First post/Beyond the broadcast bands

Welcome to the forum and I hope we will be able to help a bit.

I, too, have experimented with non broadcast band unlicensed radio, especially the 900 MHz band and I share your frustration with the need to use SMD parts. My approach as been to make breakout boards for the SMD chips/modules and use conventional through hole support components.

You mentioned not being able to have latency and you must have your reasons but latency should not be a problem for broadcasting. I am now experimenting with Bluetooth and have acquired a RN41 module from Sparkfun. Unfortunately, this is not the module used on their Bluetooth breakout board. They use the RN-42 which is a Class 1 device meaning about 30 feet of range. The RN-41 is Class two with 300 foot range. Not all Bluetooths (or is it Blueteeth?) work the same.

If you are interested in conventional FM there are some inexpensive options such as purchase a FCC certified transmitter or the Ramsey FM-25 kit.

AM is also an option if you don't need stereo. I have built a 15.219 compliant station which gives a mile range on a car receiver.

I also use Arduinos for things and we can kick around some ideas if you want.

Need to leave right now, and perhaps you can explain a bit more about what you are trying to do and we can discuss hardware, etc.

Neil

Neil,

Thank you for replying. I was afraid after I was done that people would just say tl;dr, or that this forum is only for people running community LPFMs or something. I'm grateful that my post was authorized and got a reply.

I said my reasons for these RF projects didn't matter. I've just had a lot of experience on forums where I've specifically said "I want to do X  to achieve Y," and the replies are all "Why aren't you doing A, B, and C to achieve Y" and we never get to the question I asked.

I mentioned wireless IEMs for musicians for a reason: I am a semipro (starving) musician who, among other things, could benefit from wireless monitoring. I'm not touring the country so $900 Sennheisers are out of the question, but even 100 miliseconds of delay between the time you pluck a string or hit a key and the time the sound arrives in your ears makes a system useless. With more and more of the gear we use being digital, the goal becomes removing any *additional* latency from the signal path because 20-30ms of latency is already there. Bluetooth can have up to 500ms latency, and I already have some Bluetooth consumer gear in the house, and at least analog RF gets weak before it cuts out completely. So in my current situation, Bluetooth does not perform acceptably well even if latency wasn't an issue.

Even if I'm working on a non-music audio project like old tape restoration, latency is at the very least annoying, because you still need feedback when you adjust the occasional EQ curve or whatever. I did a project last year that was hundreds of hours of tape; you could leave it and just monitor that it was okay for long peroids of time and occasionally make a minor adjustment or dial it in over time. It would have been nice to be able to get up and work around the house during those lengths of time and still have the audio in my headphones.

I also live in a fairly noisy environment, and it helps if I have even a cheap pair of isolation buds in most of the time. So that's how this whole thing got started. If I'm going to build an IEM system for myself, it had better be low-latency for the times I need it to be, even if I'm sometimes just listening to pre-recorded music while working around the house.

Therefore, by "security through obscurity," I just mean that sometimes I'd be broadcasting audio that was horrendously awful or embarassing (say I'm trying to learn a new song by playing along and the first two hours sound like Jimi Hendrix after suffering multiple strokes) and I would rather it not be available to every car that drives by, even though I am not broadcasting sensitive or classified information.

So, the application seems fairly standard to me. A base station, with enough power that it gets around the house and the yard, either an analog signal or a low-latency digital signal (the TI PurePath chips are like 25ms, which is acceptable), and enough of an interface to a microcontroller that I'm not building extremely-sensitive, voodoo-heavy RF boards that even a lifelong EE would have trouble with. Grab a bodypack-style enclosure off Digikey and build a complimentary receiver into it; again, with a nice strong detector and eventually enough of an interface to a microcontroller where I can have a little display, maybe a battery life indicator, maybe multiple frequency pairs. And something I can build more than one of. Sometimes my wife, who isn't a musician or a techie, wants to listen to a TV show while she works around the house. If I can duplicate the build for her, that would be great.

