AGC-T Offline Help?

  • 1
  • Question
  • Updated 1 year ago
  • (Edited)
Is there someone on the list that wouldn't mind a couple (or more) direct emails with questions about AGC-T particularly in CW mode? It's an old topic that comes up again and again and I don't really want to be the one to start it all over. I'm looking to see if someone might have an insider understanding or at least some intuition on how AGC-T should work - beyond what's already been printed.

My email address is kkinderen@gmail.com.

73,
Kev K4VD
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 775 Posts
  • 171 Reply Likes

Posted 1 year ago

  • 1
Photo of HCampbell  WB4IVF

HCampbell WB4IVF

  • 199 Posts
  • 45 Reply Likes

I know you’re looking for info beyond what’s already been posted, but in case you haven’t run across it, Steve Hicks wrote a white paper several years back that included a discussion of the AGC-T algorithm (under “Gain Control Considerations”):

https://www.google.com/search?q=DIRECT+CONVERSION+RECEIVERS+hicks&ie=utf-8&oe=utf-8

and followed up:

https://community.flexradio.com/flexradio/topics/questions-about-agc-t-algorithm

BTW, I hope you share your offline questions and answers.  AGC/AGC-T comes up frequently, but it’s an important concept and I know that I, and probably others, would benefit!


Howard
(Edited)
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 775 Posts
  • 171 Reply Likes
Hi Howard. I'm in a few discussions over the issue. Thanks for reminding me about Steve's papers. I need to go back and review them.

I've spent a lot of time the past few months on digital and had my AGC turned off. But I wanted to play around on CW for a bit and of course AGC was flipped to Slow. Thing is, I couldn't get the results I'm used to. The AGC-T control felt like a variable RF Gain control from end to end. I'm not getting that "signal popping out of the noise" I usually get when I hit the right spot on the curve. I'm kind of wondering if there's some other control I forgot to set.

Anyway... pretty sure it is just me. I'll keep playing and if I find anything useful I'll post it up.

73,
Kev K4VD
Photo of HCampbell  WB4IVF

HCampbell WB4IVF

  • 199 Posts
  • 45 Reply Likes

Thanks Kevin.  CW is my favorite mode and I’d be interested in any other info on AGC-T  related settings that you pick up.  My own experience probably isn’t typical (tinnitus being an especially miserable handicap for a CW op!).   (-:

Howard

Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4025 Posts
  • 1248 Reply Likes
You may perhaps be expecting something a little more dramatic than what actually happens, but the effect of the properly adjusted AGC-T is noticeable and helpful.  But it isn't "earth-shattering" and "dramatic."

I never run CW with the AGC in SLOW.  Hardly ever in Medium.  On bands with lots of fading, a SLOW AGC will not follow the fading closely enough and you will lose characters in the fade.

On CW, I use FAST AGC, and a relatively narrow filter - at least 400 Hz, preferably 250 Hz or less.  In a pileup or weak signal situation, I love the 50 Hz filter!

I have the AGC-T linked to the AUX2 button on my FlexControl Knob so I can make fine and fast adjustments when needed.  It is much more accurate than the mouse.

I begin with the AGC-T far enough to the right that the noise is at maximum.  Then I slowly bring it to the left (smaller numbers) until the background noise just begins to drop. - perhaps a little more on a noisy band or if I am not looking for weak signals and want a quieter background.  This is the "Knee" or "Sweet Spot."  if you hit the right spot on the knee, then even weak signals will rise above the background noise more noticeably than if the AGC-T was much higher.  You will also have less noise rattling around in the filter.  While the CW filters on the 6000 series do not "ring,"  blasting it with unnecessary white noise, especially on an extremely narrow filter, will generate a background noise profile that almost sounds like ringing.  It is annoying at best, and unnecessary, because with a properly set AGC-T you will not miss any weak signals, only the noise that adds to operator fatigue.

As you play with the control, you will notice a spot where the noise drops off like the side of a cliff.  By the way, this sweet spot will be at a different place in each band, which is why the band persistence links the AGC settings to the band.

Now, here is the catch that some miss..... If you activate WNB, NB, NR, APF, or adjust your filter width or RX EQ, you may need to tweak the AGC-T setting just a little to find the best sweet spot again.  All of these controls interact a little.  For example, selecting a more narrow filter will reduce your noise level, readjusting AGC-T for a new sweet spot will make sure that you are still able to hear the weakest signals, but limiting the noise passing through the filter.

