Flex / WSJT-X Users... Signal Reports?

  • 1
  • Question
  • Updated 2 years ago
For those using WSJT-X what is your experience with signal reports?

I find that my received signal report is very often (> 70%) lower than the report I'm sending. The differences are around 5 dB but often enough as much as 15 dB difference. My expectations would have been generally similar sent/received signal reports with the occasional outliers. 

I've been trying to decide what factors may be what might be the most likely cause. My primary concern is not so much a weak signal (though I'd like to understand why) but a somehow distorted or bad signal. It would help to see if my experience is similar to others.

 The Obvious:
  • Lots of poor receivers or noisy environments
  • Lots of people running JT at high power
For reference, here is my last 10 JT(65/9) reports (Sent, Received). I was running 40 watts during this run:

-06, -11  delta -05
-01, -15  delta -14
-01, -07  delta -6
-13, -09  delta +4
-09, -28  delta -19
-17, -12  delta +5
-24, -11  delta +13
-11, -23  delta -12
-11, -15  delta -4
-03, -10 delta --7

Kev K4VD
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 781 Posts
  • 172 Reply Likes

Posted 2 years ago

  • 1
Photo of Rob Fissel

Rob Fissel

  • 270 Posts
  • 48 Reply Likes
Having your report be lower in dB than the other sender is a good thing in my opinion, considering this is a weak signal (NOT QRP) mode. However, there are so many variables that it's hard to tell. 

If the majority of your reports (as you indicate) are lower, then keep doing what you're doing. While not a QRP mode, I do enjoy treating it as such, and will always follow the amateur mantra of "lowest amount of power required to complete the contact."

Typically, if I get a signal report of -10 or higher, I'll back off the power. I use the quick math method which is very very unscientific in my head (in conjunction with FlexMeter that gives me my tx power in dBm)...

Most radios are capable of decoding down to -25 pretty reliability. Say someone comes back to me with a signal report of -5. That gives me about 20 db of S/N room to play with, so I'll back down my power output by at least half and see where that gets me. 

I wouldn't worry to much about the delta between the other guy and you, because, well, to each his own. I know many people that routinely call CQ with 20-30 watts (way too much in my opinion), and will consistently receive strong signal reports. To me, that means back way down on the power output. To others, the louder their signal is, the happier they are. 

Distorted or bad signals simply wouldn't be decoded. If I had to take an educated guess, you're receiving weak signal reports because you're a more considerate operator. There are plenty out there, including myself, that will send a friendly email to amateurs when I notice issues with their distorted or noisy signal in an attempt to resolve. 

Keep in mind the other variables at play here. Type of antenna in use, efficiency of their feedline system and antenna system, their radiation angle vs. yours (while taking distance between stations into account), their potential noise floor, etc. Of course, their S/N is relative to their individual noise floor. What could be a report of -5 for me with a relatively low noise floor could be a report of -25 to someone else with an S9+10 noise floor. 

As is such in life, everything is relative. 
Photo of Martin Ewing AA6E

Martin Ewing AA6E

  • 294 Posts
  • 58 Reply Likes
I've noticed the same thing. My take is that this just says that you are receiving with a better signal to noise ratio (SNR) than the other guy. That could mean that you have a better receiver, that your receive antenna is less noisy (local QRN etc), that you're getting less QRM in your receiver (better filtering, less IMD etc), or of course your tx power or tx antenna efficiency could be worse.

So it's fairly likely that the difference is that you have a better rx setup -- a Flex! -- than average.

73 Martin AA6E
Photo of Walt


  • 236 Posts
  • 74 Reply Likes
I saw differences in what I saw and what the other stations reported and wondered what was going on, enough so that I ran my antenna through a splitter and fed both my TS-590 and the 6500 and watched dual WSJT programs process the same signals. 

Found the Flex and TS590 Rx values were the same most of the time.  Sometimes they would alternate as to who was 1 db better.  But for the most part, identical.  Weakest I could decode was -26, but mostly -25 was the bottom limit.

So I know its not the radios - I do have higher noise here than probably most stations so I expect to see better signals at the other end than what I can copy.

I find it a fun mode.  I thought the Flex would out-shine the 590, but not in this test, which I ran steady for 4 days on 20 meters.  Maybe if I had a low-noise site.  In the Burbs and so noise is all around me, unless the power goes out.  Nice to have a good battery bank for some brief, no-noise operating.

Photo of Jay / NO5J

Jay / NO5J

  • 1618 Posts
  • 233 Reply Likes

The Color gradient I'm using is a few dB hotter than SmartSDR. Which results in the weaker signals being a little more visible, but has the side effect that the transition to red occurs at a lower signal level too. But you can compensate for the difference with the gain controls in Fldigi and the Wide Graph. But it's never perfect.

