locked #config IC-7610 AGC setting #WSJTX_config


Tad Danley K3TD
 

What is the recommended AGC setting on the 7610 for the WSJT modes?
Thanks and 73,
Tad, K3TD
 


Alan <n5pa@...>
 

Tad:

 

On the IC-7610 you have the AGC set to OFF.  At least that is what I have set mine up at and a couple of other.

 

73,
Alan Clark, N5PA
Ellisville, MS
Email:  n5pa@...
URL:  http://www.n5pa.com

 

From: main@WSJTX.groups.io [mailto:main@WSJTX.groups.io] On Behalf Of Tad Danley
Sent: Friday, July 09, 2021 3:30 PM
To: main@WSJTX.groups.io
Subject: [WSJTX] #config IC-7610 AGC setting

 

What is the recommended AGC setting on the 7610 for the WSJT modes?
Thanks and 73,
Tad, K3TD
 


Virus-free. www.avg.com


Larry Menzel <pensionguy@...>
 

No AGC at all.  Pull the slider down on the right side of the WSJT screen while transmitting and take the AGC to zero.



On Fri, Jul 9, 2021 at 3:41 PM Tad Danley <tdanley@...> wrote:
What is the recommended AGC setting on the 7610 for the WSJT modes?
Thanks and 73,
Tad, K3TD
 



Bill Somerville
 

On 09/07/2021 22:33, Larry Menzel wrote:
No AGC at all.  Pull the slider down on the right side of the WSJT screen while transmitting and take the AGC to zero.



On Fri, Jul 9, 2021 at 3:41 PM Tad Danley <tdanley@...> wrote:
What is the recommended AGC setting on the 7610 for the WSJT modes?
Thanks and 73,
Tad, K3TD

Larry,

you are confusing AGC and ALC.

73
Bill
G4WJS.


Michael Black
 

AGC is a receive setting,.,..not a transmit setting.  ALC is the transmit setting.

For AGC it should be OFF for most operators unless you have a  neighbor swamping your signal.
AGC will reduce the signal level and can cause loss of low-level signals so OFF is the preferred mode.

For ALC my notes on a 7610 shows this: IC7610 MOD INPUT LEVEL 11% with 50W gives 42W
Which should provide approx 84W without tickling the ALC and make for a nice clean signal.

Mike W9MDB





On Friday, July 9, 2021, 04:38:13 PM CDT, Larry Menzel <pensionguy@...> wrote:


No AGC at all.  Pull the slider down on the right side of the WSJT screen while transmitting and take the AGC to zero.



On Fri, Jul 9, 2021 at 3:41 PM Tad Danley <tdanley@...> wrote:
What is the recommended AGC setting on the 7610 for the WSJT modes?
Thanks and 73,
Tad, K3TD
 






Larry Menzel <pensionguy@...>
 

HAHA...my bad...my eyesight never was too good.  You're correct...sorry.

ready fire aim

Larry


On Fri, Jul 9, 2021 at 4:56 PM Michael Black via groups.io <mdblack98=yahoo.com@groups.io> wrote:
AGC is a receive setting,.,..not a transmit setting.  ALC is the transmit setting.

For AGC it should be OFF for most operators unless you have a  neighbor swamping your signal.
AGC will reduce the signal level and can cause loss of low-level signals so OFF is the preferred mode.

For ALC my notes on a 7610 shows this: IC7610 MOD INPUT LEVEL 11% with 50W gives 42W
Which should provide approx 84W without tickling the ALC and make for a nice clean signal.

Mike W9MDB





On Friday, July 9, 2021, 04:38:13 PM CDT, Larry Menzel <pensionguy@...> wrote:


No AGC at all.  Pull the slider down on the right side of the WSJT screen while transmitting and take the AGC to zero.



On Fri, Jul 9, 2021 at 3:41 PM Tad Danley <tdanley@...> wrote:
What is the recommended AGC setting on the 7610 for the WSJT modes?
Thanks and 73,
Tad, K3TD
 









Jim Brown
 

On 7/9/2021 2:55 PM, Michael Black via groups.io wrote:
For AGC it should be OFF for most operators unless you have a  neighbor swamping your signal.
AGC will reduce the signal level and can cause loss of low-level signals so OFF is the preferred mode.
I use FT8 and similar modes mostly on 6M and 160M, and on both bands, nearly always have local signals driving the AGC, so I DO use Slow AGC in my K3. I also carefully ride gain on the input to my audio interface (Tascam US100) so that both the clip light on the Tascam stays green and I get as close as possible to the top of the green bar without going red. This maximizes the dynamic range of the A/D, which in turn allows WSJT-X to dig closer to the receiver's noise floor to decode weaker signals.

