Low Loss Coax for Test Equipment

[RCHRDM]

Can someone please recommend the best low-loss coax for connecting RF test equipment together (with N connectors)?

Richard

[beanflying]

The Best for what will be the obvious question from others?

Rough idea of any power requirements and frequencies would help narrow it down. But from your Type N's I gather to are up and over 1GHz?

[TheUnnamedNewbie]

What frequencies are you working at? What do you consider best? Phase stability? Loss? Robustness? Flexibility?

At work we use mostly H+S Sucoflex 103 and 102 for the generic low microwave stuff. 102 has slightly higher loss but is a thinner, more flexible cable. Both are available as stock product with N connectors.
For phase stable stuff we (like pretty much everyone, I guess?) use Gore Phaseflex.

For the very low level signals we tend to stick to RG316 and RG58 for most stuff. RG316 is tiny and so flexible.

[tautech]

You might find this chart useful for guidance:
As attached.

[mzzj]

Luckily OP just wanted best without price constraints as Gore phaseflex cables are about 1000 usd per piece 

tautech, link, not image 

[beanflying]

I remember speccing Teflon Encapsulated Goretex O'rings on a few specialty jobs in my former life - so glad I was selling not buying 

[TheUnnamedNewbie]

Luckily OP just wanted best without price constraints as Gore phaseflex cables are about 1000 usd per piece 

Where can I get those? We pay a heck of a lot more per cable than 1000 USD ;p 1000 bucks a cable is what we spend on a lot of our non-phase-stable cables.

[Berni]

Sure Gore Phaseflex or Huber+Suhner Sucoflex

But the question is if you are prepared to pay the associated price tags.

If you want the absolute best in performance you will need rigid coax bent into the form you need.

[mzzj]

Luckily OP just wanted best without price constraints as Gore phaseflex cables are about 1000 usd per piece 

Where can I get those? We pay a heck of a lot more per cable than 1000 USD ;p 1000 bucks a cable is what we spend on a lot of our non-phase-stable cables.

Mouser
https://www.mouser.fi/Gore/Wire-Cable/Cable-Assemblies/RF-Cable-Assemblies/Phaseflex-Series/_/N-bkrix?P=1yym9e5Z1z0zkrm

2.92mm connectors are lot more spendy tho

[radar_macgyver]

If you want the absolute best in performance you will need rigid coax bent into the form you need.

For phase stability, yes. One can get lower loss, but using Heliax. It's awful stuff to work with though - has a will of its own and refuses to bend the way you need it to. Of course, there's also flexible waveguide for even lower loss.

[RCHRDM]

Thanks for the replies!  I am connecting test equipment together over about three feet or less (e.g., signal generator to spectrum analyzer or oscilloscope).  I am checking the calibration of the spectrum analyzer and oscilloscope at present.  Frequencies are below 500 MHz and power might be up to 10 dBm.

Richard

[RCHRDM]

I am working mainly with ham radio equipment.  "Best" is lowest attenuation.

Richard

[RCHRDM]

How about RG142?

Richard

[Berni]

For those sorts of things good ol RG58 tends to be plenty good enough, as long as its good quality cable and not some cheap Chinese crap that has only a few strands around it as shield. At these sorts of distances and frequencies the attenuation is almost nothing. That power rating is also handled just fine by any cable out there.

But if you do want a nice low loss cable a commonly used one is LMR240. Gets you less loss and handles things into the GHz range much better, but will be a good bit more expensive while having the poor flexibility of a decently sized garden hose.

Cables that get you low loss, high frequency operation and high flexibility all at once tend to be more expensive proprietary things from people like Gore

[RCHRDM]

I checked.  RG-213 has lower loss than RG-142.

Richard

[mzzj]

I am working mainly with ham radio equipment.  "Best" is lowest attenuation.

Richard

I have some cut-offs of "perfect" cable for you:



oops. any other requirements than lowest attenuation? 

[don.r]

I am working mainly with ham radio equipment.  "Best" is lowest attenuation.

Richard

I have some cut-offs of "perfect" cable for you:

<img>

oops. any other requirements than lowest attenuation? 

A family of mice could make a nice home in that cable.

[mzzj]

Heliax also comes in 5" diameter but I have never seen that 

Ham radio fellas are  building notch filters from similar cable to the one in picture ie ~2" foam insulated.
For a good notch you need really low loss cable. 

[RCHRDM]

Wow!  Nice cable!

Richard

[tautech]

Thanks for the replies!  I am connecting test equipment together over about three feet or less (e.g., signal generator to spectrum analyzer or oscilloscope).  I am checking the calibration of the spectrum analyzer and oscilloscope at present.  Frequencies are below 500 MHz and power might be up to 10 dBm.

