Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

FCC Part 15/ISM DIY radios for full duplex microcontroller coms?

Status
Not open for further replies.

()blivion

Active Member
Hi.

Looking to build my own unlicensed radio projects. Sadly, I suspect there is a chance this idea will go over like a fart in church, you radioaficionados don't seem to take kindly to those trying to do anything without a licence, whether there are legal and good intentions or otherwise. But I have to ask the question as I have been trying to get into the science for just about as long as I have been doing electronics, and I just can't seem to get anything concrete. Finally, I would rather learn from people who know about RF interference than blindly build some random schematic I found on google that may or may not be a pollution nightmare.

To that end...

Could someone share whatever simple, legal, moral, and overall worthwhile schematics/projects they have, (or know of) for doing full duplex ISM and/or Part 15 compliant RF coms between microcontrollers? Or possibly help me design something from the ground up?

What I'm mostly after in the RF section is...
1) Frugal construction, using cheap, common, discrete parts and techniques.
2) Design immunity from causing realistically problematic RF pollution.
3) Flexibil vectors for modulation, changing carriers, and what have you using the MCU.

Know that...
1) I am fully aware of the prebuilt cheap modules. I want to build it out of parts.
2) My power and range is of course very low. A few feet/meters, no more than ten.
3) I prefer using GP BJT's. No reason in particular other than that I have a bunch.
4) I have a more or less controlled environment, and a reasonable head on my shoulders.
5) I DO NOT have a lot of radio equipment, I'm a low budget hobbyist. (I of course have basic electronic equipment.)
6) I live in a rural area, I highly doubt there is strong chance of me causing significant RF interference unless I really tried.

Here is an example of something that is about what I am looking at for transmitting. https://www.swharden.com/blog/2011-08-06-ridiculously-simple-avr-mcu-am-radio-transmitter/ I wouldn't mind building something a little more complicated. Maybe some provisions for FM modulation. I tried to build this already and it didn't even have enough power to transmit a few inches. I wasn't sure if it was radiating more power on a band that was out of range of my receiver, so I was too afraid to up the power without knowing what exactly was happening.

I like the idea of a regenerative design for the receiver portion. The SNR of superhet is cool and all, but I think I like the absolute gain AND simplicity of the regen more. I just don't know how I should proceed if I also want a BJT built MCU interfaceable design.

Let me know if you have something good to share. Thanks.
 
Attached are a few RF oscillators and some filter circuits I have been playing with in the simulator. Note that these are just me doing some tinkering in this general area, and do not necessarily represent a particular direction that I am planning on taking. Edit: Also, they are not necessarily finished and working designs.
 

Attachments

  • 1Mhz LC Osc.asc
    1.6 KB · Views: 326
  • 1M LC filter.asc
    2.9 KB · Views: 302
  • Passive Ultrasonic Bandpass filter.asc
    895 bytes · Views: 329
  • Ultrasonic Bandpass filter.asc
    3 KB · Views: 316
  • RF Osc 1.asc
    1.4 KB · Views: 339
  • RF Osc 2.asc
    1.9 KB · Views: 316
  • VCO.asc
    3.1 KB · Views: 342
  • LC parallel Osc.asc
    2.6 KB · Views: 310
OK, I will just wade in and answer a fer points as best I can.

Looking to build my own unlicensed radio projects. Sadly, I suspect there is a chance this idea will go over like a fart in church, you radioaficionados don't seem to take kindly to those trying to do anything without a licence, whether there are legal and good intentions or otherwise. But I have to ask the question as I have been trying to get into the science for just about as long as I have been doing electronics, and I just can't seem to get anything concrete. Finally, I would rather learn from people who know about RF interference than blindly build some random schematic I found on google that may or may not be a pollution nightmare.
I do not go to church, but I have been known to fart quite a bit after eating beans etc!

Could someone share whatever simple, legal, moral, and overall worthwhile schematics/projects they have, (or know of) for doing full duplex ISM and/or Part 15 compliant RF coms between microcontrollers? Or possibly help me design something from the ground up?
I assume that you are in the USA, otherwise "FCC Part 15" has no relevance to you.
Have you read the Part15 documents to see what is and is not allowed?

What I'm mostly after in the RF section is...
1) ....
2) Design immunity from causing realistically problematic RF pollution.
3) ....
The most likely problems are:
Using an inappropriate frequency.
Inadequate filtering of hamonics, resulting in signals at 2, 3, 4, 5, 6, etc times the wanted frequency.

Know that...
1) ...
2) My power and range is of course very low. A few feet/meters, no more than ten.
3) ...
4) ...
5) I DO NOT have a lot of radio equipment, I'm a low budget hobbyist. (I of course have basic electronic equipment.)
6) I live in a rural area, I highly doubt there is strong chance of me causing significant RF interference unless I really tried.
A few milliwatts can go a LONG way.
Without test equipment you are stuffed! A wide range general coverage receiver can go a long way to help.
Rural area, little chance... maybe, maybe not!

