RF over fiber CATV FTTH for antenna array feed?

[Neil Smith]

I found out at the EUCARA conference that the SKA LPDA fir-tree antennas use direct RF over glass connections back to the receivers. 

There are low-cost RF over monomode fibre interfaces intended for CATV local distribution with fibre-to-the-home that are available at low cost these days. 

RF bandwidth is quoted at 45 MHz to about 1.2GHz. I'm considering whether to use RF over fibre instead of coax for long (120 metre) runs of connections between receive antennas in an array. Also for the IF feeds to/from masthead transverters. That would allow complete galvanic isolation to help reduce noise pickup on the feeders - assuming of course that I use local battery power for the interfaces at the antennas.

The copper feedlines have to pass several large antennas that run up to a kilowatt on transmit on many bands, plus they go through a duct close to my machine shop, where there are ten multi-kilowatt inverters throwing out RF noise.

Some of the RFoF CATV interfaces have AGC, which I would need to disable. 

 I'll be using very low noise-figure LNAs or downconverters at the antenna feedpoints and aiming to get a background noise level at tye radio room end of the links with the antennas pointing at cold sky that's a few dB more than the noise floor of the fibre modules.

I guess the SKA is using fancy Huber&Suhner RFoF kit that costs 100 times as much as these cheap FTTH modules!

Has anyone tried using this approach on an amateur array?

Neil

Yorkshire UK

[Bruce Rout]

Hi Neil. I'm not as far along as you but I have what looks like the maximum length of coax to antenna. About 360 m. I'm going to use a couple of t-bias connectors to get power to LNAs at the antenna. Fortunately I'm in a very radio quiet zone. Not even cell phone coverage. Eventually I'll have to get power to the antenna for declination control but I'm first planning to see if I can get a signal. I'm planning on using LMR-600 coax. So I'm very interested in hearing of results using fiber. I didn't know a raw signal could be transmitted over fiber. I would have thought it would need to be digitized first.

-Bruce 

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/sara-list/CAGbfs6VznjHW%2BU_iN_BgEEFfzz_stBZ69y9_0rSQH_uJL1aKGA%40mail.gmail.com.

[fasleitung3]

Hi Neil,

I have not tried this but I am interested in trying out. What low cost modules are you considering? Do you have a link for that?

Thank you,

Wolfgang

--

[Neil Smith]

Hi Wolfgang, I'm thinking of transmitters like this: https://www.aliexpress.com/item/1005003738333315.html

and receivers like this:

https://www.aliexpress.com/item/1005004073845975.html

I might order a few examples to see if they are any use. Normally you need a distributed-feedback laser in the optical transmitter to get good linearity. That's probably why the "real" RFoF units are hundreds of dollars each.

I'll try some two-tone intermodulation, dynamic range and noise tests to see if there is any benefit from these simple units.

Neil

-- 

[duncan campbell-wilson]

Hi Neil,

Good that you are thinking of using  fibres.

I have used fibres in antenna arrays and as usual there  are a few things to be mind full of.

(1)  The fibres have temperature dependent performance. This is relevant for interferometers.

(2) The laser transmitters are noisey both electrically and optically compared to a copper system. Therefore the electrical  amplification prior to the laser diode is large. In arrays of closely spaced antennas RF shielding becomes important. In the array case, failure to make sure that the inter element post amplification cross talk is very low may cause RF electrical instability and oscillation in the whole array.

(3) The optical components  are heavily  dependent on the photon induced current. (photo current squared).

(4) The dynamic range (cf copper based systems) see (3) is reduced wrt remaining linear in the presence of RFI.

(5) Optical beating in the fibre due to reflections off reflecting interfaces ( photo diode) can cause grief. A fibre that doesn't  have a good optical termination is likely to induce electrical ripple  in the bandpass  across  wider band system post diode detection.

This is my experience using  low cost single mode fibre systems in a phased array.

Best wishes for your endeavour.

Rgds Duncan 

To view this discussion visit https://groups.google.com/d/msgid/sara-list/da8d7cad4476e6109e9dcbf8760ee8154844755c.camel%40googlemail.com.

[Neil Smith]

Thanks Duncan, I've been around fibre systems since the late 1970s, doing hand-cleaved joints and optical microscope physical jointing backnin the day, but haven't done anything with analog over fibre at all. 

My primary use-case is actually for single remote receive LF terrestrial antennas at 8270 Hz and 136/472/1830 kHz, not for astronomy, but that's way below what the CATV units can support. 