Richard

Up to 1 GHz these 1M long ones have been fine for me:
https://siglentna.com/product/bnc-cable/

Sure you can buy Met grade cables but unless you're chasing every 0.1dB why ?

[RCHRDM]

I am doing something kind of basic, which is to check the accuracy of my test equipment.

I have an Agilent E4400B signal generator.  I checked its output with an HP 432A for which I have three working sensors.  The E4400B seems to be accurate when I measure 2 dBm at 50 MHz and all three sensors agree.

When I make this same input to an HP E4406A Transmitter Analyzer, the measurement is about 1.95 dBm on its spectrum analyzer.  This is pretty good, I should think.  The other day when I tried this, I did not get as good a result and wondered if it was the coax cable.  I have a couple of other instruments (LeCroy WaveRunner 44Xi-A oscilloscope and a JDSU JD747A Signal Analyzer) that I would also like to check. 

I just tested again, and the spectrum analyzer measurement on the JD747A is way off (like 12 dBm!) presumably because I don't know how to adjust the settings correctly.

I am not an electronics engineer and have the equipment for ham radio work, mainly.  I am still learning!

Richard

 

[tautech]

 

I just tested again, and the spectrum analyzer measurement on the JD747A is way off (like 12 dBm!) presumably because I don't know how to adjust the settings correctly.

Yes just might be input attenuation settings.

Absolute amplitude measurement precision is another rabbit hole most don't want to go into, if I'm getting within a dB or 2 it's near enough to tell me I'm thereabouts.

Even Sharirar on the Signal Path checking accuracy to 3 GHz on a SVA1032X using reasonable quality SMA cabling and a SMA-N adapter was getting ~0.5dB losses.

Grab a few cheap BNC cables using the chart I posted earlier to give guidance on coax type and frequency specs and for your work you'll probably find them quite OK.
I got a customer 50 odd BNC cables just from an Aliexpress supplier and apart from their crimps being a little 'how ya doing' suited their needs well.
You don't need to spend big bucks for 90% of uses.

[RCHRDM]

Thanks!  That's good to know.  I don't need to obsess about it!

I learned this about RF wattmeters (like the Bird 43) that have about 10% accuracy.  As you say, that is good enough to know if there is a problem or not.

I have a few Chinese cables with N-connectors.  The crimps failed on a couple of them and the connectors fell off!  I'll get a few good quality cables.  I figure it is also a good idea to avoid adapters (e.g., N to BNC) and go straight to the device under test, if possible.

Richard

[tautech]

Thanks!  That's good to know.  I don't need to obsess about it!

I have a few Chinese cables with N-connectors.  The crimps failed on a couple of them and the connectors fell off!

 
Yes this you have to keep an eye on as there's a lot of subtle variability in measurements and a just poorly adjusted crimper can let you down here.

I figure it is also a good idea to avoid adapters (e.g., N to BNC) and go straight to the device under test, if possible.

Yes in the perfect world we'd all have the correct cable with the right ends but again for hack around work you will need some adapters and cheaper cabling.
Don't overlook all quality connecters/cables have limited mating life cycles and such good stuff needs be reserved for bestest work.
For instance the front panel bulkhead connector into quality equipment is often preserved with a sacrificial adapter, say a N-SMA then you are able to work with SMA cables that are quite cheap and typically much more flexible.

[RCHRDM]

Good idea!  I have a miniVNA Pro, which is a small VNA, as you may know.  It has SMA connectors.  I bought some short cables (1 foot) with BNC or UHF connectors for testing.  I thought this would save the SMAs from getting yanked around.

I should do something similar for the bigger test equipment, as you suggest.

Richard

[RCHRDM]

Is an "N SMA" a cable with and N connector on one end and SMA on the other?

Richard

[tautech]

Is an "N SMA" a cable with and N connector on one end and SMA on the other?

Richard

Yes that or an adaptor. Do your homework as there are many different types of SMA ends and for a while they are quite confusing.
I got this SMA adapter set recently and it covers most types:
https://www.aliexpress.com/item/32915423924.html?spm=a2g0s.9042311.0.0.14294c4dtHo85K

Some reasonable quality SMA cables here:
https://www.aliexpress.com/item/32854810729.html?spm=a2g0s.9042311.0.0.14294c4dtHo85K

[RCHRDM]

Thanks.

So, if you are connecting an instrument with an N front panel output connector to another instrument with an N input, do you use adapters to run an SMA cable between the two instruments?

Richard

[tautech]

Thanks.

So, if you are connecting an instrument with an N front panel output connector to another instrument with an N input, do you use adapters to run an SMA cable between the two instruments?

Richard

Yes that's where I slowly heading with my gear, BNC for much but slowly moving to SMA.
Depends on what you need to do. I have 6 GHz N cables and a selection of N type adapters.
Really it's more about the mating cycles lifetime of the front panel connector on your instruments plus how high frequency you're working with that dictates cable type and to some degree the quality of your adapters.