Then, eventually, the reverse: The transmitter in the bodypack and a base station receiver, for wireless instrument systems. I say "eventually" because now you're getting into the need for a super-strong, super-low-noise RF path because you're going into amplifiers and such.

This is all semi-hobby, want-not-need stuff, I think I know enough about electronics to put this together, whether or not any traditionalists or "old-timers" think it's cheating by stringing ICs together. The issue is primarily I can't seem to wade through all the RF data-link chips, mono circuits from the 80s, and SMT-only options to find something that will actually do the job. I don't think that Shure, Sennheiser, and all the pro audio wireless companies design all custom chips for their products, it doesn't make financial or business sense. For any $700 wireless system commercially available, I think there must be a $5 IC or a couple of $5 ICs that synthesize the carrier frequency, modulate an audio signal onto it, and amplify it to 40mW or whatever, so that these companies can order them by the thousands and pick-and-place them, and that these chips are available and purchasable and I would be able to wire them together to send an audio signal that hasn't changed much since Armstrong deveoped it in the 1930s.

And it just breaks my brain that I can't find one. When we all have quad-core supercomputers in our cellphones. Not one. That doesn't make sense to me. And that's where I'm hung up. That's definitely step one of the solution.

There is a company called "Schmartboard" that makes breakout/adapter boards with a very unique system for easily soldering QFNs. The boards are $10/ea so it's a little anxiety-inducing to know I'll probably ruin three or four before I get it right. The TI PurePath chips (I keep mentioning them because it's the only thing I've found that even remotely fits the bill) need a separate ADC and have an optional "range extender" (read: RF amp) so we're talking at least 6 ICs, six properly-soldered Schmartboards, and various components to have a prototype system, so it's not exactly the easiest way to go initially. Plus, if it does work, it'll be hell getting it all into a bodypack, as each Schmartboard is like 2" square and there'd be at least three in the bodypack plus a mainboard. I can't imagine that there isn't a one-chip solution for analog FM MPX on UHF that would tide me over for a year or more.

There's part of me that would love to open-source a project like this once I have a prototype, so that the next person like me that wants to do something like this doesn't have to go through all this crap. But that's way out in the future.

Now I think I understand what you want to do and why latency is a problem. (Also, the latency between when you post and when it appears has been fixed.)

Regarding if what you are discussing "fits" here I believe it does. Many of us are interested in getting signals from here to there which may not be "broadcasting" per se, for example, studio to transmitter links.

Let's see if I can summarize this correctly:

-The device is to take instrumental music and send it with no delay to a receiver/headphone arrangement.

-The wireless signal preferably should not be heard by casual listeners in the broadcast band and should have low noise.

-The range need not be great...enough to work in a house.

-Cost is a factor.

-If a build your own approach is taken it needs to be achievable with modest skills and tools.

Assuming I am close to or on target with the above we can start by firming some technical parameters. It is not hard to build a simple transmitter from discreet parts. The hard part is to build a receiver front end and I would not even try this myself. If you accept this then the starting point would be to find a receiver module which would give the required bandwidth and with some simple analog amplification would produce the output you need.

There are a number of other technical things which would need to be handled but before jumping into that let me know if I am on the right track. There may well be some of us here who know of an "off the shelf" solution but before throwing these at you we should further define the problem.

Neil

As a fan of analog radio and TV the wishes expressed by Epiglottis bring me to realize a simple thing I had not consciously thought about.... analog audio is the ONLY way to have instant real-time transmission and reception of a signal.

It appears that ALL digital methods have latency, and musicianship can be confused even by the analog reverb/echo on a large concert stage, I doubt whether digital technology can ever provide instantaneous response.

For a self brew transmitter you might think about the Big Talker shortwave transmitter designed by a group of us here at part15(dot)us, including Neil Radio8Z.

http://www.kdxradio.com/bigtalker.html

The schematics shown are for 13.560 MHz - FCC Part 15.225, amplitude modulation, and can be heard on most shortwave receivers. Caveat: not all shortwave receivers cover this band.

