Cat 5/6 instead of coax for Amateur Radio feedline ???

[KM4FER]

In my "Shack" (implies higher quality than what I actually have) there is an unused Cat 5 jack that used to be for AT&T Uverse high speed internet feed.  Outside there is enough cable that I could use to feed a dipole antenna.  So rather than make a new hole in the wall and buying coax feedline I wondered if I could use this Cat 5 cable for low power HF (30MHz max) QRP work. Having the extra pairs in the cable could be handy for power and control of remote amplifiers, antenna tuners or such.


Belkin specs for Cat 5e:

Electrical Characteristics
Impedance: 100+15 Ohms
Mutual Capacitance, Max. nf/1000 ft.: 17.1
DC Resistance, Max. Ohms/1000 ft.: 28.6
DC Resistance Unbalance of a pair: 5% Max.
Capacitance Unbalance (Pair to Ground): 330pf/100m Max.
Propagation Delay Skew: 25 nS/100M
Normal Velocity of Propagation: Plenum FEP insulated 69%,
Non-Plenum insulated 66%
MHz IMPEDANCE ATTN. NEXT ACR ELFEXT PSNEXT PSACR PSELFEXT R.L.
1 100+15 2 68.3 66.3 68.8 66.3 64.3 66.8 22
4 100+15 4.1 59.3 55.2 56.7 57.3 53.2 54.7 25
8 100+15 5.8 54.8 49 50.7 52.8 47 48.7 26.5
10 100+15 6.5 53.3 46.8 48.8 51.3 44.8 46.8 27
16 100+15 8.2 50.3 42.1 44.7 48.3 40.1 42.7 27
20 100+15 9.3 48.8 39.5 42.8 46.8 37.5 40.8 27
25 100+15 10.4 47.3 36.9 40.8 45.3 34.9 38.8 26.3
31.25 100+15 11.7 45.9 34.2 38.9 43.9 32.2 36.9 25.6
62.5 100+15 17 41.4 24.4 32.8 39.4 22.4 30.8 23.5
100 100+15 22 38.3 16.3 28.8 36.3 14.3 26.8 22.1

I'm not sure I understand transmission lines and antennas yet but it looks to me like with impedance matching at each end this ought to work.

Am I way off in my thinking?

thanks...
earl...

[T3sl4co1l]

Losses are a bit higher. For HF, probably not a concern.

You can't really connect pairs in parallel or series, because their velocities aren't matched.  (Again, may not be a problem for HF, but beware.)  Obviously, it's not coax and it's not 50 ohm, so you'll need a matching balun, or else your transmitter and antenna grounds will fry everything, even for QRP.

The additional wires probably need to be well filtered.  Theoretically they are all balanced, but if your balun isn't ideal, for example, expect coupling.  If the control signals are slow anyway (provide filtering at both ends if they are driven by switches), using them in an unbalanced manner would be okay.

Tim

[TheSteve]

You can certainly use it for control of other devices but should really stick with proper coax for your feed line. All kinds of nasty things will likely happen if you try to use it for RF.

[hendorog]

You can certainly use it for control of other devices but should really stick with proper coax for your feed line. All kinds of nasty things will likely happen if you try to use it for RF.

I'd like to know more about why it wouldn't work. Electrically its just a balanced feed line with an impedance different to 50 ohms and there are many HF installations that fit that description.

I don't know what the effect of the other pairs in the cable would be though.

[Z80]

In theory I don't see why you couldn't do it at very low power levels.  I've seen video sent over cat5 using small adapters which are just a toroidial balun in a box.  The problem is though if you try and put any kind of power down it and you get any mismatching you could get arcing inside the cable.  Coax is rated for high voltages, cat5 isn't.

[cdev]

Look up the Chris Trask *receiving loop* with varactor tuning and balanced feed line. I would do that and use the feedline to control things like some additional caps to switch in and out and possibly switching the loop so you can have 1,2 or 4 loops (series or parallel or all combined into one). You can also put a balanced matching network and a low noise preamp in it so then the length of the feedline wouldn't matter at all. But because it was balanced you also might have superior noise immunity.

[uncle_bob]

Hi

The lan cable is not as well shielded as good coax, but it's better than some of the cheap stuff out there. It most certainly does not have the power handling capability of coax so QRP needs to mean just that -- very low power. The limitation comes when you heat up the dielectric in the twisted pairs. At best, only run one antenna per cable. The cross talk isolation that is fine for a LAN (all signals are pretty strong) is not going to be that good for RF.

Bob

[vk6zgo]

It will be a lot of messing around for a fairly poor result.

Use the CAT cable as a "draw wire" & pull plain old RG58 through.
You now have a real feeder,& you have salvaged a useful length of CAT cable.

[Richard Crowley]

If you wanted to make a DIY remote-controlled transceiver mounted aloft with the antenna, that is one thing. And not terribly difficult with modern technology.
However, using UTP for RF communication seems pretty dubious. Mostly, I think because of significant line-losses in the cable. If you are running QRP you can't afford such significant losses either transmitting or receiving.