BTW.  I always use a relatively low sidetone - 500 Hz works best for me - and augment the receiver audio profile by turning on the RX EQ with all sliders to the bottom except for the 500 Hz slider at the top.  I also often use APF (Audio Peak Filter) at about 30 (no more, or it WILL get a bit pingy)  and NR at about 10-15 on CW, especially on noisy bands.  These settings, with a 500 Hz filter and carefully adjusted AGC-T have helped me copy many stations that I would have easily missed without their help.

I hope this helps.  These are MY preferences.  Other ops will report a different experience and set of preferences.  But that is the beauty of the FRS SDR - Flex-ability!

Have fun and have good luck.

Ken - NM9P
(Edited)
Photo of John - AF3K

John - AF3K

  • 120 Posts
  • 31 Reply Likes
Howard,  Thank you for posting the paper and the discussion link.  Hadn't seen that before. John AF3K
Photo of Al / NN4ZZ

Al / NN4ZZ

  • 1735 Posts
  • 590 Reply Likes
As Ken noted adjusting the AGC-T setting is helpful as you change antennas, bands, or when conditions change.  It is not usually dramatic but does help on the weak signals.  The biggest problem for me is remembering to tweak it since it doesn't need adjusting most of the time.  I really think that automating the setting would work for me 99% of the time.  Maybe for you too.  

Hopefully this idea will be implemented one day and we'll have it as an option.  It is number 2 on idea list by user votes currently.  

https://community.flexradio.com/flexradio/topics/option-to-automate-the-agc-t-setting

Regards, Al / NN4ZZ  
al (at) nn4zz (dot) com
SSDR / DAX / CAT/ 6700 -  V 1.10.16
Win10
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 775 Posts
  • 171 Reply Likes
Thanks for the help. Ken, always appreciate your step by step guides. And Al, I had previously added my vote to the auto AGC-T option and still think it would be a good idea. What would be better is to auto set AGC-T but give an indication of the current setting. This would allow the user to manually take over from where the algorithm left off. It also might help me get the proper feel and expectation for AGC on this radio. So far, it is the biggest mystery to me.

I went back over Steve Ellington's video which I found useful in the past. I still don't feel AGC is working like it used to but I guess so much of what AGC does depends on band conditions.

73,
Kev K4VD
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 775 Posts
  • 171 Reply Likes
Here's an example of the action of AGC mode on a CW signal with fairly low band noise. The sample starts with AGC mode set to slow then switches to fast.

It seems to me that AGC slow is almost peaceful for slow to medium CW. There's not enough time between elements for the noise to come back up. When fast, the noise is almost as prominent as the CW signal. From peace to chaos. For high speed CW a faster mode is probably appropriate.

Still not certain I'm setting AGC-T properly but I'm thinking I'm in OK shape. I did not change the threshold between slow/fast mode change.



Here is the associated audio file: SlowFast.mp3

If anything, I would like to have an adjustable mode maybe. In the case of this sample, a little slower might have cleaned things up even more. Another thing, seeing the effect as in the graph above actually helped me hear the affect. Not unlike the panadapter helping me find find a weak signal in the noise. I sure wish there was more control over the AGC and some sort of graphical representation of the setting right on the panadapter.

73,
Kev K4VD
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 775 Posts
  • 171 Reply Likes
I'll be jumping from AGC questions to RF Gain questions in a moment but one more comment on AGC. I still don't fully understand how it works and I don't see a "knee" or sudden drop.I think there are two features that should be added as a pair or individually:

1) Visual cues (clues?) as to where the AGC-T setting should be placed. I believe the current AGC-T is arbitrary based on what a person thinks they are hearing or seeing on various meter arrangements. If not a visual cue right on the sacred (or fragile?) panadapter how about something like an indication on the S-meter or a dedicated meter for AGC.

2) Automatic Automatic Gain Control... if you can't give us a visual cue then give us an automatic setting. Work some magical mathematics and set an optimum AGC-T. Keep the Fast/Med/Slow/Off and add AUTO to that list. Make it so the AUTO setting can transfer to the fast/med/slow setting so we can tweak from there.

Enough from me on AGC. I know these suggestions were proposed over the yeeeaaarrrrssss and are probably around #2000 on a list of 10,000.