It just struck me that it would make for an interesting experiment to avoid all of the appalling RF noise from my neighbours by running fibre from all of the antennas around 100-180 metres from my house.

 VHF pollution from solar panel is terrible here, heatpumps are worse, and car chargers, scooter chargers, noisy LED lamps and all the inverter-driven motors in washing machines and freezers make life miserable for those of us addicted to weak signals. 

If the analog fibre experiments fail, I'll just run some of my SDRs remotely and link them back over the same fibre but using gigabit ethernet extenders. I guess I'll at least be able to use the RFoF solution to distribute local oscillator signals and 100 MHz reference clocks from my Rubidium GPS-steered master reference even if it's no use for low-level signal transmission.

Neil

[duncan campbell-wilson]

My current thoughts are send a clock up the fibre and bring back digital on the fibre . I also think in arrays down conversion is safe option as there is as you have frequency separation / isolation .  Rgds Duncan

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sara-list/CAGbfs6V66mc79_rmxoUUByhMj7YJU%3DnmvhUnjmwoPqA29nBx%3DA%40mail.gmail.com.

[fasleitung3]

Please let us know how your tests go. In particular it would be interesting to see whether these units are still usable at 1.4 GHz.

Thanks,

Wolfgang

[Don Latham]

This is extremely interesting...

Don

---

From: "Neil Smith" <neil...@gmail.com>
To: "sara" <sara...@googlegroups.com>
Sent: Tuesday, September 23, 2025 3:55:57 PM
Subject: Re: [SARA] RF over fiber CATV FTTH for antenna array feed?

To view this discussion visit https://groups.google.com/d/msgid/sara-list/CAGbfs6WBQn2NOx2R77OUz8DHTa18-1JErQ_5SbPohiKhbVP0OQ%40mail.gmail.com.

--

------------
Don Latham
PO Box 404,
Frenchtown, MT, 59846
406-626-4304

[Captain Anne Flint]

A stupid question - why wouldn’t fiber [inherently?] support analog signals? Wende 

To view this discussion visit https://groups.google.com/d/msgid/sara-list/1643960803.210095660.1758725984400.JavaMail.zimbra%40blackfoot.net.

[Marcus D. Leech]

On 2025-09-24 19:29, Captain Anne Flint wrote:

A stupid question - why wouldn’t fiber [inherently?] support analog signals? Wende

Because laser diodes aren't particularly linear, from what I understand.   So, you impress some kind of modulation on
  the laser from the analog signal.

To view this discussion visit https://groups.google.com/d/msgid/sara-list/ADC216B4-C47B-4272-A928-84FD4A9C5562%40gmail.com.

[Neil Smith]

You need something to linearize the output, so distributed feedback lasers coupled tightly to PIN photodiodes are often used to get reasonably linear response, but there's threshold effects to handle and Rayleigh Backscatter is a nightmare, so fancy slanted fiber interfaces are required. Also dithering is used to help reduce the problems at lower rf frequencies. There's some good papers about the choices made for the SKA, but all behind paywalls. There is a rather dense paper on the ArXiV preprint server at  

https://arxiv.org/pdf/2005.02733

 that goes into excruciating detail. 

Lots of RFoF systems out there for thousands of dollars with fancy DFB lasers with extended Bragg mirrors or Fabry-Perot systems in fiber. I was just hoping that mass-produced CATV RFoF systems might give me a way to avoid having a zillion copper feedlines running though my transmitting antenna farm.

Simple answer, laser diodes are horribly nonlinear without countermeasures.

Neil

[Bruce Rout]

Not stupid at all.  The glass fiber does not conduct electrical current. Light flickers appropriately at one end and is detected at the other. The light is in the form of a laser. This is an enormous oversimplification 

To view this discussion visit https://groups.google.com/d/msgid/sara-list/13c370bd-2e04-4f32-be35-f95273476c3c%40gmail.com.

[Lamar Owen]

On 9/23/25 10:46, Neil Smith wrote:
> I found out at the EUCARA conference that the SKA LPDA fir-tree
> antennas use direct RF over glass connections back to the receivers.
>
> There are low-cost RF over monomode fibre interfaces intended for CATV
> local distribution with fibre-to-the-home that are available at low

> cost these days. ...

>
> Has anyone tried using this approach on an amateur array?

