AGC issues with DSP radios?

[Mark Roberts]

On the radios that I have that are based on the Silicon Labs SI4734 (or related) chip, I sometimes hear a weird pumping effect on some (but not all) AM stations. It's worst in the 6 kHz bandwidth setting, and disappears almost entirely with a 3 kHz bandwidth. I can also mitigate it by detuning from the desired station by 1 kHz.

Here are some brief samples, taken from Albuquerque a couple of weeks ago (I never encounter this in the Bay Area; I have encountered it in Sacramento and some other locations). The receiver was a Tecsun PL-310ET, but I've encountered it on other Tecsun models and on the Digitech AR-1780, the Australian version of the XHDATA D-808:

KTBL, 1050, Los Ranchos de Albuquerque: https://www.mediafire.com/file_premium/5cjuxzzx4sra3h0/KTBL-Sample.2023.05.25-1705.mp3/file

KDNF, 840, Belen, NM: https://www.mediafire.com/file_premium/igbbgyizuvd5ews/KDNF-Sample.2023.06.05-1556.mp3/file

KDNF, 840, offset by 1 kHz (i.e. tuned to 841 kHz): https://www.mediafire.com/file_premium/hcjof71aml9lc9p/KDNF-Sample-Offset.2023.06.05-1657.mp3/file - sounds a tiny bit noisy but much better than when centered on frequency.

Are the affected stations using a particular method of modulating their carrier that causes this effect? It appears that the radio's AGC is confused and is constantly trying to readjust itself.

This effect does not happen with non-DSP radios.

It's very curious, and a bit annoying.

 

[gar hi]

The fading in and out of AM or shortwave stations in sudden increments instead of being gradual is annoying but if you look up DSP radios, they claim what a great invention it has been and almost all radios now have it.

I really like my C Crane Skywave and if it does have DSP, it's effect is subtle unlike other radios I've gotten in the last 15 years or so.

 

[Mark Roberts]

The fading in and out of AM or shortwave stations in sudden increments instead of being gradual is annoying but if you look up DSP radios, they claim what a great invention it has been and almost all radios now have it.

I really like my C Crane Skywave and if it does have DSP, it's effect is subtle unlike other radios I've gotten in the last 15 years or so.

Keep in mind that the effect that I describe does not happen with all or even most stations. There seems to be something specific to stations such as KTBL and KDNF that causes the AGC not to work properly. I find it hard to believe that this is by design. I don't know much about how AM transmitters work these days - is there any form of carrier reduction going on in some transmitters?

 

[awsherrill]

There are some AM stations running MDCL (modulation-dependent carrier level), which varies the RF carrier level according to modulation.

As I recall, there are two methods of doing this. One method reduces carrier level as modulation increases, and carrier is at full strength with zero modulation. The other method is the opposite; carrier level drops as modulation decreases, with full carrier at full modulation.

The intent is to reduce electric consumption and save money. Some operators say they have realized up to 20%-30% decreases on their power bills.

Here are a couple of (old) links with information:

Crawford saving money with MDCL

WOR sees significant savings using MDCL

One of our AM stations has a transmitter with this feature built-in, and have been considering trying it out when I have some time to fool with it.

 

[Schroedingers Cat]

You can see that modulation dependent carrier level clearly on an FIM-41.

One fairly nearby 50000 watt station shows about a 30 percent decrease in field strength. There is some hysteresis with the d'Arsonval meter movement.

I have noticed that the 1000 watt stations, including Class IVs/Class Cs with no carrier reduction, sound better now than the 50000 watt stations that use it. Used to be that it was the 50000 watt stations that sounded better.

 

[Mark Roberts]

There are some AM stations running MDCL (modulation-dependent carrier level), which varies the RF carrier level according to modulation.

As I recall, there are two methods of doing this. One method reduces carrier level as modulation increases, and carrier is at full strength with zero modulation. The other method is the opposite; carrier level drops as modulation decreases, with full carrier at full modulation.

The intent is to reduce electric consumption and save money. Some operators say they have realized up to 20%-30% decreases on their power bills.

Here are a couple of (old) links with information:

Crawford saving money with MDCL

WOR sees significant savings using MDCL

One of our AM stations has a transmitter with this feature built-in, and have been considering trying it out when I have some time to fool with it.