We try first to limit the wear on the connectors on our instruments so pick a connector type for the sacrificial adapter and a cable/connector combo for your everyday use.

The sky's the limit for the cost of high end stuff however up to a GHz or so you might just measure the difference using cheaper stuff.

[RCHRDM]

Thanks.  Good advice.  I am working in the ham band frequencies so not more than 500 MHz most of the time.

Thanks for the great conversation today!

Richard

[tautech]

Thanks.  Good advice.  I am working in the ham band frequencies so not more than 500 MHz most of the time.

Thanks for the great conversation today!

Richard

No problem.
One must remember this cheap stuff is just for LF everyday usage whereas Meteorology grade gear is in a different league. 

[DaJMasta]

Cheap ebay cables can do a reasonable job with that, especially if you're not worried about high power or absolute minimum loss.  If you are worried about the higher quality connectors built onto the instrument, I'd suggest connector savers, basically good quality male to female adapters that you put on the instrument before your cable - that way you minimize the wear cycles on the instrument and prevent any bad connector geometry/torquing from damaging the built in connector.

[mzzj]

Grab a few cheap BNC cables using the chart I posted earlier to give guidance on coax type and frequency specs and for your work you'll probably find them quite OK.
I got a customer 50 odd BNC cables just from an Aliexpress supplier and apart from their crimps being a little 'how ya doing' suited their needs well.
You don't need to spend big bucks for 90% of uses.

My luck with chinese /aliexpress connectors and cables has been pretty much mixed lot. SMA center pins not making contact, plating peeling off(internal shorts, yikes), very wobbly BNC connectors etc.
Oversize center pins that ruin the front connector on your expensive test equipment is the last thing you want..
I have also got decent and usable connectors but I haven't bought enough of anything to say who would be reliable vendor.

Unless you want to get in the cable&connection inspection I'd suggest some low-cost brand-name cables and connectors at least for most commonly used test equipment.

[BravoV]

Cheap ebay cables can do a reasonable job with that, especially if you're not worried about high power or absolute minimum loss.  If you are worried about the higher quality connectors built onto the instrument, I'd suggest connector savers, basically good quality male to female adapters that you put on the instrument before your cable - that way you minimize the wear cycles on the instrument and prevent any bad connector geometry/torquing from damaging the built in connector.

+1 

Consider to equip all those N connectors at the equipment with sacrificial adapters, either native N male to N female, or N male to SMA/BNC, you wont regret it for long run

Example of the right angle N sacrificial adapter that I put at my SA input, right angle adapter is handy as the attached cable can lay down in front of the test device if its placed on table, and connected by cable to the adjacent other test device, instead of protruded out and hanging perpendicularly to the device panel, and at some cases I think the cable is mechanically stressing the connector, especially on stiff/heavy ones.

[Berni]

I agree with BravoV. 90 degree adapters are quite nice connector savers for SAs because if you are using big hefty cables its often more convenient to have the cable go off to the side, just don't rotate the 90 degree adapter when using it, make sure its tightened down and have the cable rest on something.

In my case i also have a inline DC block hanging off it since my SA can't handle any DC at all on its input.

But overall i standardize on SMA cables because SMA is a nice small compact connector that can handle pretty high frequencies, also cables and adapters are easy to find for it. To get to SMA from test equipment i have cables with SMA on one end and N or BNC on the other. For connecting to a circuit i have a box full of cheap Chinese SMA board mount connectors that i solder into my circuit permanently and then just connect my coax to that.

Using adapters is not that bad. Sure you want to use less adapters if you can, but unless you are working at 10GHz and up you are not going to see any signal deterioration from an extra adapter or two. My scopes pretty much have a BNC to SMA adapter permanently live on one of the 4 channels so that i can just plug a SMA cable right into it.

[BravoV]

I agree with BravoV. 90 degree adapters are quite nice connector savers for SAs because if you are using big hefty cables its often more convenient to have the cable go off to the side, just don't rotate the 90 degree adapter when using it, make sure its tightened down and have the cable rest on something.

In my case i also have a inline DC block hanging off it since my SA can't handle any DC at all on its input.

Yep, especially the DC block is permanently attached to the SA, instead of sticking out from the SA's panel, that is prone to accident like unintentionally bumping it, and break/bend the SA's connector  , using a 90 degrees right angle adapter will minimize that risk.

Too bad there is no animal such as a N male to SMA female adapter, but with 90 degrees orientation, that will be perfect. 

[Berni]

I agree with BravoV. 90 degree adapters are quite nice connector savers for SAs because if you are using big hefty cables its often more convenient to have the cable go off to the side, just don't rotate the 90 degree adapter when using it, make sure its tightened down and have the cable rest on something.