I'm estimating that the transmitter could be re-configured to operate under FCC 15.235, 49.82-49.90 MHz using frequency modulation, but finding a suitable receiver might be a challenge.

One of our part 15 people built a transmitter for this higher frequency and found a baby monitor that used the same channels, which he rigged as a receiver to use as a studio-transmitter link to deliver audio to his AM transmitter located some distance away.

Thank you Neil. You've summed it up perfectly. My Si4713, which is Arduino-controlled, has a terribly frustrating ~5200Hz whine on a silent signal, plus I can hear pips or ticks in the background, once for every time my Arduino polls the transmitter breakout - there's a little display that I use to see frequency, audio level, etc. Even if I take off the display I often need it to dump metrics to a debug console; if so I hear it tick every time it prints a new value. Now, it works if you run the audio up loud enough that your ears don't hear the background noises, but in quiet material and spoken material, it's agonizing. I've rebuilt the circuit a number of times, used shielded wire for the data and audio wires (the data wires are essentially producing 3.3v square waves on their lines), covered the inside of the box with copper sheet, and sometimes it's reduced, but not eliminated. The other strange thing is it drops off a bit when I'm not in the same room as the transmitter. Anyway, thank goodness it kinda works as a stopgap while we figure the rest of this out. But I've heard commercial FMs lose their STL and their carrier was super clean. Considering those transmitters are usually 25 years old or more, you'd think we'd have made some advances in the smaller ones, too.

I have never wound an inductor coil in my life so the concept of building something totally discrete scares me. I was looking for ICs because it's (supposed to be) a quick way to get a reliable experience out of a circuit that was designed by professionals and EEs. Other thing is, unless you've got a great circuit for an FM encoder (L+R, 19kHz pilot, L-R on DSBSC), we'll have to build two transmitters and two receivers into each unit, one for left and right channels. I thought I mentioned that. Stereo is important.

I can't see AM ever being an option in any band. Having grown up listening to AM and shortwave, I love the asthetic of it, the noises and the crackles and pops but it makes for a terrible signal path when you're trying to replace wires.

I see how a receiver front-end could be a problem, but I don't know enough to know why, not that it matters. Have I stumbled into an area where it's simple to broadcast in 900MHz but challenging to receive? Again, there must be a chip somewhere that does most if not all of it.  I assume it's not as simple as grafting an FM decoder/demuxer onto the output of a detector listening to the transmit frequency?

Seriously, you're spot-on, except that I do need stereo, and every pro-audio "wireless" system I can find specs on claims their "modulation method" is "MPX FM," (but on 400, 500, 600, or 900MHz bands) and claim the receiver frequency response is 35-15kHz. We know that bandwidth, don't we?!

So yes, you got the idea. I'm looking forward to any ideas you might have. I've been looking into all the CAD programs I'd have to learn and all the datasheets I'd have to learn to interpret to even have a snowball's chance of designing a board that will take SMT chips and components, and it seems to get more complicated the deeper I go. All the CAD programs are pre-stuffed with models of NAND gates and shift registers; but when you're looking at this sort of stuff they've got nothing. And it scares me, frankly, that I'd have to best-guess where the solder pads go, and the spend the money to get the board made and assembled someplace, to only know it didn't work after I'd paid the money and got the box in the mail. Not a very good way to work. But I can solder all day long. There's nothing big enough for me to solder anymore.

Anyway, over to you. Ready and willing to learn.

ej

Carl, you're certainly right on the latency. Even in the mid-2000s, audio drivers on Windows for the built-in sound chip had completely unusable latency, even though the computer itself was more than fast enough to do the processing, it got hung up in the audio hardware. I bought my first MacBook Pro in 2008, took it right over to my MIDI controller, plugged it into the USB port, fired up GarageBand, and I was stunned when I could play through, using the Mac as a sound generator, with near-zero latency. Now, in both situations, you buy a pro-audio interface that guarantees the latency is low and stable; but on Windows they have one kind of driver for built-in sound and one kind of driver for pro audio gear. On the Mac it's all CoreAudio.