I'm suspecting it's something like this. Whatever it is, it's confusing the AGC.

I may be too far away from KCBC to see if I notice any effects there, but I'm going to try receiving it over the next few days and will follow up here.

 

[boombox4]

It depends on the radio, and probably the way the SiLabs chip is programmed, or even the firmware inside the SiLabs chip, which I would guess has been changed over the years that their DSP chips have become common in radios.

My Grundig G2 (2012 model, bought in 2014) is Degen made and probably has an Si4731, and it has fairly tight AGC and pumps on MW. It doesn't pump on SW. My Sangean PR-D5, which is a post-2012 model definitely has an Si4731 chip, and the AGC doesn't pump. However, the AGC on that radio is too slow.

My Sangean PR-D18, which has a SiLabs chip, doesn't pump. The AGC is better than the PR-D5, but not too tight like my G2.

My new Sangean PR-D4W has a Si473__ chip and doesn't pump, and the AGC is not too slow. In fact, all aspects of the audio with the PR-D4W are perfect. It's probably the best MW radio you can get. I was surprised.

My Tecsun PL-398MP, which has a Si4734 chip, doesn't pump on MW or SW.

 

[KenoGuy]

My PL-310ET will occasionally do the same thing, but the bigger issue is that the AGC is just too aggressive. It soft mutes on signals that are plenty strong to not need AGC.

I'm pretty disappointed with that radio and seldom use it, although the FM is very good. My PL-380 is a MUCH better radio.

 

[timetraveller]

If it's not the AGC, then the problem here might also be the soft-mute software "hunting" for the right audio attenuation value under certain varying signal strength conditions, like the MDCL carrier variations. I had a radio that did this "buffeting" (I call it that) quite a bit, and I can't remember which radio that was at the moment.

AGC and soft-mute are separate functions on these chips, and are not related to each other. There are separate software controls for each. Base values are set at chip power-up. The radio designer may even elect to disable them if he wants. That would be a poor idea for AGC of course, but the DSP radios issued over the last few years have generally toned down soft-mute quite a bit. The first of the DSPs, like the PL-380, implemented soft-mute at full force. Newer ones, usually not so much.

Bill WE7W
RADIO-TIMETRAVELLER

 

[Mark Roberts]

If it's not the AGC, then the problem here might also be the soft-mute software "hunting" for the right audio attenuation value under certain varying signal strength conditions, like the MDCL carrier variations. I had a radio that did this "buffeting" (I call it that) quite a bit, and I can't remember which radio that was at the moment.

AGC and soft-mute are separate functions on these chips, and are not related to each other. There are separate software controls for each. Base values are set at chip power-up. The radio designer may even elect to disable them if he wants. That would be a poor idea for AGC of course, but the DSP radios issued over the last few years have generally toned down soft-mute quite a bit. The first of the DSPs, like the PL-380, implemented soft-mute at full force. Newer ones, usually not so much.

Thanks, Bill. Before I write anything else, I want to let you know how much I've enjoyed your site over the years.

As for soft muting, that's an interesting possibility. Would it apply to strong signals, though? One of my examples, KTBL, was an Albuquerque station (1 kw DA aimed north from a site in Albuquerque's South Valley) received just outside Albuquerque (near the Rio Rancho/Corrales border on the south side of Rio Rancho).

I've also noticed soft muting being scaled back. The Sangean DT-800 lets you turn it off altogether. For Tecsun models, I have both a PL-380 and a PL-320, the latter just purchases, and the PL-320 seems to have less soft muting than the 380.

 

[boombox4]

My PL-310ET will occasionally do the same thing, but the bigger issue is that the AGC is just too aggressive. It soft mutes on signals that are plenty strong to not need AGC.

I'm pretty disappointed with that radio and seldom use it, although the FM is very good. My PL-380 is a MUCH better radio.

Like Bill noted, it almost sounds like two different issues. Soft muting is when a DSP radio drastically cuts the volume almost to zero when a signal drops below a certain threshold. My Radio Shack Pocket Radio (which has a SiLabs chip) does that. The soft muting is obvious in that radio's case.

Then you have the 'capture' effect which some DSP chips seem to exhibit (my PR-D5 has that), where the chip is attempting to resolve a weak signal. It can sound like soft muting, but it's not soft muting. My PR-D5 has slow AGC, which exaggerates the sound of the capture effect.