In my case i also have a inline DC block hanging off it since my SA can't handle any DC at all on its input.

Yep, especially the DC block is permanently attached to the SA, instead of sticking out from the SA's panel, that is prone to accident like unintentionally bumping it, and break/bend the SA's connector  , using a 90 degrees right angle adapter will minimize that risk.

Too bad there is no animal such as a N male to SMA female adapter, but with 90 degrees orientation, that will be perfect. 

Exactly, especially since my SA sits in a 42U rack where its possible to walk by it.

I do sort of have such an right angle "adapter". Its a 50cm piece of nice thick microwave cable that has a 90 degree N connector on one end and 3.5mm SMA on the other (I have some 3.5mm female to female bulkheads to fit on the end of it). That one sort of permanently lives on my Gigatronics 20GHz synthesizer. I think i got that as part of a pile of old RF cables that looked fancy enough to buy.

[mzzj]

Too bad there is no animal such as a N male to SMA female adapter, but with 90 degrees orientation, that will be perfect. 

Never say never
https://www.centricrf.com/clearance/clearance-adapters-0-18ghz/c3553-sma-f-to-n-m-right-angle-adapter-6-ghz-vswr-1-15-brass/

[BravoV]

As pointed out by others, those various adapters can be cheap ones, but for the sacrificial connector, spare no expense, buy quality top tier brand as they're much-much better than those no brand Aliexpress ones.

Examples of N male to male adapter.
Left one is from Aliexpress vs right Amphenol, I just leave yourself to judge. 

[BravoV]

Too bad there is no animal such as a N male to SMA female adapter, but with 90 degrees orientation, that will be perfect. 

Never say never
https://www.centricrf.com/clearance/clearance-adapters-0-18ghz/c3553-sma-f-to-n-m-right-angle-adapter-6-ghz-vswr-1-15-brass/

Wow ... thank you ! 

[mzzj]

CentricRF webshop looks like treasure trove for anything RF related.
Digikey seem to also have pretty good selection of SMA cables from 10usd to 1000usd

[Berni]

Too bad there is no animal such as a N male to SMA female adapter, but with 90 degrees orientation, that will be perfect. 

Never say never
https://www.centricrf.com/clearance/clearance-adapters-0-18ghz/c3553-sma-f-to-n-m-right-angle-adapter-6-ghz-vswr-1-15-brass/

Oh i might pick up a few of those. Seams to be a reasonable balance between quality and price.

Since yeah be careful with crappy Chinese RF connectors as they can be bad enough to damage connectors they mate it. For microwave stuff i tend to buy old used high quality stuff and inspect them under magnification to make sure they don't look damaged. I don't have the budget to buy only brand new fancy high quality RF connectors.

[BravoV]

I checked.  RG-213 has lower loss than RG-142.

Richard

If you opted for RG-213 cable, which is quite thick & stiff, make sure you have the one that has double shielded as you want minimal lost at RF.

Even at top branded ones, depends on the variants, some have single braided shield only.



Here example of my RG-213 equivalent cable manufactured by Leoni, watch the dual shieldings, the outer 1st layer with braided wires + 2nd layer inner metal foil.

[eb4fbz]

RG-223 is my usual choice for <1GHz lab cables. You will allways have some loss, so i prefer to focus on cable flexibility and shielding quality.

For higher frequencies or precise VNA phase measurements i stick to H+S Sucoflex.

[nctnico]

RG-316 is easy to work with and good up to 1GHz.

[RCHRDM]

RG316 seems to have higher loss that RG58.  Not sure my reference is accurate, though.

Richard

[RCHRDM]

I saw some cables on eBay using LMR-240:

https://www.ebay.com/itm/LMR240-Silver-BNC-MALE-to-BNC-MALE-Coax-RF-Cable-USA-Lot/143067828057?var=442029207133

That seems to be low-loss cable.  I kind of prefer using BNC to an N-to BNC adapter on the instrument.  It seems like it would be faster to connect a BNC rather than an SMA.

Richard

[tkamiya]

I buy patch cables from a local cal-house.  I think they were 10 dollars a piece.  Cable itself is TimesMicrowave 195.  I'd be surprised you can actually measure loss at frequency less than 500MHz.  I'd be glad to sell you what I have, since I can easily buy more.

[RCHRDM]

Thanks for the offer!  What connectors?  Lengths?

Richard

[tkamiya]

Connectors are both N.  80 inches long.  I have many other types, too.
Let's take this offline, since it will be no interest to others in the thread.

tkamiya9(at)yahoo(dot)com

[eb4fbz]

I would avoid solid core, foam dielectric or aluminum film shield cables for bench use. They degrade with use and multiple bendings, unless they are those expensive VNA grade cables from Gore.