My next question in my series of dumb questions is about RF Gain. According to the manual, setting RF Gain involves checking the signal level without an antenna (dummy load ok?) and with an antenna. If the difference is less than 8 dB increase in noise on the antenna then additional RF Gain is appropriate. How much?
Dummy Load
  0 dB	-121.0 dBm
  
Antenna
-10 dB	-108.0 dBm  +13.0 above dummy load
  0 dB	-119.5 dBm  + 1.5 above dummy load
 10 dB	-123.0 dBm  - 0.2 below dummy load
 20 dB	-124.5 dBm  - 3.5 below dummy load
Here are my RF Gain my questions:

1) Going from dummy load to antenna (0 dB) only showed 1.5 dB increase so additional RF Gain is indicated. How much? I thought if I added 10 dB the difference would be +11.5 dB but instead it is -0.2 dB from dummy load.

2) The whole scale looks backwards to me. As I increase the RF Gain from -10 dB to +20 dB the noise drops from -108 to -124 dBm. How does adding 30 dB of gain (-10 to +20) translate to a 16.5 dB drop in the noise level? How does that drop in the signal strength as indicated by the meter translate to an increase in the noise I hear in the headphones?

Not really questioning if things are working right. I'm assuming they are. Just looking for an explanation to help me understand what I'm seeing. In the meantime I'll be pulling out the handbook and doing some web searching to see if I can't find an explanation. 

 I know as soon as I hit the submit button I'll see my mistake but here it goes...

73,
Kev K4VD
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9152 Posts
  • 3478 Reply Likes
Based on your numbers above, you need no RF gain applied.  As per the procedure, you look at the gain drop going from the antenna being connected to the value when it is disconnected.  At the 0, +10 and +20 values, the difference does not exceed 8 dB so your antenna is noise limited.  

You are not adding gain to the input signal.  The most important sentence in the article is:

Remember that when you engage the preamp, you have not changed the signal level at the antenna terminal. You have, however; improved the signal to noise ratio (SNR) with the preamp and therefore the noise will go down in reference to a fixed signal level. The previous statement holds true in all cases except one. That is if the radio is antenna noise limited.  
(Edited)
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 775 Posts
  • 171 Reply Likes
Am I misreading?
If the band noise measurement without a signal present in the passband is 8 dB or more than that with no antenna, additional gain is not needed.
My interpretation is that if I switch from no antenna to antenna and don't show at least 8 dB more signal level then gain is needed. Right?

In the case above, the band noise measurement without a signal present in the passband is 1.5 dB more than with no antenna. This is not 8 dB so additional gain is needed. 

If I went from -121 to -113 then no RF gain because it is 8 dB (or more).

Kev
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9152 Posts
  • 3478 Reply Likes
It may be easier to look at the difference between the antenna connected and disconnected.  Whether or not it is a decrease or an increase depends on if you start with the base line value or the final noise value to determine the difference.

The disconnected the antenna and the noise floor reading was -121.0 dBm.  This is your base line.

At the  0 dB RF preamp setting with the antenna connected, the S-meter value is -119.5 dBm.  This is an increase in the noise.  

Any time the noise floor rises with the antenna connected when you add RF gain, you have a worse SNR (more noise with the same signal level)

You added 10 dB of gain and with the antenna connected, the noise (signal) recorded was -123.0 dBm.  Compared to your baseline, this is a - 0.2 dBm decrease in the noise level.  Decreases are always preferred, but the difference is not 8 dBm or greater, so you are still noise limited.

You added 20 dB of gain and with the antenna connected, the noise recorded was -124.5 dBm.  Compared to your baseline, this is a - 3.5 dBm decrease in the noise level.  As before, decreases are always preferred, but the difference is still not 8 dBm or greater, so you are still noise limited at the antenna.

So, based on these numbers, you get no SNR reception enhancement by enabling the RF preamp.

What you will find is that unless you are located in a very RF quiet (very rural) location, you will probably not see any benefit to adding gain using the RF preamp until you get to the higher bands, like 12m and above. 

Or you are using unity or negative gain antennas (like beverages) that can benefit from RF preamplification in certain circumstances.

I suspect your antenna has gain based on the numbers above.
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 775 Posts
  • 171 Reply Likes
So if I got it right (and hope I do because it is starting to make sense), when I switch from no antenna to antenna and see ANY increase then it is safe to assume no RF Gain is really required. I'm already above the receiver's noise floor.

But if I see no gain then adding a little RF Gain may help.

My measurements were done on 20 meters. This indicates I don't need any RF Gain on 20.

I think I'm good with that understanding.

Why does the book use 8 dB? It seems to go back to a message you posted from Gerald four years ago.

Kev
(Edited)