73, Jim K9YC


Martin G0HDB
 

On Fri, Jul 9, 2021 at 10:56 PM, Michael Black wrote:
AGC is a receive setting,.,..not a transmit setting.  ALC is the transmit setting.
 
For AGC it should be OFF for most operators unless you have a  neighbor swamping your signal.
AGC will reduce the signal level and can cause loss of low-level signals so OFF is the preferred mode.
 
You don't need a very strong neighbourhood signal to cause problems with the 7610's AGC off even when the pre-amps are off - the audio output from the rig, either from the audio codec via USB or from the analogue ACC1 socket (or the speaker/phones outputs), will begin to distort when the RF input signal reaches approx -85dBm which equates to around S3-4.  Once you start getting severe distortion on the audio going into the WSJT-X decoders you're likely to suffer all sorts of decoding issues.

I'd strongly recommend operating with AGC on, with a fast time constant (I use 0.8secs) and setting the audio levels so that the WSJT-X audio level 'thermometer' reads ~60dB when receiving only band noise.  This is broadly the same solution as advocated by Rob Sherwood NC0B; Rob performed some AGC-on tests when receiving WSPR signals on 630m and demonstrated that the WSJT-X decoders would cope without degradation to the decoding of the wanted WSPR signals when an in-passband interfering signal that was 75-80dB stronger than the off-air WSPR signals was present. 

Having AGC on also should completely eliminate the possibility of the 7610's front-end ADC being driven into clipping, as indicated by the 'OVF' annunciator coming on, by strong signals (local or otherwise).

--
Martin G0HDB


Michael Black
 

Why 60db instead of the recommended 30dB on the level?
That's limiting your dynamic range.

Mike W9MDB




On Saturday, July 10, 2021, 09:55:39 AM CDT, Martin G0HDB <marting0hdb@...> wrote:


On Fri, Jul 9, 2021 at 10:56 PM, Michael Black wrote:

AGC is a receive setting,.,..not a transmit setting.  ALC is the transmit setting.
 
For AGC it should be OFF for most operators unless you have a  neighbor swamping your signal.
AGC will reduce the signal level and can cause loss of low-level signals so OFF is the preferred mode.
 
You don't need a very strong neighbourhood signal to cause problems with the 7610's AGC off even when the pre-amps are off - the audio output from the rig, either from the audio codec via USB or from the analogue ACC1 socket (or the speaker/phones outputs), will begin to distort when the RF input signal reaches approx -85dBm which equates to around S3-4.  Once you start getting severe distortion on the audio going into the WSJT-X decoders you're likely to suffer all sorts of decoding issues.

I'd strongly recommend operating with AGC on, with a fast time constant (I use 0.8secs) and setting the audio levels so that the WSJT-X audio level 'thermometer' reads ~60dB when receiving only band noise.  This is broadly the same solution as advocated by Rob Sherwood NC0B; Rob performed some AGC-on tests when receiving WSPR signals on 630m and demonstrated that the WSJT-X decoders would cope without degradation to the decoding of the wanted WSPR signals when an in-passband interfering signal that was 75-80dB stronger than the off-air WSPR signals was present. 

Having AGC on also should completely eliminate the possibility of the 7610's front-end ADC being driven into clipping, as indicated by the 'OVF' annunciator coming on, by strong signals (local or otherwise).

--
Martin G0HDB




Martin G0HDB
 
Edited

On Sat, Jul 10, 2021 at 04:19 PM, Michael Black wrote:
Why 60db instead of the recommended 30dB on the level?
That's limiting your dynamic range.
 
The rig's AGC ensures that the signals going into WSJT-X never result in a reading of greater than approx 74dB on the WSJT-X audio level 'thermometer', and setting the audio level to give a 60dB reading on band noise means that there's ample headroom above the 30dB mark so that even weak signals will be above the threshold resulting from ignoring the bottom few (5 or 6?) bits coming out of the soundcard digitisation.

