NO AGC when AGC-T Properly Adjusted. See Video

Steve N4LQ

Flex-6500 et al. SmartSDR 1.4.16 or any version

I perceive this as a problem. Others may not. 

Adjusting AGC-T "properly" results in no AGC at all. 

Reason: AGC-T is really RF Gain and the AGC Threshold is fixed at such a high level that constant volume changes occur due to fading and various signal strengths. 
I demonstrate this in this video. 
https://www.youtube.com/watch?v=9z_c4S-LV0g 
We end up "riding" the AGC-T control to compensate. It's exactly like using a receiver from the 1930's.

I would like to see AGC-T to be a true AGC Threshold setting. We should be able to move it down - left and lower the threshold so that AGC worked on weaker signals. 

Threshold = Point where things start to work.
Low threshold should provide AGC at LOW signal levels. 
What we have is not a threshold at all. 

Turning AGC-T all the way up - to the right should result in NO AGC. Threshold raised! 
The current design is faulty. 

KC2QMA_John

Yes I noticed the same here. They just need to move the threshold point down and things should be fine. I would also like to see an advanced menu where you can adjust the AGC parameters and save as custom AGC presets.

George KF2T

What setting was the AGC-T? Slow, medium, fast, off? Sounds like off. The setting is not an RF gain control, although it sounds like it sometimes.

Steve N4LQ

I'm even sure that moving the threshold down will cure this. Not even an S9 signal triggers it now! 

Next we need to look at the jumpy S meter that is impossible to read. AGC has no effect on it either. 
My Drake 2B's meter works better!

George KF2T

Sorry, I noticed "Fast" a little late. It does seem to be working properly.

George KF2T

Love the 2-B!! But why should AGC affect an S-Meter that is supposed to read signal strength, not perceived level?

Steve N4LQ

George....AGC was set to FAST...I think that can be seen in the video. 
AGC-T is indeed RF Gain. It has nothing to do with AGC. I think I proved this. 

Steve N4LQ

Signal level is normally measured at the IF. 
IF gain is normally regulated by AGC. 
AGC decay normally controls S meter hang time.
This ain't normal. 

Yes I know there's no IF. Not a problem for software though

George KF2T

You're quite right, Steve. I'd suggest that the 6000 Series operates differently than legacy radios, and understand that can be discomfiting. I enjoy the way it works, now. It did take a while for 30-odd years of habits and expectations to change. Of course, "a simple matter of programming" could provide legacy-like features, if desired. Sounds like an excellent suggestion for the FRS Lab. For me, I prefer the s-meter to show input signal level, and RF gain to vary signals on the scope. Perception after demodulation is important, but probably (again, for me) should be downstream. Dang it, now you're making me want to troll eBay for a used but good 2-B. Fine receiver, it was. Loved the copper chassis, cheery glow, and overall look and feel. Have a great weekend & thanks for the exchange.

Steve N4LQ

Yes George, the 6000 takes you back to how the SX-28 worked. Turn off the AGC, ride the RFG and ignore the S meter. Retro at its best!

Al_NN4ZZ

Steve,
What do you think about this idea to automate the AGC-T setting?   It doesn't address the issue you raised but at least reduces/eliminates the frequent adjusting that is currently required.  

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

Regards, Al / NN4ZZ  
al (at) nn4zz (dot) com

k0eoo

Hi Steve,  I don't think the F6000 ACG-T works like the AGC in the SX-28 but it does work like the THRESHOLD RF Gain control on the Collins 75A4 and all S-Line, KWM-2 and later Collins rigs....  Very advanced for the time.  You can set it to cut the band noise without effecting the audio level of the signal, a problem earlier RF Gain controls had, like that used in National, Hallicrafters and Hammarlund receivers.  I think you can find a description of how THRESHOLD AGC works in any manual for the early Collins Receivers.

Of course, any such implementation isn't going to be perfect but what Flex has done with there AGC-T works better in software than the Collins Radio implementation does in hardware, at least on my radio's....

I for one really like how the AGC-T control works on the Flex.  I would only like them to add a horizontal dotted line across the panadapter screen to indicate where the THRESHOLD is set, that would be helpful....

I might add, in the Flex, the AGC-T does not effect the S-meter reading like it does in all older and big box radios, which I see as a distinct advantage over other implementations.

Steve, hope I didn't miss the point of your post.  If I did I apologize...