I have a couple of inexpensive units: an HY-21-R38 receiver and matching
YH-21-ST15P10B transmitter, that I have thought about using for a
project with one of our radio telescopes.

For our 26 meter antennas here at PARI, we actively use Ortel/Emcore RF
over fiber units for S-band and X-band work.  For S-band, we have
Ortel/Emcore WiBa 10382B 10-2700MHz units which have internal pre-laser
amplifiers and adjustable attenuation; for X-band, we have 10341B
0.1-13GHz transmitters which do not have adjustments.  The 10341Bs were
$10,000 each back in 2006-2007 when they were purchased under a grant
for the Dedicated Interferometer for Rapid Variability (DIRV -
https://www.researchgate.net/publication/233476450_The_dedicated_interferometer_for_rapid_variability
and https://www.nsf.gov/awardsearch/showAward?AWD_ID=0520928 for
reference data).  Both of our 26 meter antennas were outfitted with feed
cartridges with dual-band-dual-polarization feeds for S and X bands, and
on the last day of the grant term in August of 2013, fringes were
finally observed at X-band, with S-band fringes observed one year
prior.  Those same Ortel units can be found on eBay these days for far
less money; we bought a couple of populated rack units for a bit less
than $1,000 each about three years ago.

Ortel/Emcore has a design guide that goes into great detail about the
challenges; you can find a copy on Scribd at
https://www.scribd.com/document/146479987/RF-and-Microwave-Fiber-Optic-Design-Guide
(I found a few other copies online, but from shadier websites.  I have a
printed copy from way back when).

These Ortel links have around 30dB of dynamic range and are purely
analog in their operation.  The fiber itself that we use, Optical Cable
Corporation military tactical with Siecor SMF-28 at the core, is the
only cable we could find that is rated for the repetitive bending of an
antenna system, and has been extremely reliable in practice.  But it is
a fact that the fiber run (600+ meters for one antenna) has variable
attenuation depending upon the pointing angle.  So calibration for every
observation is a must.

For S-band, the worst interference here lies at 2.125, 2.150, and 2.175
GHz in the form of 4G/LTE towers for Verizon; the next worst signals are
just above 2.3GHz in the form of Sirius/XM.  Pointing directly at a
Sirius/XM satellite with a 26 meter dish can dramatically overload the
RF chain without post-LNA filtering.  Even with the filters that we
have, the LTE and Sirius/XM RFI is so strong that we run in the low end
of the dynamic range band.  On S-band the original design has 30 dB of
gain on the LNA, then an isolator and bandpass filter, then 40dB of gain
prior to the fiber input, and I later added another filter before the
fiber transmitter to knock down the 2.1-2.2GHz LTE RFI.  So 70dB gain,
with isolators where needed, from feed to fiber transmitter.  On X-band,
the LNA has 22 dB of gain, then there's the isolator and filter, then
two 30dB gain stages for a total of 82dB of gain prior to the fiber
transmitters.

Charles Osborne did the original design, and it has worked well in
practice, with some additional filtering and tuning of the S-band fiber
transmitter attenuation control to deal with the extremely high LTE RFI
(20dB above the Sun!).

Since the WiBa S-band links have built in amplification, the receiver
output is coupled through an isolator to the mixer and then the IF is
filtered and amplified appropriately.  On X-band, there is a 30 dB gain
stage after the fiber receivers before hitting a 2 port power divider,
isolator, and mixer input, since those fiber receivers have no built-in
amplification.

So I have quite a bit of experience with how these higher-end devices
react to RFI for comparison purposes.

So, now to these inexpensive units. First, you need to find the dynamic
range of the devices at RF, if it's not sepcified.  This is best done on
the bench using a spectrum analyzer with tracking generator, like what
you would use to sweep filters.  Set up the fiber links, and play around
with signal levels, watching and recording the gain to the output.  Use
a marker generator to check for distortion.

For the actual link design, use fiber attenuators to get the optical
signal link into the rated optical level range, which is pretty narrow
for inexpensive devices, and be sure to use the correct connectors. 
SC/APC and FC/APC are pretty standard for RF over fiber units; the
angled faces of the connector dramatically improve return loss
performance, and would be used through the system in an ideal world (our
26 meter systems work ok with SC/UPC connectors in the line, but again
those are the higher end Ortel units).  The inexpensive units I have
state they have AGC on the optical link, and there is no obvious way to
remove that.

