Wireless Transmission of a 200 kHz Bandwidth Analog Signal.

[Benta]

Hi All,
I'm involved in a new project and need suggestions.

Short version:
I need to transmit an analog sensor signal with a bandwidth of 5...200 kHz over a 20...300 m distance. Right now it's done over shielded cable at up to 10 m, but for longer distances that's not viable. It has to be wireless, and as line-of-sight is not quaranteed, I only see RF as an option.
Suggestions for band, modulation, module availability anyone?

TIA.


Long (TL,DR) version:
I can't say anything about sensor type etc., sorry (customer confidential).
But I can say the following:
Baseband signal is well-defined and works perfectly over cable. The output signal is:
300 mV_RMS. This can be scaled or buffered for level and output impedance, no problem there.
5...200 kHz bandwidth. Think of it as a standard audio signal, but with much higher bandwidth.
Distortion <5% is fine, it's not HiFi.

The ideal situation would be to take the baseband signal as-is, and simply modulate it on to an AM or FM VHF/UHF carrier, just like a small private radio station.

This where my problem begins:
I can't find a frequency band where that's possible. I'd need a channel width of at least 500 kHz, which doesn't seem to be available for free use. Along with that, no off-the-shelf radio modules are available.

I want to avoid signal processing/compression on the transmitter side for several reasons (complexity, delay...).
I might be able to reduce the baseband bandwidth to 50...150 kHz and use SSB-SC transmission to get the channel width down, but that opens a can of worms on complexity.

I thought about an A/D converter connected to WLAN module of some kind, but on that I'm in over my head. And it raises concerns about signal delays.

Any suggestion appreciated.

Thanks.

/TL,DR

[T3sl4co1l]

Perhaps 2.45GHz band, many services use wide channels there (e.g. WiFi 10-20MHz).  WBFM would be fine I guess -- give or take any regional limits on channel use/allocation, ERP, etc.  I don't know the rules and exceptions (or what kinds of licensing) define that though.

Think you'll have a hard time with line-of-sight, and ERP limits at that range, even with a sensitive receiver antenna (a good Yagi is easy enough to make, at least).

Perhaps you can arrange some multipath reflectors to ease reception, but that's still costing transmission path gain.

No bands with easier transmission (really just, longer waves --> looser diffraction) have as much bandwidth, AFAIK.  Or, let's see,
https://en.wikipedia.org/wiki/ISM_radio_band
maybe 27.12MHz if you can use the whole thing, or 433 or 900 if available.  Otherwise, have fun licensing that..?

Tim

[Buriedcode]

Older 2.4GHz analogue systems - the ones used for AV will have the bandwidth - the video channel is about 5MHz - but given the crowded band, and the fact they are pure analogue you're bound to get interference.

200kHz is close enough to the 192kHz/384kHz of audio gear that you may be able to use an off-the-shelf product.  Not one for instruments which has much limited bandwidth and is designed for minimal latency, but something that sends audio over wifi channels. Your tihnking of A/D converter + WiFi is probably your best bet.   As a basic on-the-shelf WiFi solution goes, the ESP32/ES8266 are so cheap is silly, and they have diverse development enviroments (Arduino, Eclipse, custom tool chains) thats all free.

So I tihnk I'll just echo your line of thinking - and ADC + ESP32 module.  Is this signal AC coupled?  If so you can use an audio sigma-Delta ADC designed for 384kHz and just slightly over clock it to 400kHz.  These have digital antialias filters making support circuitry very simple- but will not carry a DC component since there is a minimum frequency.

As the bandwidth is only 200kHz though, any reasonble SPI ADC will do (SAR?) with a front end filter that meets your requirements for out-of-band nosie.  This may require a fair bit of prototyping to get it right, but transfering ADC results from one ESP32 to another shouldn't be too difficult to get up and running.

For others, it might help if you narrow down more specs:

1) Latency.  Are you just recording this signal in another location? Or is it being used as feedback for a real-time process?  Can you tolerate 100ms delay? 1000ms? or should it be as fast as possible?
2) Resolution.  200kHz bandwidth means a minimum of >400kHz sample rate, but how accurate/fine grained should the reconstruction be? sending 8-bit values are 400kHz is 3.2Mbps (0.4MBps)
    but if you need highers resolution, 12-bit samples increases this to 4.8Mbps (0.6MBps)
3) Power.  Is power consumption going to be a problem? Is either end of this link portable ie: battery powered?  I ask this because WiFi solutions are power hungry (>200mA).


Edit: So many typos.

[Benta]

Thanks.
The 27.12 MHz idea is intriguing. I imagine telling the customer "We'll be using the epilation frequency for the communication link." He'll be thrilled.
But jokes aside, it could be perfect. Do you know of any TX modules? I could imagine that modulation excursion could be an issue.

Intriguing is also the analog use of 2.4 GHz, which needs a closer look. Same question: know of any modules?

Thanks again, and I'm open for all suggestions. These two weren't even on my map until now. 