73, Jay - NO5J

Photo of Jay / NO5J

Jay / NO5J

  • 1618 Posts
  • 233 Reply Likes
I just noticed that when the Capture was done. I had AGC=Off, AGC-T=27 set for the Fldigi slice. For the WSJT-X slice I had AGC=Off, AGC-T=10. What a goof! Normally I try to use AGC=Fast, AGC-T=27 for both slices. Sometimes I turn on Auto for the SmartSDR waterfall palettes, Sometimes I turn it off. It's never a set it once, always right, and no need for any further tweaking thing.

73, Jay - NO5J

Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9153 Posts
  • 3492 Reply Likes
I normally run all digital modes with AGC=off unless I am operating a contest.  Since there is a direct relationship between the source audio level (volume) and the setting of the AGC-T for a maximum SNR, I do not want an in-band strong signal dynamically adjusting the volume and degrading the SNR.
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 781 Posts
  • 172 Reply Likes
I'm not so sure.  If adjusted properly:

1) Signals above the noise floor are amplified while the noise is not (i.e. improved SNR).
2) The strongest signals may be attenuated while the weaker signals are not.
3) A strong signal nearby will not have an adverse affect on a weak signal nor on overall SNR.

In other words, I understood a properly adjusted AGC actually improves SNR for weaker signals while stronger signals are so far above the noise that it doesn't make a difference regardless of mode.

This was demonstrated nicely in a video by Steve Ellington, Flex-Radio 6000 Series AGC-T Adjustment Procedure. It's a procedure I follow   in any mode and seems to work for me. It even makes sense to me (and that's always a challenge).
Photo of Jay / NO5J

Jay / NO5J

  • 1618 Posts
  • 233 Reply Likes
I think both AGC=Off and AGC=Fast, have their uses. With both, I set AGC-T low, to get 30dB during the quiet periods on the WSJT-X level meter. When the noise floor is bouncing I use AGC=Fast, in hope that the AGC action is helping by auto riding the amount of gain applied to the signals above the noise floor. I try to set the threshold so that as much of the the noise floor as possible is eliminated from the DAX stream fed to WSJT-X. Raising the threshold a dB or two might net me more decodes, but also adds probably a few  dB of additional noise for the decoder to have to deal with. I want the decoder to see the cleanest signals possible. 

On the other hand, if there's someone either close by, or overdriven active, then AGC=Off will eliminate any AGC-pumping, that drives the weaker signals out of the DAX Stream.

I try to let the signals in the passband be the determining factor in how I set the AGC.
There is an additional @30 dB of headroom above the 60 dB meter readings on the WSJT-X level meter. It's not unusual for me to see the WSJT-X level meter pegged at 60 dB, when signals are present and yet still dropping to 30 dB during the quiet periods.

Signals the decode at a -1 dB level will still get a -1 dB report even if they are 30 dB over -1 dB. Per the WSJT-X documentation JT65 signal reports range -30 dB to -1 dB. Although I rarely see < -26 dB reports. I do see more -1 dB reports than I believe.

We should only need 30 dB of headroom to receive anything decodable.

The trick is deciding how much noise you are willing to let WSJT-X, try and then fail to decode.

I try to avoid working the noise, signals net me more QSO's


73, Jay - NO5J

Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 781 Posts
  • 172 Reply Likes
I just worked up the averages of the 1500+ JT contacts in my log:

average sent: -7.72
average received: -10.33
average difference: -2.60

Walt's comparison of the Flex and TS590 could account for 1 dB of that number and then the other 1.6 dB could just be noise in the data. I have to admit I was expecting the difference to be a lot higher just based on eyeballing the data over time. 

The largest differences I saw in the data were a JT-9 conversation with AI5II where I sent a -19 and received a +10. On the other end of the scale was a JT9 Q with K9TJN where I sent a +6 and received a -29.

With only a -2.60 difference on average this didn't turn out to be quite as exciting as I thought it would be.
Photo of KM4CQG


  • 229 Posts
  • 37 Reply Likes
Check the box for deep decoding perhaps?

Photo of George Molnar, KF2T

George Molnar, KF2T, Elmer

  • 1602 Posts
  • 567 Reply Likes
Echoing others, I almost always send a better report than I get. At least on HF. That serves as a reasonableness test for my power setting. Given the robust nature of JT modes, I wouldn't expect too much difference between modern radios, all things being equal. The Flex excels with a wonderful clean signal and flat passband for both transmit and receive, helping me be a "good neighbor" on the bands.
Photo of Andrew Thall

Andrew Thall

  • 151 Posts
  • 16 Reply Likes
I can decode signals easily with my Flex 6300 that I couldn't even see on my Yaesu FTdx3000.
Photo of EA4GLI - 8P9EH - Salvador

EA4GLI - 8P9EH - Salvador

  • 1768 Posts
  • 541 Reply Likes
A couple of comments, make sure flatten is NOT selected on wsjt-x, and 20 to 50w is alot in those modes. 5 or 10w should be your max power output. I find more interesting to see how far I can get with the lowest power possible.

Try the Transverter port one day you will be amazed!