Here is an example of something that is about what I am looking at for transmitting. https://www.swharden.com/blog/2011-08-06-ridiculously-simple-avr-mcu-am-radio-transmitter/ I wouldn't mind building something a little more complicated. Maybe some provisions for FM modulation. I tried to build this already and it didn't even have enough power to transmit a few inches. I wasn't sure if it was radiating more power on a band that was out of range of my receiver, so I was too afraid to up the power without knowing what exactly was happening.
If you could not hear that from a few inches away, either:
It was not working.
Your receiver is not working.
You were listening on the wrong frequency.

I like the idea of a regenerative design for the receiver portion. The SNR of superhet is cool and all, but I think I like the absolute gain AND simplicity of the regen more. I just don't know how I should proceed if I also want a BJT built MCU interfaceable design.
Let me know if you have something good to share. Thanks.

The various types of regenerative receiver are ok, but they do have limitations.


JimB

PS, if you want to post a circuit schematic for comment, please consider those of us who do not play with simulators and post a simple picture (eg JPG) not a simulator config file.
 
OK, I will just wade in and answer a fer points as best I can.
Much thanks :)

I do not go to church, but I have been known to fart quite a bit after eating beans etc!
Excellent, keep up the good work, church is probably worse for you than beans anyway.

I assume that you are in the USA, otherwise "FCC Part 15" has no relevance to you.
Yes, USA. And I apologise on behalf of my country. Were not all terrible people. :)

Have you read the Part15 documents to see what is and is not allowed?
From what I have seen, under 100mW is the max power for unlicenced but legal use of otherwise licence only space. There are still some critical areas that are absolute no fly zones, aeronautics(no pun intended), military, fire, medical, police, etc etc, very strict, but I will be staying well away from there. Plenty of radio I could potentially use that can't possibly kill people if something goes terribly wrong.

ISM bands may be another story, very laxed restrictions it would seem. I think, (and this is pure conjecture), the ISM bands exist because there are some devices that can't help but spew massive radio noise. So they allocate the ISM bands for those devices, to sort of sweep it under the rug. So there is less reason to limit power emitted, impossible task when it's already there. It follows from this, because you can't expect using those bands to work well, to do so is "at your own risk". It appears there are still power limits for some reason. I don't really understand why, but not knowing why a rule exists is no excuse to break it.

The most likely problems are:
1) Using an inappropriate frequency.
2) Inadequate filtering of harmonics, resulting in signals at 2, 3, 4, 5, 6, etc times the wanted frequency.
True enough.
I'm not so worried about case 1, there are lots of bands that are perfectly acceptable to use regardless of occupancy, provided your power is below the legal limit. And of course I will listen to a band before using it anyway. I'm most worried about case 2, which is why I killed that little transmitter project when I did... I'll get back to this.

A few milliwatts can go a LONG way.
I know, I love this...https://en.wikipedia.org/wiki/QRP_operation, Of course, this is an extreme example. In reality, we are more concerned "practical interference". In practice, some of the electronics projects I have already done involving high power PWM, or medium-high power + high voltage Edit: [had potential to] radiate many watts of power into free space. They didn't cause interference because the energy was concentrated on frequencies I knew would not radiate with the size of the wiring, and/or were properly shielded. Edit: Well... The FCC hasn't come knocking on my door yet.

Without test equipment you are stuffed! A wide range general coverage receiver can go a long way to help.
I'm not opposed to building basic test equipment. My big hurdle is not having the ability to examine/measure/determine RF frequencies. I have a spectrum analyzer/o-scope like "thing" that's good for frequencies up to 96khz, maybe 192khz. I also have a AM/FM radio of decent sensitivity... sort of :)/)... it could use an antenna, but I have no clue where to put one. Many many connections to the tuning tank, not sure where the RF is actually coming in.

I was more hoping for proven/foolproof schematics that someone had gone through all the trouble and made working already. And hoping to have them give advice on using the parts that I already have to make/modify it. My biggest problem so far has been, all the radio-esq projects out there need very specific RF parts by the numbers. I'm fairly sure though that a more generic design could be made, or an existing design modified for general purpose parts.

Rural area, little chance... maybe, maybe not!
The concept of "rural area" is debatable when talking radio, but we can operate under the assumption that I live next to, medical, police, fire, military, and aeronautical establishments if it helps.

If you could not hear that from a few inches away, either:
It was not working.
Your receiver is not working.
You were listening on the wrong frequency.
Sorry, that was a typo on my part.

"it didn't even have enough power to transmit [more than] a few inches."

And by a few inches I mean, from my elbow to my wrist was about the realistic limit, and the SNR was terrible. But it did work as advertised... well... I used a BJT amp, not a JFET, and it was a PIC programed in my own ASM code... so it was not exactly the same project. I just remade it in spirit, same principles. Clock out is carrier, GPIO pin has audio, modulates gain of amp. Fairly straightforward contraption, just too shaky for my taste.

