Re: Precision Modulation Monitor - NOT! Author: MWeiss7401 Email: mweiss7401@aol.com Date: 1998/01/24 Forums: alt.radio.pirate _________________________________________________________________ >Why is it [LF353] strained to the limit [at 50 kHz]? Because of the high "open loop" gain >potential? >Because the device isn't fast enough to recover between cycles at rates above about 50 kHz. At 50 kHz, a precision rectifier using the LF353 can still catch the peaks, but the waveform is highly distorted. Take those 3.3 uFs out and look w/your scope.< I'll have to take your word for it on that, since the caps are soldered in on my prototype and would be a bitch to remove. I thought the LF353 was good to 2MHz at unity gain, that's why I use it in so many ultrasonic applications. I think if it distorted 50KHz, then my stereo generator would suffer awful channel separation, instead of topping 60dB from 20-15KHz on the Day-Sequerra Model 1 FM tuner I tested it with. I will do some testing of the chip on a diagnostic level this week and report findings. >I don't use a current limiting resistor to drive the meter, strictly for >damping reasons. The fast transient response of the meter is made possible by >the use of maximum current drive of the op amp. It was too sluggish in my >original design using a current limit resistor, so I eliminated it and chose my >op-amp carefully, for one with the best offset characteristics. In fact, I was >thinking of bootstrapping additional op-amp stages on the output to increase >the drive current for more instantaneous power to accelerate the meter from a >zero reading to full scale. Under normal program conditions this is quite >unnecessary, but I'm always striving for perfection here. >>I think ya got me on the point about the current-limiting resistor. Glad we agree. :-) >deleted< >Okay, here's my proposal for a simple modification: > >Install a switch which allows one to sample just the negative rectifier or the >positive rectifier, or both, so in effect three choices would be available for >monitoring: > >1. negatve deviation >2. positive deviation >3. total deviation (the default mode) >>Sounds good to me. I'm reworking the design this month. >Just a comment on the NCR tx, if that sounds as good as most commercial >stations, you must be in a market run by PDs and GMs engaged in volume wars >with no concern over excessive distortion. The Minneapolis market is suffering >from this malady at the moment. >'Taint "scientific", I know, but the tx does sound good. Maybe it's because I'm running it with no audio processing ahead of the modulator. I tacked in a little circuit that flashes an LED when the baseband level input to the mod diode exceeds a predetermined peak voltage. I ride the gain and try to keep that LED just barely flickering. I need to build up some kind of mod limiter. Want to keep it as simple as possible, though.< I conduct measurements both with my ears and with a scope attached to the tuner's baseband output. You're running without a limiter? Not even a simple one? Sheesh... To keep it within design allocations for the US spectrum, the volume levels achieved would be pointless on a LPFM. Get some processing, you'll double your listenable range. >{I would also make the following point about modulation: I will grant you that we must make EVERY effort not to interfere with commercial broadcast stations, and this would include not allowing your modulation peaks to slop over into the next guy's channel, but sticking slavishly to the +/- 75 kHz FCC-imposed legal limit seems a little silly. After all, aren't we, by definition, "illegal" just by the fact that we're on the air in the first place! If it sounds good on a typical receiver [i.e. does not exceed the modulation bandwidth of the receiver], who cares if it's +/- 75 kHz, or 85, or 95?? (anything over +/- 100 kHz, of course, and you ARE into the next guy's channel)}< This I have to disagree with. Yes, there was a time when I though that it was technically getting something for nothing by modulating to 120%, then when I tightened up the equipment to make some serious effort at a saleable product, I made it a personal challenge to achieve loudness with 100% limits. In all the times in the past when a listener was just outside my fringe area, the observation was that the signal was "scratchy" and noisey. Since I locked the upper limit strictly to 100%, that same listener still hears lots of noise when the music stops, but when modulation exists, it's perfectly listenable now. All that gained by cutting deviation by 20%. Tuners in fringe areas are less forgiving of overmod. I remember for years, there was this famous LI station that used to overmod all the time, and they were always scratchy on my Karg Labs tuner I had back then, while the other stations, though weak, were clean. Furthermore, in a multipath-ridden mountainous area, overmodulation causes gross distortion when the motor-vehicle receiving the broadcast is passing though one of those dips in the signal. Again, since observing strict 100% limits, it's as it I raised the antenna another 100', the weak areas don't break up anymore, and the modulation is loud enough to mask the noise. Heck the treble output of my design is so strong that even when the radio's hi-cut filter kicks in on weak stations, mine still sounds crisp and bright as the strong stations. Overmod only "seems" to work on local receivers with a wide IF bandpass, but it effectively reduces your useable coverage area. Nobody will listen to scratchy sound, but they will listen despite a lot of hash in between songs. With my processing, it only takes a 20dB s/n ratio of reception to make the audio's apparent dynamic range seem as good as a strong local station. You can't hear the noise at all with active program material. >Also a tx with .2 % distortion is not very suitable to good stereo >transmission. With that much distortion, if it's truly that severe, maximum >theoretical channel separation possible with an ideal stereo generator would be >about 25dB. Perhaps the .2% was in stereo, and of the individual audio >channels, not the basband modulator. >Yeah, 0.2% is the distortion figure I got for the NCR tx. Test conditions: 1 kHz test tone driving both channels simultaneously in stereo mode, input level adjusted for 100% modulation, 4 kHz LP filter switched in on service monitor (so the distortion meter would not see the pilot). The same unit tested 40+ dB for separation. I dunno. What's the formula that specifies theoretical separation vs. distortion? Max< There is a direct relationship between the powerlevel of the distortion biproduct and the intended signal. In stereo, you have a matrix, which is balanced precisely. Equal amounts of subcarrier and main combine out of phase to create a perfect null, cross-channel, this is the basic foundation of stereo as it exists today. When you add a distortion biproduct to either main or sub, let's say, -20dB, that's equivalent to 1% distortion. But that -20dB will appear in the output, may not quite at -20dB, because it only makes up half of the signal, so at best, channel nulling will provide -23dB output, which would be the effective channel separation on a system with 1% THD. Looks like you distortion readings are taken through the whole chain. The modulator must be about .02% distortion to achieve the 40dB. What is the distortion at 20 hertz? Separation at 20Hz? What are they at 15,000 hz? These are the figures that matter. Do you hear several beat tones on the 15KHz, or is it relatively pure tone? (Of course that depends mainly on your receiver, since most have too much non-linearity in the IF strip to maintain good intermod between tones and pilot signals, but it's useful data if the tuner is a laboratory grade.) Mark The "Peg-legged" Bass Pig ^ ^ _o^O-/ \ (oo)_ ?) _/ ) \_ SPAM-Pruf http://users.aol.com/amn92/amn.htm "I support the micropower broadcasting movement and freedom of the airwaves" _________________________________________________________________