FWIW, I've regularly seen decodes where there's a >50dB difference between the weakest and strongest (FT8) signals within the same 15sec timeslice.  For example, I've seen several decodes of -24dB when there's also been one or occasionally two decodes at around +30dB in the same passband and timeslice.  I'd much rather accept a slight degree of compression of the dynamic range than suffer potentially catastrophic decoding failures resulting from WSJT-X being fed with grossly-distorted audio signals for much of the time.

As Rob Sherwood NC0B will be able to confirm, the IC-7300 doesn't suffer from the same audio output overload problem as the 7610 (and its 7600 predecessor); IIRC Rob's tests indicated that the 7300 would accept an enormous input signal (of around -30dBm = approx S9+40dB) before its audio output began to exhibit distortion products.  On this basis the 7300 would be a much more suitable rig than the 7610 for operating WSJT-X with AGC off and a 30dB band-noise level setting.

--
Martin G0HDB


Jim Brown
 

On 7/10/2021 8:19 AM, Michael Black via groups.io wrote:
Why 60db instead of the recommended 30dB on the level?
That's limiting your dynamic range.
Exactly the opposite, Mike. For max dynamic range, gains must be set so that with the strongest signals present, the level indicator (green bar) is as close as possible to 80 dB and just below digital clip.

73, Jim K9YC


Michael Black
 

Where do you get that from Jim?

The WSJT-X team did measurements that clearly showed 30dB on a quiet spot was all that was needed to overcome quantization errors.

The dB meter in WSJT-X does not measure peak envelope and cannot be relied on to show maximum signal -- it shows RMS level which is more than 10dB below max signal level typically.

Mike W9MDB


On Saturday, July 10, 2021, 01:34:20 PM CDT, Jim Brown <k9yc@...> wrote:


On 7/10/2021 8:19 AM, Michael Black via groups.io wrote:

> Why 60db instead of the recommended 30dB on the level?
> That's limiting your dynamic range.


Exactly the opposite, Mike. For max dynamic range, gains must be set so
that with the strongest signals present, the level indicator (green bar)
is as close as possible to 80 dB and just below digital clip.

73, Jim K9YC






Jim Brown
 

On 7/10/2021 2:03 PM, Michael Black via groups.io wrote:
Where do you get that from Jim?
The WSJT-X team did measurements that clearly showed 30dB on a quiet spot was all that was needed to overcome quantization errors.
The dB meter in WSJT-X does not measure peak envelope and cannot be relied on to show maximum signal -- it shows RMS level which is more than 10dB below max signal level typically.
From my professional life in pro audio. And I find that it correlates pretty well with whether or not there are distortion products on the waterfall from strong signals. And the signal LED on the Tascam US100 interface that I use DOES read peaks, changing from Green to Red at or near digital clip.

73, Jim K9YC


Michael Black
 

If you have something that proves 30dB on a quiet spot is not the best setting would be nice to see it.  That's what the WSJT-X manual recommends and was backed up by testing.

Mike W9MDB





On Saturday, July 10, 2021, 05:29:05 PM CDT, Jim Brown <k9yc@...> wrote:


On 7/10/2021 2:03 PM, Michael Black via groups.io wrote:
> Where do you get that from Jim?
>
> The WSJT-X team did measurements that clearly showed 30dB on a quiet
> spot was all that was needed to overcome quantization errors.
>
> The dB meter in WSJT-X does not measure peak envelope and cannot be
> relied on to show maximum signal -- it shows RMS level which is more
> than 10dB below max signal level typically.

From my professional life in pro audio. And I find that it correlates
pretty well with whether or not there are distortion products on the
waterfall from strong signals. And the signal LED on the Tascam US100
interface that I use DOES read peaks, changing from Green to Red at or
near digital clip.


73, Jim K9YC




Jim Brown
 

On 7/10/2021 3:34 PM, Michael Black via groups.io wrote:
If you have something that proves 30dB on a quiet spot is not the best setting would be nice to see it. That's what the WSJT-X manual recommends and was backed up by testing.
I know what the manual says -- I read it carefully many years ago, and consult it for new stuff and to refresh my memory.

Consider a station that's 50 dB over S9 on a calibrated S-meter (in my case, a K3 with a GAs-FET preamp) and, at 6 dB/S-unit, how weak a signal you could decode on a 16-bit system. On 6M, my noise is S3 in noisy directions and S1 in quiet directions, and in noisy directions, at quiet times).