BangGood have a couple of QRP transceiver kits for $10 or less. They could easily be adapted for remote-control via Cat5 UTP.

[uncle_bob]

Hi

One thing to note, your Belden table of attenuation versus frequency is based on a fairly short run of wire. You already have lost 6db at 10 MHz based on that table. If your transmitter puts out 1W, only 1/4W is getting to the antenna. The rest either goes up as heat (losses) or is radiated to who knows where. By the time you get to 30 MHz, it's really pretty crazy.

Assuming you are running 10 to 30 feet, all that may not be an issue. If the antenna location involves a run of over 100 feet ... you likely are stuck at 3 MHz.

Bob

[cdev]

Two conductor balanced feedline for receive might be far superior to coax in terms of noise immunity, for the same reasons that pro microphones like to use balanced XLR audio connections.  Also, you can send DC up that feedline and put a receive preamp at the far end so then no amount of loss is going to matter because the first receive gain stage is going to be the important one and that is outside.

That solves a great many receive problems.


here is a two varactor remotely tuned loop design..You can put two more in there in parallel to get it to tune lower.

http://home.earthlink.net/~christrask/Varactor%20Tuned%20Loop%20Antenna.pdf

[hendorog]

Hi

One thing to note, your Belden table of attenuation versus frequency is based on a fairly short run of wire. You already have lost 6db at 10 MHz based on that table. If your transmitter puts out 1W, only 1/4W is getting to the antenna. The rest either goes up as heat (losses) or is radiated to who knows where. By the time you get to 30 MHz, it's really pretty crazy.

Assuming you are running 10 to 30 feet, all that may not be an issue. If the antenna location involves a run of over 100 feet ... you likely are stuck at 3 MHz.

Bob

According to the spec sheet the table is based on a 100m run - thats 300 feet for the American chaps

At 30MHz its 1.17 dB per 10m (30ft).
At 10MHz its 0.65 dB per 10m.
At 10MHz RG58 is 0.42 dB per 10m.

Other considerations:
The cable is rated for 300V RMS Max according to the spec sheet.
Its only rated for use indoors

[DimitriP]

It's like this:
You can try it.
Worse case you'll blow your antenna tuner or the radios' output stage or the wire will heat up and either smell funny or start a fire, or you'll be left wondering "where is everyboady" because the receiver can't hear anything.
Best case none of these things happen, you saved 50 bucks,  they hear you just fine in JA land and you are left scratching your head "why do we need coaxial cable anyway"?

Necessity is, after all, the mother of invention...

[vk6zgo]

Hi

One thing to note, your Belden table of attenuation versus frequency is based on a fairly short run of wire. You already have lost 6db at 10 MHz based on that table. If your transmitter puts out 1W, only 1/4W is getting to the antenna. The rest either goes up as heat (losses) or is radiated to who knows where. By the time you get to 30 MHz, it's really pretty crazy.

Assuming you are running 10 to 30 feet, all that may not be an issue. If the antenna location involves a run of over 100 feet ... you likely are stuck at 3 MHz.

Bob

According to the spec sheet the table is based on a 100m run - thats 300 feet for the American chaps

At 30MHz its 1.17 dB per 10m (30ft).
At 10MHz its 0.65 dB per 10m.
At 10MHz RG58 is 0.42 dB per 10m.

Other considerations:
The cable is rated for 300V RMS Max according to the spec sheet.
Its only rated for use indoors

The "only rated for indoors" bit is what worried me.
Perhaps if is a special "outdoors" version,seeing it went outside in its original use.

Of course,if it was just cut off when the AT & T guys decommissioned the original service,it has been sitting out in the weather deteriorating.

I'm also a bit concerned about the thickness of the conductors------from memory,they are quite thin.
Googling,some cat5 is shielded,some isn't.
( there is some out in the back room I could check,but there's too much junk in the way)

The main argument (& a very valid one),as far as I can see,is "nobody else uses it!"

I guess KM4FER  could  just try it & "see what happens".

[cdev]

Its my understanding that shielded twisted pair is quite finicky and often results in more noise problems (radiating noise from Ethernet) than UTP.

Also, all of the Cat 5/6 cable I have ever seen uses wire that's fairly thin so I would not expect it would handle power well at all. It likely would melt. OTOH, if used for receive only, with a low noise preamp at the actual antenna, neither that nor any losses (within the range that would be made up for by the LNA) wouldn't matter and the balanced nature of the system would have noise advantages.

[crun]

An old topic, but I recently had reason to consider it, and try to calculate power handling and loss based on skin resistance.
.
TL:DR: P.max 0.8W@10MHz, loss 1.42db/30m@10MHz ~=RG58 

Here is the live smath calculation (updated 8Jan18)

While you might be able to parallel two pairs to make 50ohms, the pairs deliberately have different twists (to reduce crosstalk), so might have poor delay matching. Stripping back some length and matching two pairs with closest twist would be a start.