Speaking of shaky, back on the harmonics issue, the text for that project/idea made specific mention of harmonics, do to the frequencies being derived from square waves. I found that the transmitter project that I made was located on the upper end of what my little AM/FM radio could receive, and I was seeing it multiple times in that area! This gave me the instant impression that not only was I transmitting a good deal of harmonics, but It was also possible that the fundamental was above the max range of my reciver. With aviation bands directly above the AM band, I wisely decided that it would be best not to assume that the range issue was a result of not radiating enough power.

The various types of regenerative receiver are ok, but they do have limitations.
I'm told the superhet is superior in pretty much every practical way, but I am fairly certain that it is far more complex and not quite as sensitive, which can't be anything but a downside. Keep in mind we are also talking digital use here, if that changes things. I would like to think it being digital is one of quite a few relevant circumstantial reasons to prefer a regen over a superhet.

PS, if you want to post a circuit schematic for comment, please consider those of us who do not play with simulators and post a simple picture (eg JPG) not a simulator config file.
Sorry, my bad. I also used to avoid LTspice not that long ago. Then I tried to get along with it and found out that it is (A) small, (B) free (C) not hard to use, (D) accurate (enough) (E) very informative (F) a bunch of other stuff I'm probably forgetting. Don't get me wrong, LTspice, or any SPICE program really, is not the be-all-end-all of EEing, it has weaknesses. But when used in conjunction with other tool and testing, it's a life saver. Really speeds things up and can track down issues. I would highly recommend you use it if you don't but have the power to.

In any case, I put up the schematic in PNG format for you. Just know that some of the little important bits will be missing. You will only be getting 70~90% of the picture... per picture. Edit: The top right hand one is particularly uninformative without seeing it in the sim.

1M LC filters.png1M LC Osc.pngLC Par Osc.pngPassive (Ultrasonic) Bandpass.png
RF Osc 1.pngRF Osc 2.pngUltrasonic Bandpass filter.pngVCO.png
 
Last edited:
It is all about what frequency, antennas, receivers, modulation methods. Transmitters are the easiest part.
 
Without test equipment you are stuffed! A wide range general coverage receiver can go a long way to help.
I'm not opposed to building basic test equipment. My big hurdle is not having the ability to examine/measure/determine RF frequencies. I have a spectrum analyzer/o-scope like "thing" that's good for frequencies up to 96khz, maybe 192khz. I also have a AM/FM radio of decent sensitivity... sort of :)/)... it could use an antenna, but I have no clue where to put one. Many many connections to the tuning tank, not sure where the RF is actually coming in.
The point which I am trying to make is that without suitable test equipment, it can be difficult to get even so called "reliable" circuits working correctly.
I guess that your scope/spectrum analyser is a PC sound card device, fine for audio but just does not cut the custard for RF.
This is why I suggest a "general coverage receiver".
OK we are talking $$$ here which you may not have available to spend on a hobby, but if you want to play with radio, at some point you need a receiver.


Speaking of shaky, back on the harmonics issue, the text for that project/idea made specific mention of harmonics, do to the frequencies being derived from square waves. I found that the transmitter project that I made was located on the upper end of what my little AM/FM radio could receive, and I was seeing it multiple times in that area!
If the transmitter project was working, I would expect that you would hear it best at around 1000kHz on the AM band.
Multiple responses could also be due to a strong signal from your transmitter and poor performance from your receiver.

This gave me the instant impression that not only was I transmitting a good deal of harmonics, but It was also possible that the fundamental was above the max range of my reciver. With aviation bands directly above the AM band, I wisely decided that it would be best not to assume that the range issue was a result of not radiating enough power.
The Aeronautical bands are not "directly above the AM band".
There are various bands in the short wave region which are used by aircraft on trans-ocean routes. You would have a hard job to upset those with your little transmitter.
For more local communications, aircraft use VHF from about 110 to 135 Mhz.

I think that many people do not realise just how wide the radio spectrum actually is.

As for your simulator circuits, I have only had a brief look as I have been rather occupied with "work stuff" today.

JimB
 
MikeMl said:
It is all about what frequency, antennas, receivers, modulation methods. Transmitters are the easiest part.
Hi Mike

Not that I object too much, transmitters certainly are constructively easiest, but they have the distinct problem that you can't tell they are working properly without a receiver or special tools or calculations, I think that's my biggest problem anyway. I'll probably have to build some tools first, before transmitters do become easy.

JimB said:
The point which I am trying to make is that without suitable test equipment, it can be difficult to get even so called "reliable" circuits working correctly.
Of course, and this point was not entirely lost on me. But the need for specific RF tools should drastically decrease the more one is using proven designs. Case in point, if I just used an internet design, part for part, and the design was known to work, and I built it correctly, I wouldn't need more than a soldering iron, the parts, and the time.

JimB said:
I guess that your scope/spectrum analyser is a PC sound card device, fine for audio but just does not cut the custard for RF.
Yeah, I can't afford a real scope. I made this one, and it works wonderfully well considering. I was thinking that if I had some kind of adjustable RF source of significant precision, I could heterodyne into the audio range and sort of kludge together a RF Scope. This adds some interesting challenges though.