Even hardware (which is just microcontrollers and DSPs these days) has latency. It might be 5-10ms, but it's there, and you've got to add up all the latency in your situation. And live on a big stage, I saw a figure once that for every meter you are away from a sound source like your guitar amplifier, or the drummer, it adds 10ms of latency. I am probably wrong on those numbers but close. Since audio moves at near light speed in cabling but not in air, you get back some latency running monitoring systems. But if those monitoring systems are digital, that's another 10-15ms you have to worry about as the audio takes the path through an ADC, a transmission protocol that has the ability to re-send dropped packets, and a DAC at the other end. Most of my effects gear and keyboard sounds are computer-based these days, but you can get the latency down to 20-25ms or less, and at least on the Apple side the CoreAudio module prioritizes itself pretty high in the OS, so random multitasking things happening in the background are not allowed to slow it down. But once you get into the 30-50ms range, that starts to become an effect, called slapback echo (well, it technically is comb filtering at that point, and slapback above 50).

I'll tell you a weird one. In the past I've performed out with both my instruments and vocals were run through live-performance software that could handle realtime effects, level switching, mixing, backing tracks, etc. And whatever latency that was in the round trip, from the inputs to my in-ears (wired at the time) was low enough that I didn't notice it while playing guitar, but it was just high enough that, with isolation earbuds in, the sound of my voice resonating in my head was comb filtering with - and *cancelling* to a degree - the voice that was coming back into my in-ears. I kept turning up my vocal in the monitor mix, not understanding, and another guy playing with me listening to the same mix, says "why is your voice 10x louder than everything else?" and I said, "it's not, I can barely hear it!" I'd never run into anything like that before.

Anyway, yeah, latency should always be in your mind whenever you're working with a digital signal path. Calculate and add the latencies at each new piece of gear where ADC-> DSP-> DAC. In a computer based situation, be aware of how many realitime plug-ins you're running, every time you give the CPU more to do to a live audio feed, the longer it's going to take and your latency can get out of control. When you're just mixing or playing back, most software compensates for latency. But even if I'm recording a vocal with effects on the lane, I have to mute the lane and just listen to the mixer channel, so I get an immediate copy of my voice in my ears. Otherwise it sounds off, dull, and strange, and I can't do any sort of articulation or nuance with it. And over Bluetooth? Forget it.

FCC15.237 Might be a very useful place to look.

Operation in the bands 72.0-73.0 MHz, 74.6-74.8 MHz, and 75.2-76.0 MHz.

(a)  The intentional radiator shall be restricted to use as an auditory assistance device.

(b)  Emissions from the intentional radiator shall be confined within a band 200 kHz wide centered on the operating frequency.

(c)  The field strength of any emissions within the permitted 200 kHz band shall not exceed 80 millivolts/meter at 3 meters.

It seems to me that there's enough leeway in those specs to build a working device for your use.

Or, here is one supplier of equipment for such an application which you might enjoy knowing about even if it's too expensive:

http://comtek.com/bst-25/

The law seems to say that these frequencies are only for the transmission of speech to the handicapped. Doesn't sound good. Plus the maximum field strength is 1/3 what it is in FM, so we're going the wrong direction there. Also, I have a need for stereo in this project, and any professional auditory assistance device is likely to be mono only.

FCC 15.237 is restricted to "auditory assitance device" for any application where any situation might benefit from auditory asistance. This does not mean it's only meant for handicapped persons.

Because there are several frequencies available for this, you have two choices to achieve stereo...

First, you could use two discreet transmitters for left and right;

Or, second, you could employ the multiplex method exactly the same as an FM station because of the 200 kHz bandwidth allowance.

