT-X 'thermometer', which isn't too dissimilar to the setting of 60dB that I use.  I also understand that Rob's approach was validated by Joe K1JT.

73
--
Martin G0HDB


Jim Brown
 

On 5/12/2021 5:43 AM, Martin G0HDB wrote:
Finally, I believe that when Rob NC0B operates his Rx with AGC on he sets the audio levels to be at around the 68dB mark on the WSJT-X 'thermometer', which isn't too dissimilar to the setting of 60dB that I use.
It's important to realize that distortion can be produced at multiple points in the audio chain between detector and A/D converter, and, as suggested by a post in this series, even in the IF due to poorly implemented AGC.

My rig is an Elecraft K3, which has AGC that is pretty well behaved, so I use slow AGC. As a retired audio professional, I set RX audio level driving WSJT-X so that with the strongest signals present (that is, during decode), the green bar is as close as possible to the top without flashing red. I use an external sound card that has a front panel gain adjustment, so I can "ride gain" on it when loud locals fire up on 6M or 160M, which they often do. If I did not have that adjustment, I would "ride gain" on the RF gain.

It's also important to realize that there are multiple gain stages between the detector and the A/D, and if any one of them is poorly adjusted, distortion can be produced, even if the green bar is fairly low. A good clue that this is happening is to look for harmonics of a very loud signal fairly low in the AF passband (for example, 550 HZ, with harmonics at 1100, 1650, 2200 Hz), or for intermod distortion (multiple thin vertical lines in the waterfall for very strong signals anywhere in the audio passband).

One way this can occur is if Record Gain in Windoze is set too low, so that the input to the A/D is overdriven and hits analog clip.

73, Jim K9YC