Once the optical signal levels are within spec, you can determine how
much gain you need at the LNA to get the expected noise floor into the
dynamic range window.  The inexpensive transmitter I have specifies 80
+/- 5dBuV at the input, with a 75 ohm input impedance; this converts
(using the handy
https://www.everythingrf.com/rf-calculators/db-v-to-dbm-calculator
converter) to -28.75dBm.  The dynamic range is 10dB (+/- 5dB), which
works out to a range of -33.75 to -23.75dBm .  So, take your expected
power to the feed in dBm and add enough gain to get into the range of
the input of the fiber transmitter (at 75 ohms; an isolator immediately
prior to the fiber transmitter might be enough to keep passband ripple
at bay, but either using 75-ohm amplifiers or a 50-75 ohm transformer
would be best).

It's hard enough to keep a 30dB dynamic range fiber link happy; 10dB is
likely going to be much more difficult.

There is another purely analog option, however, short of just putting
the SDR at the dish and using gigabit/ten gigabit/ hundred gigabit
ethernet links.  You could downconvert at the dish and then frequency
modulate the IF there, and transmit the FM signal over the fiber.  FM
doesn't need a large dynamic range over the fiber to carry a large
dyamic range of the signal.  Your linearity is then limited by the
dynamic range of the FM modulator/demodulator pair. This would be a
custom device; I don't know of any off-the-shelf units.

Hope that's helpful.

[Neil Smith]

Very helpful indeed. Many thanks.

Some of these low-cost devices appear to have at least 20dB linear dynamic range and depending how you read the spec, could do as much as 40dB with AGC disabled.

It will be amusing to see how they behave in reality. I've seen a few Huber&Suhner RFoF devices at not-terrible prices, but my specific interest is in feeding a biggish (32 antenna) phased array with cheap SDR devices as the receivers, so price is important.

My other use-case is for VLF receive using e-field capacitative probes without needing any digital noise-makers nearby. The dynamic range requirements are much greater though, so even using the baseband signal to modulate a VHF carrier with FM couldn't even start to approach the dynamic range of my 16-bit SDRs (at least 100 dB).

Neil

-

[Bruce Rout]

Thank you for your detailed description. I'm blocked from the PARI.edu site. Is it because I'm outside of the US?

--

--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sara-list/CAGbfs6XwCZOke8wgyxTcoON87Qs7cdHw93YHeBiu%2BC9_N%3DpbVQ%40mail.gmail.com.

[Lamar Owen]

On 9/26/25 09:38, Bruce Rout wrote:
> Thank you for your detailed description. I'm blocked from the PARI.edu
> site. Is it because I'm outside of the US?
>

I replied directly to Bruce, but wanted to put this out on the list, too.

No doubt because of what the PARI site used to be, our website is
constantly under attack.  No defense is perfect; there will be false
positives as well as false negatives, and we have had several damaging
false negatives over the years.  So if you can't see our site and want
to, please let me know by email.  The email server's firewalling is less
aggressive than the website.  I neither need nor want your individual IP
address, just the IP address range you're coming from.  It will then be
researched and the firewalls' criteria can then be appropriately
adjusted.  We use several layers of dynamic firewalling, and it's a
really difficult balance to achieve of allowing every one who should be
allowed access while blocking the bots and scripts that would otherwise
hit us thousands of times per minute.

Most of the abusers actually are within the US, and use Amazon AWS ans
Microsoft Azure cloud servers, among other providers.

The big downer is that working on the firewalling and website security
takes a lot of my time that I would rather be spending on the antennas. 
I wish the malicious actors would just leave us alone, but that's not
going to happen.  We're too enticing a target.

In any case, I apologize for any inconvenience our unfortunately
necessary firewalling causes users of this list.

You can use the Internet Archive's Wayback machine to look at our site;
the most recent snapshot is
https://web.archive.org/web/20250905132909/https://www.pari.edu/  Of
course, the Wayback Machine can get pretty slow at times, but this is
one way around our firewall for indirect access.

[Bruce Rout]

Thank you. I want to investigate your site. I am on the road now and will get you the range ip when I get back. 

Again, thanks for your efforts.

Bruce 

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sara-list/f47b6cd7-1b70-4af6-95f3-0ba1fe89eccc%40pari.edu.

[Bruce Rout]

Lamar, I did not get an email from you. My external ip begins with 

209.142. if that helps.
I was able to use the wayback link.

this is the error I get with pari.edu

You are not allowed to visit this website.

Your IP: 209.142.xxx.xxx