This suggests that crosstalk and return loss are pretty good:
http://www.ieee802.org/3/10GBT/public/mar03/cohen_1_0303.pdf

I plan to use CAT5 to separately feed each dipole of a multiband fan dipole from a separate qrp beacon transmitter.
http://www.hamuniverse.com/multidipole.html

I have black sheathed for outside use, and also some gel filled underground I could use too.

[David Hess]

The problem is the loss.  Unless both ends of the cable are impedance matched, the high SWR leads to loss greater than the one way loss because the signal is bouncing back and forth.  Twin lead or ladder line has such low loss that it can work with impedance matching at only one end.

The SWR and matching problem also lowers the power handing capability.  It is like running a poor power factor on a power line resulting in circulating current losses.

[cdev]

This works really well I gather for a very low noise active receiving antenna.

Chris Trask on his web site has a design for a remotely tuned magnetic loop (using MV-109 varactor diodes) and a preamp (which is unnecessary because the signal is super strong even without it) that uses cat-5 as its feedline.

[David Hess]

If the signal to noise ratio is high and the power levels are low, then it is not a problem at all.

[crun]

The problem is the loss.  Unless both ends of the cable are impedance matched, the high SWR leads to loss greater than the one way loss because the signal is bouncing back and forth. 

The problem of loss is overstated. The forward loss is much like RG58, and much better than RG174.

A thin wire dipole is 70ohms , basically as mismatched to 100ohms as it is to 50ohms
VSWR is 1.4 - considered good, and mismatch loss is 0.14

In any case 50ohms is not obligatory

• Since someone using a 50ohm unbalanced transmitter would be using a balun, it could obviously be a 2:1 balun
• use a 4:1 Ruthroff balun (wrap the twisted pair around a core) and make the transmitter for 25ohm
  
• anyone using an L or pi match can match to 100ohms as easy as 50.

The SWR and matching problem also lowers the power handing capability.

In the case of 1dB feedline loss, power handling is only reduced by 12% (less for greater feedline loss, as the return power is attenuated before it reaches the source).

[crun]

An interesting thought is that you could use the R.dc to tell you the actual temperature rise of the conductors, and thus find what DC power -> MaxTemp, or let you measure the temperature live when running voice to see when it gets too hot. The cable is rated 60C, but that probably allows POE. RG174 is rated to 85C, so probably the hard limit is 85C in the conductors.

With a single pair a choke could join the far end, or using a double pair 50ohm setup, the whole lot could be DC looped.

Of course as T rises, loss rises. Might add T to the calculations

[deBug]

Yes, it will work at low power.
You need to impedance match both sides. That is pretty easy to do with a couple of ferrite cores.
The turn relationship needs to be 1:1.41 (square root of 100/50) and the reactance of the winding should be at least 5 times larger than the impedance. As this is dependent on frequency there will be a lowest usable frequency for a certain number of turns. If you want to go lower in frequency it generally means you need more turns.
So if you for example use an FT50-61 core make the 50ohm side 10 turns and the 100ohms side 14 turns. 10 turns on FT50-61 gives about 7uH
for 7MHz the reactance is about 300ohm so that is fine down to 7MHz. Use these at both ends.

//Harry

[crun]

My previous calculations were based on ~0.6A current limit, leading to ~1W limit.

So actually I measured the max continuous DC current capacity at 5A for two pairs, 50C temp rise of the actual conductors, using the 4 port resistance and copper temperature coefficient method

From this I calculate an RF power limit, from heating, of >60W RMS per pair @10MHz. Using two pairs thats >120WRMS.

The Vbreakdown rating leads to a limit of 78W Peak per pair or 160W total.

So an SSB transmitter of 100W PEP is probably OK, even allowing for some SWR increasing the voltage.

These are theoretical calculations from DC measurements - see calculation worksheet here

So my statement that its only good for low power looks likely to be bollocks. Next up: actually trying it with RF

[tkamiya]

I would try it.  Ethernet signal is really in VHF region.  On 1Gbit network, 4 pairs of VHF signals are sent and received simultaneously and it is working fine.

In old days, I've used things like TV ribbon cables just fine for HF.  Some of us also experimented with zip cord.... 

Just make sure your antenna coupler can handle this though.  Twisted pair is a balanced line.  Coax output on your radio is unbalanced.  So a balun or specially made antenna tuner would be necessary.  Also, sharp bend and touching metal would cause issues. 

I'm not sure if I'd trust it for 100 watts.  Dipole being 75ohm is in free space.  In reality, it's much lower in realistic height.  So SWR will be high.  You could have breakdown in insulation.

Hum....  Now I want to try this.  Muahahaha..... (evil scientist style ham radio'ing)

[T3sl4co1l]

FWIW, CAT6 is good for almost 100W at DC.  The voltage limit there is driven by safety, but the ultimate limit for the cable isn't a whole lot much more than that (300V?).

I would think a 1-to-4 splitter balun would do fine, maybe not up to 100W at modest frequencies (say ~100MHz), but certainly more than 10s of W.

Attenuation is in the datasheet; it's not as good as twin-lead though.

Tim