JimB said:
This is why I suggest a "general coverage receiver".
OK we are talking $$$ here which you may not have available to spend on a hobby, but if you want to play with radio, at some point you need a receiver.
Yeah, budget is pathetic. If I have the parts, it's fine. But as soon as I have to make orders... It's pretty much a dead project to me. I have to be in serious dire need with no way out before I resort to making purchases. To put it in perspective, there are lots of equally interesting projects I have that cost me nothing. I think the goals here can be achieved just as freely and easily though.

I do have a receiver, and it does work, picks up stations I didn't know existed. And although the receiver is not perfect, a design should exist that will precisely hit the right spot on the dial if it is built correctly. Also simulators can shoehorn that in as well.

There is also building a receiver. The only problem here is that it will be marginally difficult to know exactly what frequency I am seeing/hearing. I would have to measure the resonant frequency of the tank at a few specific points, then interpolate beyond my ability to measure frequency. It's not 100% accurate, but it's better than trusting the dial on my current radio.

JimB said:
If the transmitter project was working, I would expect that you would hear it best at around 1000kHz on the AM band.
The problem with this expectation is the accuracy of hitting the selected carrier depends on the unknown Q of the internal RC oscillator. Claims of Q are around %1, but this is not talking about creep/jitter, just how close the average frequency is to the intended. So real Q could be way worse.

More importantly, the fact is the internal calibration value was lost, and I have no metric for knowing how far this calibration can pull the internal oscillator. A single step could be a significant deviation in frequency or marginal. And hitting the max/min of the scale could make the effective output frequency ±2Mhz, or only ±1Khz... who knows really.

JimB said:
Multiple responses could also be due to a strong signal from your transmitter and poor performance from your receiver.
Could be, I was getting responses more and more faintly as I moved farther from the top, and they seemed to be exponential-esq. This fits fairly well with a harmonic series, but that doesn't mean there is not another explanation.

As for receiver, it works surprisingly well for being a POS clock-radio. It's just that it has a limited range, only covering the AM/FM bands, which of course, is to be expected. It's selectivity and sensitivity are great though. Of course, a professional receiver would destroy it hands down. But at the cost of a few hundred of these AM/FM radios, I've no choice but to stick with this for now.

JimB said:
The Aeronautical bands are not "directly above the AM band".
There are various bands in the short wave region which are used by aircraft on trans-ocean routes. You would have a hard job to upset those with your little transmitter.
For more local communications, aircraft use VHF from about 110 to 135 Mhz.
Ah yes, thats what it was. Above the FM bands is what I was thinking. I got my AM and FM broadcasting bands mixed up. There are some danger zones above the AM band, but as you said, not so easy to interfere with.

JimB said:
I think that many people do not realise just how wide the radio spectrum actually is.
I like this as a good reference. Also gives you a decent idea of exactly who gets the biggest slices of the pie.
The_US_Radio_Spectrum.jpg

JimB said:
As for your simulator circuits, I have only had a brief look as I have been rather occupied with "work stuff" today.
Take your time, no rush. Work comes first of course.


Other thoughts
I'm probably going to make a Mark II of the before mentioned transmitter, and be a bit more thorough with the design. I'm guna want to add some harmonic filtering, and/or a proper tank and such. With a proper tank, I should be able to adjust the carrier around, to make sure I have actually found the fundamental when I hear it. And I won't need to μC to run at the carrier's frequency, which is a good thing.

When building the first MCU based transmitter, I didn't have a way to infer how much power was getting to/through the antenna. I need some formula or simulation that works with energy radiating systems. I know that the power radiated depends on how the antenna length compares to the carriers wavelength. And I know that at 1Mhz, the antenna would be a very small fraction of the wavelength. And I knew that meant there was not going to be much load from the antenna. But beyond that, without changing the carrier or the antenna length, I had no particularly decisive way how to figure (or adjust) how much power I was actually transmitting.

I think my problem here is impedance matching, as I keep forgetting about it/don't know where to start.

Moving beyond the above, I would ideally have a discrete VCO (or full PLL) building block that can be used for both my transmitters and receivers. This would make μC control of the radio powerful and trivial, which is going to be essential if I am to do some of the projects I have been thinking about. The problem is such things are more complicated, and I specifically want simple. Then again, everything becomes "simple" when you thoroughly understand it.

For a cheater PLL (with admittedly bad Q) I was thinking of just measuring the running frequency of a simple VCO with the μC, and comparing that to the μC's crystal every now and then. This would allow me to (roughly) program the output frequency of the VCO with the μC, and significantly reduce external parts count/complexity.
 
Last edited:
Of course, and this point was not entirely lost on me. But the need for specific RF tools should drastically decrease the more one is using proven designs. Case in point, if I just used an internet design, part for part, and the design was known to work, and I built it correctly, I wouldn't need more than a soldering iron, the parts, and the time.