73, Jim K9YC


Michael Black
 

I suppose if you have neighbors causing that kind of front end loading you have no choice.
I've never seen such a thing here or on any of the hundreds of people I've helped.

That would seem the time for AGC though.

Mike W9MDB




On Saturday, July 10, 2021, 08:34:51 PM CDT, Jim Brown <k9yc@...> wrote:


On 7/10/2021 3:34 PM, Michael Black via groups.io wrote:
> If you have something that proves 30dB on a quiet spot is not the best
> setting would be nice to see it. That's what the WSJT-X manual
> recommends and was backed up by testing.

I know what the manual says -- I read it carefully many years ago, and
consult it for new stuff and to refresh my memory.

Consider a station that's 50 dB over S9 on a calibrated S-meter (in my
case, a K3 with a GAs-FET preamp) and, at 6 dB/S-unit, how weak a signal
you could decode on a 16-bit system. On 6M, my noise is S3 in noisy
directions and S1 in quiet directions, and in noisy directions, at quiet
times).


73, Jim K9YC





Jim Brown
 

On 7/10/2021 8:20 PM, Michael Black via groups.io wrote:
I suppose if you have neighbors causing that kind of front end loading you have no choice.
I've never seen such a thing here or on any of the hundreds of people I've helped.
That would seem the time for AGC though.
Yep, which is why I settled on it a few years ago. Like I said, my uses of FT8 are weak signal work on 160 and 6M. When you're trying to work EU from NorCal on 160, you're running legal limit, and so are the other guys out here trying to do the same thing. And it DOES require careful gain-riding of the A/D unit driving the computer.

Some ops choose to deal with this sort of issue by narrowing the IF, but as soon as Joe answered my question about doing that by talking about the phase shift produced by the filters degrading decoding, I immediately understood. In the world of pro sound reinforcement, we learned 40 years ago that phase shift also degrades the ability of the human ear/brain to understand speech.

73, Jim K9YC


Martin G0HDB
 

On Sat, Jul 10, 2021 at 11:34 PM, Michael Black wrote:
If you have something that proves 30dB on a quiet spot is not the best setting would be nice to see it.  That's what the WSJT-X manual recommends and was backed up by testing.
 
Mike:

Setting the audio level to give a 30dB reading on band noise is fine if you can be absolutely certain that, with AGC off, the audio output level from the Rx will follow the RF input level in a linear manner over approaching a 60db range, ie. to somewhere near the 90dB mark on the WSJT-X 'thermometer'.  However, this demonstrably does *not* happen with some rigs, eg. my previous IC-7600 and my present IC-7610.  With these, the audio output starts to become distorted when the RF input level reaches around -85dBm which is only about 40dB above the receiver's noise floor and possibly only 20-30dB above the level of band noise even on 6m.

As already mentioned, the tests performed by Rob Sherwood NC0B on his IC-7300 demonstrated that it does not exhibit the same characteristics as his IC-7610 on its audio output until a very high RF input level is reached; there must be some differences in the (analogue and/or DSP) designs of the 7300 and 7610 that give rise to the different behaviours.  I have no knowledge of how other makes of rig perform in this respect - it's not a parameter that I can recall ever seeing mentioned in test reviews - but I wouldn't be at all surprised if some of them also can't handle a 60dB range of input signal without going into distortion on the audio output when AGC is off.  I suspect the presence of significant distortion products on the audio signal going into the WSJT-X decoders will have a rather deleterious effect on the decoders' performance...

I know it's possible ride a rig's input attenuator and/or the RF gain control to ensure that the audio output never becomes distorted but how many operators have the desire, understanding, and skills, to do that all the time, when the received signal strengths might be widely varying over short periods of time?

I would assert that unless someone is *absolutely confident* that their receiver doesn't begin to exhibit distortion on its audio output at high signal levels then they'll obtain better overall decoding performance if they operate with AGC on and set the audio levels so that band noise gives a reading around the 60-65dB mark.

The only scenario in which I operate my 7610 with AGC off and a WSJT-X reading of 30dB on band noise is when I'm using MSK144 - I can be fairly confident that the received signal levels are unlikely to cause distortion on the 7610's audio output...  :-)

--
Martin G0HDB