Also, the POWER output at 15.237 is better than for 15.239 because millivolts are larger increments than microvolts.

The transmitter linked operates per a different section of the rules, Part 74, and has an output of 50 mW, which is magnitudes higher than allowed for the FM band at Part 15.

You're right. I had read the wrong thing, part of a request by some company to add verbiage to the rule, and in there it said handicapped. I was also confusing it with 76-88MHz, which is available on most IC FM transmitters meant for worldwide sale because the Japan FM band is 76-90MHz. Which brings 15.231 into question, because 231a says you can't use the band for ANY continuous or non-data transmissions, but then 231e says you can use the band for "periodic" transmissions that are disallowed by 231a if you keep the field strength below 500uV/meter @ 3m, which doesn't make much sense, except that a quick flip of the "World FM" register on the Si4703 verifies nothing within the range of my reciever is going on at all, periodic or otherwise, from 76-88MHz. Sounds like a minefield.

I do see now that the auditory assistance channels are 80,000uV/m @ 3m, and that's awesome. But we're going lower in frequency which means longer antennas - though I imagine at 80mV I wouldn't need an antenna for half a mile.

But, now I'm extra frustrated, beacuse the intentional radiator I have goes right to the edge of the highest auditory assistance band and stops. No 75.9MHz, in other words. Which brings me back to my original question, and again, what I'd really like is parts, integrated circuits, preferably something hand-solderable, before we run out of time to even acquire new old stock. If I was going to buy something I'd buy a pro-audio IEM system (there are honestly decent off-brand ones for $400, but the good name-brand ones are more than twice that). So I'm really not looking for a retail solution, because I'll go there instead of a real niche-market thing.

Again, I'm really looking for assistance in finding a modern component that will radiate on a frequency at a decent strength - I'll try 75.2-76 if I can find something that will radiate on that frequency, but I want to try the 900MHz ISM band as well - and a modern FM stereo encoder to stick in the circuit in front of it. I've got LiPo batteries, bodypacks, discretes, and so on to build the actual units, I just can't seem to find anything short of a voodoo-laden discrete circuit from the 1980s with a bunch of delicate aircore inductors that will radiate a continuous carrier on ANY frequency, besides the little chips that Taiwan uses to flood the world with those things that put your MP3 player over your car stereo and they sound horrific.

So I'm grateful that I was set straight, unintentionally so, on my misinterpretation of 76-88MHz (though like I said, severe clear out here in the country, and you sort of have to interere with another signal to get busted), and it's nice to know that I can go to 80,000uV/m @ 3 in the "auditory assistance" bands, though I imagine someone at the FCC has the clout to send me away on the notion that I oughta know that doesn't mean personal property transmitters for someone with most of his hearing left. I'm not a big fan of operating at the edges of ambiguously-written laws, since I'm pretty sure in this day and age Part 15 amounts to table scraps for us common folk as they make their money and maintain their power dealing with wealthy corporate types and fining Opie and Anthony. Maybe I'm a cynic, but I think the FCC sees unlicensed low-power broadcasting as "we don't have the manpower to police every 10uV/m spurious emission, so we're going to look the other way as long as no one complains," not as "hey, America, your taxes pay our salaries, the airwaves are yours, go to town!"

Anyway, I'm getting off topic here. Having a legitimate, palpable reason, sitting on my workbench, for researching which frequencies I can and can't transmit on is a problem I want to have, and don't have yet. From a component-level perspective (ie no kits or finished products at the retail level), how do I get started?

Thanks much for all the great information.

When I did videorecording as a profession I experimented with wireless mics and learned something that could be useful now.

A comparison was made between a $1,000 Lectrosonic Professional System and a budget Sony WCS-990. I was able to get them performing alike with no observable difference. How did I do it?

At first the Sony was disappointing because the mic included with the package sounded like a bad telephone call.

So I replaced it with my own mini-microphone and it sounded identical to the Lectrosonic.