Thats a seriously naieve view :( .... without test gear you would never know if its on the correct frequency, what the power output was, if it was producing masses of spurii and harmonics. Whether you build the circuit exactly to plan is almost irrelevent. A transmitter or receiver still needs to be tuned up once its built

One particular commercial 576MHz local oscillator I was working on some years back was diabolical ( horrific) in its operation.
The tiniest tweek of the trimmer capacitor took the output from a pretty clean carrier out with the first 2 harmonics 40dB or more down on the main carrier to a horrible picket fence of spurri from one of the spectrum to the other ..... well over a couple of GHz of the spectrum anyway.!!
The point is ... even with just a power meter and tuning for peak power output would not have alerted me to the bad quality of the oscillator's signal. ONLY a spectrum analyser allowed me to continuously monitor the output as I tuned the oscillator to get it to produce the cleanest signal possible and get rid of all the crud.

The second point is ..... once you move from AF circuit construction to RF construction decent test equip is ESSENTIAL else you can cause all sorts of hell for other users of the spectrum. A soldering iron and a multimeter just doesnt cut it. You should be looking at test equip first and circuit construction second
You shouldnt be building any RF circuit TX or RX without the ability to analyse it operation

regards
Dave
 
Sorry it took forever to reply. I lost internet con for a while and was otherwise also distracted.
Thats a seriously naieve view:( .... without test gear you would never know if its on the correct frequency, what the power output was, if it was producing masses of spurii and harmonics. Whether you build the circuit exactly to plan is almost irrelevent. A transmitter or receiver still needs to be tuned up once its built

I respectfully disagree, many RF circuits are mass produced with little or no tuning and testing and work perfectly fine. TX and RX.

There are RF circuit designs that can be made that can ONLY oscillate on a single frequency with an immutable max output power. Take a crystal osc circuit that is designed to operate on only the fundamental, and only has 100 milliwatt available input power. Other than absolutely absurd statistical anomalies in the realm of getting struck by lightning while winning the lottery, such a design has no chance of putting out RF on the wrong frequency that is over the 100 milliwatt legal limit.

Though I confess that this logic is somewhat of an impractical over exaggeration. The above circuit would probably be less than useful, and I clearly would require some kind of test equipment to make any not so immutable circuits.

My only point is there is a limit to what is reasonably required for such low tier grade of circuit. One can't expect it's reasonable for a competent dabbler to have to get a full RF clean room, professional spectrum analyzer, professional receiver, RF circuit design software, and whatever other spendy RF tools are out there, just for some ~100mw toy radios that will probably simply fail to work or self destruct if they have any sort of problems. Least likely problem of all would be to cause anyone else RFI problems.

Now, if by "proper tools" we are talking reasonably easy to build or obtain, or something that relies on some simple tricks/maths that one can easily use to infer the necessary info using other parameters, then I would clearly have no reason not to want to know/have these available to me.

I think JimB's suggestion of a "general coverage receiver" is probably the best example of a solid general purpose RF building tool, and probably more in the realm of what is practical. A receiver is a good allround tool to have for RF, and should be fairly easy to come by or build. Wheeler's formula and similar precision RF inductance formula are another reasonable general purpose RF tool compromise that could be employed, as opposed to having to use an expensive inductance meter that may be 5%~15% more accurate. Finally, if a resistor intended for RF tests perfectly on DC, you can expect that 99.99999999999% of the time it will be just as good at RF, without ever having to go near RF test equipment.


One particular commercial 576MHz local oscillator I was working on some years back was diabolical ( horrific) in its operation. The tiniest tweek of the trimmer capacitor took the output from a pretty clean carrier out with the first 2 harmonics 40dB or more down on the main carrier to a horrible picket fence of spurri from one of the spectrum to the other ..... well over a couple of GHz of the spectrum anyway.!!

The point is ... even with just a power meter and tuning for peak power output would not have alerted me to the bad quality of the oscillator's signal. ONLY a spectrum analyser allowed me to continuously monitor the output as I tuned the oscillator to get it to produce the cleanest signal possible and get rid of all the crud.

The second point is ..... once you move from AF circuit construction to RF construction decent test equip is ESSENTIAL else you can cause all sorts of hell for other users of the spectrum. A soldering iron and a multimeter just doesnt cut it. You should be looking at test equip first and circuit construction second

I understand what you and everyone else is saying, and your specific issue pretty much is an impossible situation if you don't have the exact right tool, but that all seems like a non issue to my quest.

Your particular case seems like a design inadequacy, almost certainly because it was a commercial interest. Generally, when a circuit is designed with robustness in mind above all else, it will either work right, or it will go failsafe when something is amiss. If the above LO was designed such that it would function, but could possibly give absolutely woeful output with very a minute trigger, it wasn't designed to the standard I would hold for a simple low parts belt and braces design. A good design will be intrinsically failsafe, or have extra parts and warnings to fill that role.