Then you got good ol' AGC, which, when it's tight and pumps on a DSP radio, you really notice it.

My Radio Shack DX-398, an analog chip radio from the late 1990s, had very tight AGC, so tight, pumping AGC isn't just the realm of DSP radios. There were articles mentioning the 398's tight AGC when the radio first got popular.

The main difference I notice in DSP tight AGC (like my G2 seems to have) is that the tight AGC distorts the audio slightly when the AGC is really pumping.

 

[timetraveller]

Thanks, guys.

Mark-
"Would it apply to strong signals, though?" Yes, it could. The software knows how much soft mute attenuation to apply by testing the signal to noise figure. It's constantly testing this figure, and has the ability to check it too on adjacent channels.

I also read that the Sangean ATS-405 had the ability to defeat soft mute by key presses. I didn't know about the DT-800.

boombox4-
Great observation on the "capture effect". I find this capture effect to be even more evident on the analog tuned DSP radios. I've been trying for a few years now to pin down the tuning and software details on how this is done. The XHData D-219 I bought gave a lot of clues. I wrote a blog article on the D-219, which might be of interest.

Notes On The XHData D-219 Analog DSP Radio

My D-219 came a couple of weeks ago and I've been checking it out. A fair amount of discussion has ensued about this radio, and I think feel...

radio-timetraveller.blogspot.com

Interesting things, these DSP radios.

Bill WE7W
RADIO-TIMETRAVELLER

 

[Mark Roberts]

Thanks, guys.

Mark-
"Would it apply to strong signals, though?" Yes, it could. The software knows how much soft mute attenuation to apply by testing the signal to noise figure. It's constantly testing this figure, and has the ability to check it too on adjacent channels.

I also read that the Sangean ATS-405 had the ability to defeat soft mute by key presses. I didn't know about the DT-800.

Thanks for the insight. As for the DT-800, the control to defeat soft muting is on the front panel! With the radio on, just press PAGE until you hear a beep, then press 3, which is also labeled "S. MUTE". If it's on, "S-MUTE" will appear in the lower right corner of the frequency display.

Interesting things, these DSP radios.

I'm much more of an FM listener than an AM listener and, for FM, they're amazing. The picture for AM is a little more mixed, but they certainly add tuning and various other conveniences I would have loved to have had when I was much more of a DXer than I am now.

 

[RadioFan2J3]

There are some AM stations running MDCL (modulation-dependent carrier level), which varies the RF carrier level according to modulation.

As I recall, there are two methods of doing this. One method reduces carrier level as modulation increases, and carrier is at full strength with zero modulation. The other method is the opposite; carrier level drops as modulation decreases, with full carrier at full modulation.

The intent is to reduce electric consumption and save money. Some operators say they have realized up to 20%-30% decreases on their power bills.

The interesting part about the power savings is that they may be very dependent on the programming.

If we were to compare AM broadcasting back in the days of the modulation level wars, typically, with the rockers keeping the modulation monitor levels near 100% negative and even higher positive modulation levels, running the transmitter in the MDCL mode where the carrier is at full strength at full modulation, it would seem there would be very little power savings.

On the other hand, using the MDCL process where the carrier level drops under modulation, there could be considerable power savings as the typical carrier level could be 3dB to 6dB below the normal licensed power. J. Fred Riley wrote an opinion piece on this very issue claiming that if a station could achieve satisfactory coverage area running a power cutback, perhaps the licensed power level is too high. (Yes, I tried to find the article - seems to me it was a front pager on Radio World.)

If the programming content is far more moderate in modulation levels, perhaps an example could be talk radio audio, with pauses, then the MDCL process where the power increases with modulation, there would be some savings versus full power operation, and perhaps equal power savings using the opposite MDCL process.

Another aspect of MDCL operation is if the station is on generator power. Perhaps again, dependent on type of programming, but at one location I worked at, we had 100% backup power with diesel generator sets. I did some quick and dirty tests looking at the governor systems for the generator sets and the reactions of the fuel racks controlling diesel fuel into the cylinders, I decided if we had to go to the generator power, that we'd go to full power, no MDCL operation, just to save a bit of wear and tear on the generator sets.