The range was exactly the same between both of them.

The Sony comes switchable between 2-channels, in the range 912-916 MHz.

So I got a second one and was able to run two Sony systems, each with its own channel and separate receiver, ideal for stereo.

Today MCM Electronics has the Sony, now upgraded to WCS-999

http://electronics.mcmelectronics.com/search?cataf=&view=list&w=Sony+Wireless+Microphone&x=19&y=9

Carl,

I appreciate all the great information and advice you've given. I'm afraid though that you're not hearing what I'm asking. The Sony unit; can you crack it open for me and give me the numbers on all the ICs? Chances are there's a $4 IC in that thing that synthesizes a carrier frequency in the 912-916MHz band, and perhaps another $4 IC that modulates that carrier. I bet you they're not custom or proprietary Sony chips, I bet you they're made by a semiconductor company and they are or were available through retailers like Digikey. I also bet you that the circuit in the Sony unit is not very different from one of the circuits in the chip's datasheet. And everything else is op-amps and discretes, and potentially a number of sub-circuits I don't need (like AGC), or stuff I'd want to do differently (like multiplex stereo, or a rechargable battery instead of an AA or 9V cavity). I want to buy those chips, learn how to use them, and build my own device. Period, end of sentence. I haven't been able to find those chips, or similar ones; I don't know what they're called, I don't know how they're classified, I don't know if I'm looking for a complete system on a chip or an oscillator or a VCO or a synthesizer or if I need a separate chip to do PLL or I need an external crystal or what the frequency of the crystal might be. Where's the prescaler? Does it come in the oscillator chip or do I build one? Optimally I'd love one-chip solutions, as it seems to still take a year or more to find one and wire it up - trust me, it would be cheating at this point to be able to do it with the crap in my parts drawer - but I am looking to understand how these devices work inside, and use it to roll my own; and other than, for example, an all-in-one IC on a breakout board, I am NOT interested in buying any commercially available unit that isn't a high-quality, well-reviewed, pro audio grade purpose-built IEM system. I am not interested in hacking a so-so cheap device so it works better (trust me, I have done enough of that for ten lifetimes). My goal is to learn how to do it and built it myself, and if I can't I will wait until I can afford the right thing.

I hope that we can talk more about that.

ej

ej,

When I commented that I would not try to design a receiver here's why. A receiver must work with very small signals and select them from the rest of the spectrum and amplify up to usable levels. Lots of tricky engineering.

I found two modules which you can consider (linked below). I have used 900 MHz modules from this company and they worked as expected. At least here there was no interference (sub-urban environment).

These work with analog up to 28 kHz which is good but the bad news is this will be a mono channel. Perhaps use a 900 MHz for one channel and a 869 MHz for another. I am not sure of the regulations for 869 MHz and this is very close to the public service band so caution is advised.

These modules use castellated half vias for connection and at first this seems daunting but these are on .1" centers which is large compared to most. Some builders report that this size can be soldered to an IC socket with machined pins and used as a dip part but I would prefer to simply solder some wires to only the half vias I need to use. This would make connections pretty easy.

Costwise, each is about $13 and $17, so not too bad.

Links:

http://www.mouser.com/ProductDetail/Linx-Technologies/TXM-916-ES/?qs=K5ta8V%252bWhtbMJKFukW4tHQ==

http://www.mouser.com/ProductDetail/Linx-Technologies/RXM-916-ES/?qs=K5ta8V%252bWhtavrWQcR/nf/g==

Click on the datasheet button in the links for technical stuff.

Neil

FCC 15.249 seems to be the section needed for this use...

Operation within the bands of 902-928 MHz

(a)  The field strength of emissions from intentional radiators operated in this frequency band shall comply with the following: Field strength of fundamental 50 millivolts/meter; Field strength of harmonics 500 microvolts/meter.

(b)  Field strength limits are specified at a distance of 3 meters.

Further detail is found in the rule.