Now... you are probably thinking, you can't have your cake and have also eaten it. It has to either be simple, or robust, but can't be both. But I myself have found that more complicated systems actually tend to fail more often, or at least go into a half working intermittent failure mode more often. That this "simple is better" philosophy may be applicable to a RF circuit, I can't say for sure, but I can't see why it wouldn't. Fewer parts means fewer things to have to check. One can always swap suspect parts for new ones, and that is a lot easier with a low parts count simple design.


Final thoughts.
It's not as if there are no ways to make an RF designs that can't be built correctly without RF test equipment. And it's not like there are no cases where we can test a part at DC, and not be reasonably assured that it will also function the same at RF. If any of these were not true, then we wouldn't have anything RF in the first place.

Let's not forget that I am talking about very low range, low power, simple "toy" radios. This is not DX, I am not trying to reach across continents here. If I can get across the room, I will be happy. And to me, this low expectation and limited resource relaxes much of the requirements for professional test equipment, and does so enough to make such a circuit more than doable without stepping on anyones toes. It's really not super important that it work 100%.

But that's just how I continue to see it...

In the end though, I did come here for advice and opinion, it would be stupid to subsequently throw that away simply because I don't like or don't agree with what I am hearing. And since more than one person has claimed it's not possible without pro equipment, and no one has said otherwise, then I have no choice but to concede defeat on the matter, or take it into my own hands elsewhere.

As I said, if making the project possible comes down to having to buy really expensive equipment or similar, then it's utterly pointless. I am better off buying the $2 mass produced μC RF boards on Ebay and modifying them for my needs. Or simply moving to other projects I have going that won't cost me as much as a new car.

Anyway, if any of you happen to think of, or want to discuss, any way that can actually "achieve the impossible", post it and we will talk. And if I come up with something that works on my own, I'll let you know as well. Otherwise, I consider the issue closed.

Thanks for your input.
 
I have had time to do some playing with electronics today, and so developing on your oscillator circuit which you posted earlier, I created my own circuit and built it and tested it.

In the attachment, I have drawn the tuned circuit separated from the amplifier.

For a test, I intended that the oscillator should run at 1MHz.

In a book somewhere long ago, I found some "RULES OF THUMB" for a Colpitts type circuit.

The rules were to set the reactance of the capacitors, at the required frequency, so that:

X C parallel = 140 Ohm

X C coupling = 100 Ohm

X C feedback = 45 Ohm

And obviously the reactance of the Inductor should be the same as that of the Parallel capacitor.

Calculating these out gave:

C parallel = 299pF
C coupling = 1592pF
C feedback = 3537pF
Inductor = 22.28uH

These are obviously not standard values, so juggling around and using what I had on hand, I used:

C parallel = 330pF
C coupling = 1500pF
C feedback = 3000pF
which meant that the inductor should be 23.45uH

For the inductor, I used the standard formula for a single layer coil and calculated that I would need 24turns on a 1.23inch dia former in order to get 23.5uH.

The amplifier is probably not as good as it could be.
I set the DC conditions to give 1volt at the emitter and 3.5v at the collector.

On reflection, I could probably have done better by connecting the collector directly to the supply and increasing the emitter volts to around 2.5v.

Anyway, I built it and it worked.
Measuring the frequency at the emitter using a x10 divider probe to connect to the counter gave the frequency as 964 kHz, a bit lower than the design frequency of 1000kHz.

For a cross check, I measured the actual value of the capacitors used to build the circuit and found that two of them were quite a bit off.

Code:
Nominal            Measured
   Value                 Value
  330pF                      351pF
1500pf                    1611pF
3000pf                    2977pF
3000pF                   3007pF

Recalculating the frequency using these capacitance values and the calculated value of the inductor gave the frequency as 979kHz.
This to me is as near to the actual measured frequency as anyone has any right to expect from a simple test build.


So, coming back to your request for a "reliable" circuit.
What I have built is probably reliably reproducible in terms of frequency accuracy and ability to oscillate, but the waveform at the emitter is distorted when viewed on the 'scope, and has harmonics to 15Mhz or more when viewed on the spectrum analyser.

Having all the "big boys radio toys" certainly helps when playing with RF circuits, but for a low power, low frequency (?) thing which you have in mind it is possible to make things work, just don't connect it to a big long antenna and no one will know the difference.

JimB
 

Attachments

  • Colpitts.JPG
    Colpitts.JPG
    29.8 KB · Views: 573
  • Colpitts Oscillator 001.jpg
    Colpitts Oscillator 001.jpg
    326 KB · Views: 545
  • Colpitts Oscillator 003.jpg
    Colpitts Oscillator 003.jpg
    328.9 KB · Views: 435
  • Colpitts Oscillator 012.jpg
    Colpitts Oscillator 012.jpg
    297.3 KB · Views: 464
  • Colpitts Oscillator 018a.JPG
    Colpitts Oscillator 018a.JPG
    222.1 KB · Views: 463
Beautiful work JimB.