Generallly, I'm not too worried about interference in my environment. For A/D conversion, 8 bits is enough, but 10 bits would be nice for signal strength tolerance.

[Postal2]

Older 2.4GHz analogue systems - the ones used for AV will have the bandwidth - the video channel is about 5MHz ....

Correct. I have such system on 5GHz - works perfectly.

... Intriguing is also the analog use of 2.4 GHz, which needs a closer look. Same question: know of any modules? .....

Yes I have (5GHz), I bought that on Aliexpress.

It's called Pat-530 200m 5.8 GHz. (US $40.90)

[mag_therm]

Hi Benta,  Can you use UDP? (--considering its occasional disadvantages which should be slight if you can set up a dedicated sub net.)

I use UDP for link between ham radio sdr and computer. (wspr  bidirectional running 24/7)  It worked well on Netgear Nighthawk 5GHz,
although I now use cables and a switch to save energy.
If the link is over private property (you mention 2c screened), routers can be added as relays. I have done that too

[Benta]

@mag_therm.
Thanks, but you lost me completely.
If I could do this with a couple of transistors, that would be ideal. But it's a customer project.
It sounds a bit like you think I have UDP, Netgear Nighthawk, ham radio sdr, routers, relays and stuff lying around my property.
I don't, but Thanks.

[T3sl4co1l]

You can do it with a couple transistors, absolutely; find the last surviving, oh say BFR92 or better, put one or two together in the boring old FM microphone / tape player to radio adapter circuit, tune it to 2.4x GHz instead, and there you go.

Won't be at all stable, lord knows power output, purity, etc.  All that is left as an exercise to the reader

Actually, need not even be too awful if it's done with a dielectric puck, transmission line or cavity resonator.

But ready-to-go things, made responsibly with properly designed channels, limited bandwidth, well defined SNR, etc., yeah dunno; a few possibilities but most will be too narrow band or low bitrate (audio?) or too specialized (video?).  Though there's a possibility a video converter is just that, a converter, and doesn't care about what composite or other signal is fed into it -- could be worth trying to feed one with your signal anyway, or maybe with some minor encoding if you can afford to (mimicking sync pulses to set DC offset, perhaps?), and see if you get the same thing out the other end.

Kind of the same principle as, there's at least one USB (or maybe it was Thunderbolt) to VGA converter, that's literally just a triple-channel DAC fed by a frame buffer and DMA; essentially 100% software defined / CPU based graphics adapter, how horrible right, but, I mean PCs can get away with that kind of thing.  (Remember the days of "soft modems" dragging down the Win98 PC? Needless to say, days long gone now!)  This had been noticed by some avid hackers, who repurposed the frame buffer into an SDR transmitter, just plug a strip of wire into the VGA jack and go.   Wouldn't have been possible if it was in any way VGA-purposed, i.e. internal frame buffer, sync generator, automatic blanking, etc..

Tim

[BrianHG]

(Cheap and dirty method...)

Find yourself an old TV UHF wireless sender and an old TV tuner.

For North American versions, you will get a really clean 2MHz bandwidth with flat signal integrity dropping as you approach 4MHz.

Interference can be an issue as analog television is an AM modulated system, but if you send digital through the video channel with error redundancy, you can send a good 2 megabit easily using simple MCU's UART port driving the video transmitter's input at 0.7vp-p.  On the receiver side, you will need a fast comparator with a few filter caps to get the weak analog ~1.5vp-p video signal back into a digital form.

On the other hand, if you can get you hands on an old analog satellite transmitter in the LNB IF band.  (950-1450MHz), You will now have a pure FM modulated signal with a 16/27/36MHz bandwidth options on the cheapest of satellite LNB IF tuners (obviously you do not need the LNB, just a little wire as an antenna as you are working right at the tuner's 950mhz band instead of the satellite 10ghz band...).

[BrianHG]

Maybe something like this:
Garbage transmitter

You will need a cheap receiver with no 'blue-screen' electronics.  Or, if you are testing with an old VCR tuner, you may need to tap the tuner's video output before the blue-screen 'no-video' electronics come into play.

[paul@yahrprobert.com]

There's an entire industry that does this:  aerospace flight test telemetry.  I used to work on S band (2.5 GHz) telemetry transmitters.  We used mostly 5 Mbps FM NRZL encoding.  Transmitters typically did 5 watts from a package the size of a pack of cigarettes but cost a fortune (they had to survive all kinds of vibe and shock over big temperature ranges).  The bandwidth was right around 5 MHz IIRC.
I would hope somewhere some company is adopting these things for cheaper commercial applications.....

[Buriedcode]

I'm not sure the "quick and dirty" methods are really applicable - unless this is a one-off DIY home project, surely you want something thats reliable, and perhaps also legal in their country.

Telling OP to DIY analogue radio links, or buy dubious modules with no regard for FCC or EU regulations is just plain daft.

Intriguing is also the analog use of 2.4 GHz, which needs a closer look. Same question: know of any modules?