Regards, Dennis, k0eoo

Steve N4LQ

Dennis.
I've worked on Collins equipment for many years. The AGC in them is just typical of most any radio. The IF signal is amplified, rectified and the resulting voltage is fed back to RF and IF stages. DC bias for those stages is provided by the power supply and made adjustable by the RF gain control. There are no "threshold" adjustments but there is the usual slow, med, fast switch positions for AGC decay time. 
The problem with that era of Collins receiver was slow AGC attack time which resulted in thumps on strong cw stations. Various mods were attempted over the years.
The problem with our Flex machines is there seems to be virtually no AGC at all until the signal is very strong. This results in annoying volume fluctuations as received signals change strength.

KC2QMA_John

Like I said, That's why especially in a SDR you must have AGC Parameter controls (Attack Time, Hold Time, Decay time & Threshold). Because because the flex has soooo much RF Dynamic Range these controls become much more critical than a traditional receiver And yes having a dotted line or some other graphical representation would be a nice touch.

There is another way to approach this. Use Audio Compression. With audio compression you are only limiting the audio not the RF signal. The best would be a combination of RF & AF AGC! Then you could decide do I want RF AGC, AF AGC or Both and guess what with SDR the possibilities are almost endless!

That was one of the things on my wishlist https://community.flexradio.com/flexradio/topics/kc2qma-smartsdr-wish-list-like-who-doesnt-have-one

73/KC2QMA

John  

k0eoo

Steve,

R9 in the 75A4 and R57 in the 75S1 set the AGC Threshold just above the receiver noise floor, plus, in these receivers the RF Gain is controlled by adjusting the AGC negative bias, base line, up or down.  That's why you see the S-meter in those radios move up or down indicating the input signal threshold. point.

In other radios of the time the RF Gain control adjusted the gain of the RF and IF amplifier tubes and was totally independent of the AGC system... In those radios the RF gain control was usually in the cathode circuit of the RF and IF amplifier tubes.

Steve, I would like take any further discussion on this subject off line as not to bore the community with such OLD information.  My email address is on QRZ....

Best Regards, Dennis, k0eoo

k0eoo

Hi John,

You bring up a vary good point.  Because the Flex has such great dynamic range a RF GAIN control is really NOT necessary.  That may be another reason why Flex implemented a design whereby they vary the AGC threshold and not the apparent gain of the signal.  I like the fact that the S-meter measurement doesn't change as you vary the AGC-T control or as you tune across a signal...

And, yes John, maybe some form of receive audio compression would be a good solution.

Regards, Dennis, k0eoo

KC2QMA_John

I also prefer the s-meter to be independent from AGC as it is now. I want to know what the signal is at the ANT Input because that’s what counts to me. I also think of my 6500 like an RF analyzer measurement device.

I think at a minimum if they can adjust the ratio of the threshold to track stronger signals better things will work better. But for me control over AGC parameters are a must!

73/KC2QMA

John  

Tim - W4TME

The AGC-T and AGC algorithm is working as designed.  There are some opportunities for automation since the current implementation is fully manual.  I anticiapte that we'll eventually focus on this at some point once we have a few other things in the software addressed.

Steve N4LQ

Here's a little video I made showing how to adjust AGC-T per the manual - Page 36- and per Ken's detailed and helpful instructions. 
https://www.youtube.com/watch?v=Hwm6rszuMis&feature=youtu.be

KC2QMA_John

The video clearly shows the flaw with the AGC algorithm. It also shows that you have to be willing to deal with a lot more band noise in order to have any AGC action. You don’t want to have to unnecessarily raise your nose floor up so high just so that the AGC works. That's why you must have AGC Parameter controls.

Like I said the other option is to use Audio Limiting to deal with the differences in sound level.

73/KC2QMA

John

KC2QMA_John

https://community.flexradio.com/flexradio/topics/kc2qma-smartsdr-wish-list-like-who-doesnt-have-one

Al_NN4ZZ

Steve,
Good video, I only operate CW and thought FAST was the proper setting so never tried SLOW but from your video it looks like it might be better.  I'll give it a try.   I was adjusting AGC-T a little differently but ended up close to the same settings as you.    (when I start hearing the band noise, I adjust up about 20-30 more). 

It sounds like (based on Tim's comments) that the request for an automated AGC-T feature might be added at some point.  It would sure take a lot of the fiddling out of the equation.   You could just set the threshold you prefer and it would adjust automatically.   In addition to the variations due to the band in use, and the band conditions at the time, if you have a gain antenna the noise level changes depending on the direction your antenna is aimed.   

Regards, Al / NN4ZZ  
al (at) nn4zz (dot) com