I'm a little disappointed that a pure LC resonant circuit has that much harmonics, is the amp causing it? I expect an overdriven amp could inject excess noise into the tank, causing it to go somewhat askew. Or is it just the tank is actually partly resonating on those frequencies of it's own will, do too having a low Q or something?

Do you have a better explanation for the harmonics? I don't get any in the simulator, but that doesn't say a whole lot.

Do you think another tuned tank after the amp might strip off the harmonics? Or can you think of another simple way to remove them?

I'm sure for across the room toy radio, that is more than acceptable quality. But if there is a simple way to make it better, there is no reason not to do it. Well, other than making it more complicated of course.


Again, most appreciated work sir. Very useful information.
 
I'm a little disappointed that a pure LC resonant circuit has that much harmonics, is the amp causing it?
Yes the distortion is caused by the amplifier.
I took my measurements from the emitter connection.
Connecting the probe to the base connection shows a much lower distortion level. This is to be expected as it is closer to the tuned circuit.
Using a coupling loop to take a sample from the coil itself shows an even lower level of distortion.

To reduce the distortion in the amplifier, the thing to do would be to reduce the gain so that the circuit was only just oscillating.
In a practical day to day circuit this could cause problems due to the oscillator not starting reliably under all conditions.

In a real transmitter, the harmonic output is suppressed by a low pass filter which is connected between the final amplifier and the antenna.
Have a look at the attachment. This is the output stage of a VHF Tx/Rx, a rather ancient type, but it shows the principle quite clearly.

JimB
 

Attachments

  • MX294 LPF.JPG
    MX294 LPF.JPG
    57.3 KB · Views: 447
Excellent, a low pass filter is easy enough to do. Though the one depicted in your attached image is a little bulky for my tastes. Is it entirely necessary to have so many stages for our simple circuit? Is all that just to get the sharpest roll-off possible? Would just one stage not suffice? It at least should significantly reduce the harmonics, right? Maybe there is a greater probability some of the low pass components will begin to resonate at one of the harmonics if it does not have enough stages. But glancing at the circuit, I would just expect that it was only for getting a sharper roll-off.


As for the amp, that's about what I thought was happening.
How about some more, possibly exponential negative feedback? This should let it have serious gain during start up, but quickly drop off as the tank starts to gain momentum... at least in theory that is. I can't think of exactly how to do this right now, but something simple should be possible. Maybe a resistor/capacitor from the inverted side of the amp back to the input... IDK. We should probably look into it though. Seems like this would be more constructive than the low pass filter, or maybe just augment one.
 
Last edited:
I think doing this adds more feedback, stabilizing the amp and hopefully helping...

more feedback.png

You may (probably) need to readjust the resistor values to make it work right.
 
Removing the 0.1uF decoupling capacitor from the collector reduces the distortion. Also reduces the output.

The LPF in my previous attachment was just to show an example of a "real one", in an application where a clean output is mandatory.

One or two "pi" sections would be more than enough for an application like this.
If I get overwhelmed with enthusiasm in the next day or two, I will have a go at knocking one up.

JimB
 
Interesting thread guys! And nice work JimB. :)

Yes the distortion is caused by the amplifier. I took my measurements from the emitter connection.
Connecting the probe to the base connection shows a much lower distortion level. This is to be expected as it is closer to the tuned circuit.
Using a coupling loop to take a sample from the coil itself shows an even lower level of distortion.
...

Maybe you could run a second transistor as a common emitter amp etc, amplifying the clean signal from the coil? The output of that amp might need much less filtering and so reduce the total parts count?
 
Praise indeed!
Thank you RB.

If I get as far as playing with an LPF, I would probably use a buffer amplifier anyway.
Connecting anything which can be variable to a free-running oscillator is usually a recipe for confusion.

JimB
 
Hi RB, welcome.

I'm glad you like this thread. I myself can see a whole host of useful indoor microcontroller applications for this kind of thing.

I actually did briefly think to myself what would RB come up with if he had such minimalistic radio circuits in his projects. And now that I look, you actually have a recent addition to your website about the aforementioned cheap Ebay RF modules. So looks like you were thinking about similar things too. :)

However, with just abstracted OOK, there is not a whole lot you can do other than just go with the abstraction and treat each module as a single wire. But to have available something one can make themselves, and thus something more flexible than a cookie cutter RF module, would open a lot of doors for those of us wanting to experiment with some advanced concepts. Like alternate modulation techniques, frequency hopping, mesh networking, stealth radio, etc, etc. Possibly some pipedream sci fi BS.... but some interesting sci fi pipedream BS no doubt.

The trick is, how do we get something acceptably stable and acceptably simple. Or can we even?


JimB said:
If I get as far as playing with an LPF, I would probably use a buffer amplifier anyway.
Connecting anything which can be variable to a free-running oscillator is usually a recipe for confusion.

Agreed, from what I know about radio (and that is admittedly very little) there is always one source of gain to maintain the oscillation, then an altogether independent buffer amp for the actual output signal. This is done because you absolutely must not put load on the tank part of the circuit, which I gather is essentially what you're saying.



