Time for new Speed Limit? 150% AM

[Tom Wells]

I have read this over before, a quick look at the chart of sample radios is very telling.

Why is the 100th percentile bandwidth radio 10 db down at 9600 khz?

Top of my head response -- it's probably one of the Wal-Mart cheapies with a minimal amount of IF filtering. If so, it will squeal like the proverbial stuck pig after sundown because of 10 kHz beats.

Making a usable wideband radio used to be expensive because of the need for a stable 10 kHz notch filter and a quick rolloff above 10 kHz. DSP-based radios make this cheaper, but they are still not down to the $20 price point that many customers look for.

A passive filter in the low-impedance path of the speaker is simple, stable and cheap. It can achieve a >very< narrow -42 db notch with only 2 db insertion loss. I found the design in a Radio-Electronics magazine from 1947 presented as an improvemnt over trying to create a tight notch in high-impedance
circuits, where you can't achieve high enough Q. It was designed for use in AM hi-fis of the day, most of which whistled pretty well.

You don't want to roll off the high frequencies, it is the audio equivalent of sharp vision. You do want to kill the whistle itself.
We correct vision loss and hearing loss in individuals. I cannot accept creating "hearing loss" intentionally to an extent that it's crippling intelligibility
and makes music sound awful.

I know parts count in cheapies is critical, but this requires one inductor, two caps and a resistance calibrated to R of the voice coil.
I have added them to almost all my AM radios and leave them in circuit, where they do affect the FM sound minimally.
When you hear AM with high frequency response above 10 khz with the beat only stripped out, you'd never believe how much
info is there, masked by the whistle or in the roll-off of narrowband radios.
For a decent quality radio, the parts count increase is well worth the result. There is no good reason why such ready-to-go networks could not
be available to match standard speaker R values, resistance, not impedance. It would be one more part on a board.
If a low-r trimmer were used, one design would match all speakers. I can find the values and post this simple circuit if any are interested in adding
it to your radios. I also wound one up for 5 khz for shortwave listening on the '36 Philco, but made it switchable.

And yes, there's monkey chatter on overlap of program material when signal ratio of desired/adjacent gets too low.
Rotate the antenna, try sidetuning to see if one sideband is cleaner,
then consider that maybe the radio needs alignment or is hopelessly wide.
Or that you're listening to a pretty weak signal.


Cheap radios aren't necessarily insensitive. Every single radio is subject to being a "good one" or a dog, unless the design itself
offers no hope due to inadequacy.

 

[rrsounds]

A passive filter in the low-impedance path of the speaker is simple, stable and cheap. It can achieve a >very< narrow -42 db notch with only 2 db insertion loss. I found the design in a Radio-Electronics magazine from 1947 presented as an improvemnt over trying to create a tight notch in high-impedance
circuits, where you can't achieve high enough Q. It was designed for use in AM hi-fis of the day, most of which whistled pretty well.

I think the need for extreme depth of the 10kHz notch filter is over-rated.
IMHO, the beat needs to be reduced to below the rest of the audio, that's all. Even 10-15 dB reduction would be more than enough in most cases.

In an AM radio with DSP onboard such a notch filter could be dynamically applied, only coming into play when interference warrants it.

One of these days, DSP will be cheap enough to put into even low-end radios, then AM will have a chance. All kinds of tricks can be applied to reduce annoyance factors. At this point it comes down to economics.


Kind Regards,
David

 

[OKCRadioGuy]

"One of these days, DSP will be cheap enough to put into even low-end radios, then AM will have a chance. All kinds of tricks can be applied to reduce annoyance factors. At this point it comes down to economics."

Let's hope so. With DSP radios the radio could auto-configure for local and distant adjusting bandwidth. For the current bunch of junk out there, would there be anything wrong with making a radio with a narrow/wide switch that defaults to narrow every time a channel is changed? That way the "return rate" wouldn't be a problem for the truely stupid comsumers out there, but it would allow the rest of us with our hearing still in tact to push a button and listen to radio instead of crappy telephone audio.

 

[rorban]

A passive filter in the low-impedance path of the speaker is simple, stable and cheap. It can achieve a >very< narrow -42 db notch with only 2 db insertion loss. I found the design in a Radio-Electronics magazine from 1947 presented as an improvemnt over trying to create a tight notch in high-impedance
circuits, where you can't achieve high enough Q. It was designed for use in AM hi-fis of the day, most of which whistled pretty well.

Exercise of the reader: Assuming each filter is hand-aligned at 70 degrees F, what temperature stability (percentage change in capacitance and inductance per degree Fahrenheit) is necessary for the inductors and capacitors to create a minimum 25 dB notch over the typical dashboard temperatures found in a vehicle (at least 30 degree F variation)?

This sort of thing is called "sensitivity analysis." It's a vital part of the design of mass-produced electronics. Drawing the schematic and calculating the nominal values of the components is just the beginning. The real thing that separates the men from the boys in the world of mass-produced electronics is determining if their circuits can be manufactured using components of a given tolerance and how much long-term drift can occur before the circuit performance goes out of spec.

DSP has made this less important because stability now depends on one more crystal oscillators instead of a bunch of analog components. However, DSP design requires its own kind of sensitivity analysis because a finite word length effects. The essential difference between this kind of sensitivity analysis and the kind done on analog circuits is that once the analysis is done, every DSP works the same.