A random google shows you can sitll get 5.8GHz "analogue" modules for drones/RC stuff:
https://drones.altigator.com/5-8ghz-fpv-wireless-video-transmitter-mini-module-200-p-40932.html

I've used these to send digital audio through a few floors - albeit without any error detection/correction/masking. And whilst these are great for high bandwidth comms - if you're dsoing it digital you'll need to packet up your data, channel coding, error detection, CRC sync etc.. it very quickly becomes quite complicated.  Where-as an existing standard like Wifi, or even bluetooth (3Mbps for BT 3.0) do all of that for you, and have the added bonus of being legal, having widespread availability and guaranteed not to cause random interference with other equipment.

Generallly, I'm not too worried about interference in my environment. For A/D conversion, 8 bits is enough, but 10 bits would be nice for signal strength tolerance.

Righty, so that puts the low limit of data at ~3.2Mbps - beyond the limit of the NRF24L01 (which can have a range of a few hundred metres with a front end amp).  And probably even bluetooth.  But again, I'm still thinking digital which has many advantages but may be overkill depending on the noise requirements.  It's just analogue radio links aren't as popular these days given their limitations, and as such, finding modules/systems that accept analogue and are compliant are rare. 

Thats why I mentioned the 2.4GHz/5.8GHz AV modules - they are the only high bandwidth analogue systems I could think of. If your signal has no DC component, you could try just passing it straight through the video channel of one of those links.

[dmendesf]

Stm32F4 ADC (2.4msps @ 12 bits) -> 8b/10b -> I/Q modulation with DACs -> any analog module.with enought bandwidth cited before -> everything in reverse

8b/10b gives you frame syncronization for 20% more bandwitdh

[Poroit]

I suggest have a closer look at A/D and use standard Wireless Mesh products.

https://hinco.com.au/shop/brand/define-instruments/pathfinder/

https://defineinstruments.com/

It's ashame you cant divulge what parameter you are measuring.

[Phil1977]

Why exactly does it need to be analogue? Is it due to latency?

Though the selection is tough, there are also digital wireless standards with <10ms of latency.

[T3sl4co1l]

There's an entire industry that does this:  aerospace flight test telemetry.  I used to work on S band (2.5 GHz) telemetry transmitters.  We used mostly 5 Mbps FM NRZL encoding.  Transmitters typically did 5 watts from a package the size of a pack of cigarettes but cost a fortune (they had to survive all kinds of vibe and shock over big temperature ranges).  The bandwidth was right around 5 MHz IIRC.
I would hope somewhere some company is adopting these things for cheaper commercial applications.....

For those curious, this is... well, with the bitstream input, I guess actually quite a lot simpler, but it depends what's going into the ICs preparing said bitstream -- but, the current/ongoing video series by CuriousMarc,
https://www.youtube.com/@CuriousMarc/videos
concerns the communications and telemetry used by the Apollo space program (featuring authentic hardware).  They used, as I understand it, an FM microwave link, containing multiple FM/PM/QAM and digital subcarriers, which was only a little easier at the time to compose than a contiguous bitstream; but, suffice it to say, if one were to compactify all those discrete and low-integration-IC modules into a couple chips, it would probably come out to about that size and capability.  So it's no accident that they'd have used something like that.

Tim

[ftg]

I'm not sure the "quick and dirty" methods are really applicable - unless this is a one-off DIY home project, surely you want something thats reliable, and perhaps also legal in their country.

Telling OP to DIY analogue radio links, or buy dubious modules with no regard for FCC or EU regulations is just plain daft.

Intriguing is also the analog use of 2.4 GHz, which needs a closer look. Same question: know of any modules?

.snip.

Thats why I mentioned the 2.4GHz/5.8GHz AV modules - they are the only high bandwidth analogue systems I could think of. If your signal has no DC component, you could try just passing it straight through the video channel of one of those links.

This would have been my first pass suggestion as well. Just trying it with a cheap 5.8GHz or 2.4GHz analog video transmitter and receiver pair.

Baseband video signals have stuff near DC so the low frequency response could be usable as is.

If not, the baseband could be mixed up 1MHz to avoid any LF roll-off.
Some modules also have one or two audio channels that are FM modulated subcarriers, but those will likely have some pre-emphasis and de-emphasis on the signal path.
Along with the FM IF filters likely being 180 - 280kHz wide and likely specced for something similar to the +-50kHz deviation seen in analog television.

[Benta]

Thanks for all your suggestions.
I used the term "delay" in my post, but "latency" is the correct term, thanks Phil1977.

Yes, that's actually one reason for not preferring digital transmission. Directly modulating an AM or FM transmitter with my baseband signal eliminates latency concerns.
I'm veering towards AV modules at 2.4 GHz, like these:
https://www.farnell.com/datasheets/64198.pdf

But whether such things are available these days needs to be researched.

I also thought about AV UHF modulators like back in the early home computer days, but they also seem fo be gone.

Perhaps I need to rethink the concept completely.

Thanks to you all.