Other thoughts.
I was looking at inductor wrapping techniques to improve Q, what do you all think about this?

I was thinking that if you double wrapped with one strand enamel coated wire, and an equal diameter strand of some non conductive cordage parallel to that, you could put a controlled distance between the windings and significantly decrease the interwinding capacitance. This should significantly improve the Q at the very small cost of needing to make the coil just a little bit longer to get the same inductance.

The other winding techniques seem a little too extreme, they have complicated weave patterns and special coil formers, and also need to be sprayed with some kind of adhesive to keep them mechanically stable. You can't argue with the results, Q's in the hundreds, but far too excessive an effort for these simple radios.

Another Q improving concept for inductors is "litz wire". It can be expensive, and seems ineffective for less than 10 Mhz, but I don't really know for sure what to make of it. I expect one could get the same or similar effect by doubling up a smaller diameter wire. This may also decrease interwinding capacitance.


Edit: Improving Q of the tank should directly reduce harmonics, significantly. That's the point of bring it up.
 
Last edited:
Hi RB, welcome.

I'm glad you like this thread. I myself can see a whole host of useful indoor microcontroller applications for this kind of thing.
... And now that I look, you actually have a recent addition to your website about the aforementioned cheap Ebay RF modules. So looks like you were thinking about similar things too. :)
...

What you are proposing sounds like fun, but you are kindof re-inventing the wheel, after wheels got really cheap...

The other day I saw a 10 pack of TX/RF module pairs on ebay for $16. That's 80 cents for each module, as cheap as a common IC, AND they have been built, tested and tuned ready to go.

To make a TX from scratch requires 1-2 transistor and a xtal etc, and some discretes. That's exactly what the ebay module has, apart from it uses a 433 MHz resonator instead of a xtal.

To make your RX device from scratch will take at least 2 transistors for the receiver and an opamp for the automatic gain adjustment. And that's exactly what the ebay module has.

So I think you might end up designing what is basically the same thing you could buy assembled and tested for 80 cents...

If you changed direction, maybe you could reverse engineer the ebay modules, to (say) change the resonator to run it on a different freq, to tweak the transmitter power/freq, or on the receiver exactly how the auto gain works and how to tune it for best performance. Mods that people might want to do, or even build their own to save 80 cents.

I think that kind of information would have a higher value than just designing one from scratch as it would be of use to more people and opens up more possibilities. :)
 
Last edited:
Hi :)

For me, the difference in purchased modules vs something simple that's built from the ground up is you get to know every single piece, it's purpose, and ways you can modify the circuit and substitute parts... hopefully. Though I admit your idea of just REing one and figuring out how to duplicate or modify it for one's needs is tempting and probably just as good while reaching a wider audience. So that's probably the more logical way to go. But that could be said about a lot of the electronic projects we do. Excluding very subjective circumstances, there is almost nothing we can can come up with that hasn't already been hashed out and mass produced into a common commodity. But being practical is arguably not the point of DIY, though I'm not actually playing that card here.


My objective was to maybe, if possible, go beyond the confines of what those simple module pairs can do, without breaking any laws or getting too difficult.

Unadulterated, those modules are really just wireless wires. You put 5V to a pin on the TX side, and a 5v signal shows up somewhere on the RX side. That's great and all, because that makes them really cheap. However they obviously have limits, you can't change frequencies, you can't use more than one pair in the same space/time, you can't do full duplex, you can't spread spectrum, you can't change the modulation scheme; you can't do lot's of things. If you want to do this stuff, you can go full out and start reversing and rebuilding the entire thing. But then, you essentially wind up back to where we are now.

Now going with a design built from the ground up...

For the TX side, making the oscillator a simple LC design, we can in theory add something resembling a varactor and pull the frequency over a wide range. This lets us create and select channels, lets us FM modulate, lets us do frequency hopping, let's us do full duplex, it lets us do crude PLL; it lets us do lots of things. And that's just with the TX side, and right off the bat because we started with a LC resonator. I can't even think of the other simple things we can do to make it better. Maybe there isn't anything.

For the RX, there are the two methods to discuss that are usually employed, superhet and regen. Most of the RF modules I have seen are superhet, which is fine, the superhet really is the gold standard for a good reason. And though it is somewhat less common, they do make RF boards that are regen. But I think a DIY regen design, built from the ground up with the same philosophy the TX is built from, and with obvious emphasis on the two being a pair, has the potential to give just as many feature gains over the cookie cutter modules as doing the TX side from scratch does.



Anyway, that's all just my open ended theories. I don't do radio a lot, so I expect I don't have that critical insight needed to judge these theories properly. It's just, to me, having a simple "wireless-wire" is less than thrilling or useful, I can do that with infrared Laser diodes from a longer distance, with higher bandwidth, and cheaper, so the RF modules are mostly useless in that regard. Their only purpose for me is just to do things at range and out of line of sight. Which, in the same room, line of sight is almost always trivial to obtain.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top