If you look at the schematic of an 8100 and wonder why some components have different tolerances than others, even within a given filter, the answer is that the circuits all went through an extensive sensitivity analysis. That's one reason why 8100s worked correctly after coming off the assembly line and kept working for many years.

 

[Tom Wells]

While ageed that DSP and reference to stable sources make this an easy fix upstream, and part tolerances
in some points are critical, I must say that I find no drift in the notch frequency of the filter in use in my own car, where I attached the filter
to the outside of the radio. Temperature is not an issue with the Q achieved at a 8 or 4 ohm impedance level.
If I had to be under the dash, adjusting the trimmer (ever), or found it to be be ineffective/drifted in winter/ summer due to impedance or capacitance change, I would not be advancing this as a good idea.

Hand-aligned circuits are good and bad, depending on what aspect of performance this may help/hurt.
I'd rather have a piece of equipment I can align than one which can only be thrown away.
I'd also rather have the simplicity of 4 discrete parts over an untold number of parts and firmware to make a DSP notch filter.

The points regarding stability certainly do apply to high-impedance circuits.
I built an op-amp "infinite gain" notch filter back in 1981 for 60 hz, 14758 (TV sweep noise) and 19 khz.
6 op amps for 2 channels and 3 frequencies, worked great, but was very hard to keep the variable R in the circuit peaked for
perfect notches on 14758 and 19khz..
Whether the trimmer pots were too touchy or the caps changed with temp, I can't say.
The 60 hz notch did not drift for obvious reasons.
I should dig out and reactivate that board just for hum reduction, it was stunning in its ability to strip hum.

 

[amfmsw]

RE: GM Kennedy.

First, I'm not a blow hard. Second, I'm speaking as a dissapointed consumer of his product. It's AM audio performance was, and is widely agreed upon, dreadful! The #1 selling radio in the world is Delco, and it was terrible. How can anyone defend a 3Kc bandwidth. What were they thinking? Here's an idea, lets force the FM sections to meet that GM standard, and the industry would have revolted.

I'm sure Mr. Kennedy's a fine fellow, but after buying his radio, it's clear the thinking behind that design is flawed.

We, the US AM Broadcasters have accepted the NRSC standard. the issue before us here is to get NRSC compliant radios into cars and home receivers, not $20 throw away radios. Although my $16 Sony AM/FM Walkman from K-Mart has great sensitivity, selectivity, and crisp clear, intelligible AM! Whay cant that be in my GM car?

 

[thebeaver32]

I've skimmed over this thread, and here's my $.02 worth.

You guys are re-hashing the past. They're NOT going to eliminate NRSC filters. Car radio makers AREN'T going to increase the bandwidth of the AM receivers.

AM is dead. Yes, I know that some cities' top-rated stations are AM. But AM is dead. Even the 50 KW's, you can only hear them in the cities, you get a few miles into the country, and it's static city. AM receiver quality is ZILCH, as you've been talking about here. The noise floor for the standard broadcast band (the AM band) has risen dramatically since radio's inception. It is NOT going to get better. The FCC let THAT horse out of the barn beginning 50 or more years ago. And they're even piling on: advocating BPL to send Internet signals over power lines is going to make it much worse, if they ever get around to actually doing it.

What needs to happen is, the FCC needs to open a new broadcast band. The best suggestion I've heard is the newly freed up TV channels 5 and 6, or at LEAST channel 6. It's already adjacent to the present FM band anyway. Give EXISTING AM STATIONS the first priority in applying for stations in the new band. Daytimers first, 24-hour stations next. AM stations that get new FM assignments will have a few years to simulcast, then the AM license disappears. Permanently. The rest of the details can be hashed out, such as the power levels/coverage of the new stations. There are LOTS of advantages for AM stations to do this: 24 hour operation, stereo, better quality audio, all the advantages of FM over AM.

It's time to let AM die a peaceful, easy death and move on to the next big thing.

 

[OKCRadioGuy]

The think I wonder is why iBiquity and NAB haven't put any force into getting the TV channels for this use. My guess is the NAB has it's nose up the TV companies' asses so far they could care less what happens to AM broadcasters. It would be slightly inconvenent to the TV folks, so the rest of us be dammed. Forget what would be best for the consumer.. We want to protect the TV companies on this issue. iBiquity knows most AM broadcasters are poor so they could care less themselves. Sad...

What exactly borders us in the 1700-1800kc region anyway? There is some merit to diplexing on something like that band. The patterns would be pretty much the same. Pure iBuz or DRM up there would do wonders for a local market at something like 1kw day 500 nights.

 

[FredRichards]

Bev & OK, putting radio in the TV spectrum does have some merit, but I would suggest that anything new go in the UHF band for several reasons. 1) If you read the FCC's comments, and replies to comments, you'll see that all the dodos wanting digital have stated one thing openly. They don't care about distant listening. 2) Therefore, put the the signals in an area which is pro-portable propagation and anti distance. As I have mentioned before, Ch 14-18 are ideal. 3) With newer portable devices it would be easier to manufacture the devices with antennas that favor the UHF spectrum. Keep in mind the size of iPod and MP3 players people use. And 4) there is already a move to populate open UHF channels with other services. Google "white spaces UHF".

But again I would like to point out that the reason AM is dead is not a technology issue, it's a content issue. If you have something that people want, they will go it it. Problem is; there is nothing on the AM or the FM dial which really is relevant, other than some people with an over inflated impression of what they have to say, thinking that everyone else is hanging on their every word.