Self made test cables for VNA......

[Squarewave]

Hello, I'm wondering if someone can suggest a better cable for my test leads for use with a VNA and SA.

I've been using some I made with RG58, which appear to be ok, but RG58 isn't really the best quality. I have more cables I'd like to make, so was thinking of something like LMR195, probably would need the 195 ultra flex. What have you guys used?
Thanks.

[KungFuJosh]

This is a giant, and potentially very expensive, can of worms.

What frequency range are you looking for?

Also, check out this related thread: https://www.eevblog.com/forum/testgear/cables-connectors-(nsmabnc)-adapters-with-measured-results-for-comparisons/

[Squarewave]

So I realise any losses are calibrated out, but I'm generally using half a metre to 1 metre max. I used crimp/solder SMA plugs. I usually only go up to 500mhz, I've no work in the Ghz range, but perhaps I should consider the VNA does go up to 1.3Ghz.

I'll have a read of that thread, thanks.

[KungFuJosh]

In that frequency range, the M&P cables are good. Gore cables are better. But it really depends on what you want to budget. It's easy to spend a ton of $$$.

[Rubo]

I would (and am) using RG400 cables for anything that I make myself for the lab. You can use connectors for RG58 (it's a close enough fit).
They are on the stiffer side, but have a much better shielding & PIM compared to any of the LMR stuff.

[mawyatt]

Get double shielded cables for more precise work.

Best

[Squarewave]

In that frequency range, the M&P cables are good. Gore cables are better. But it really depends on what you want to budget. It's easy to spend a ton of $$$.

Have you got a link to M&P? I was able to find Gore.

[KungFuJosh]

In that frequency range, the M&P cables are good. Gore cables are better. But it really depends on what you want to budget. It's easy to spend a ton of $$$.

Have you got a link to M&P? I was able to find Gore.

here ya go: https://messi.it/en/catalogue/50-ohm-rf-coaxial-cable/n-male-connectors/hyperflex5/solder.htm

HF5 should be fine in the frequencies you're using, but going up to 7 or 10 gets better performance. I've shown them, and Gore, in the other thread.

[Squarewave]

In that frequency range, the M&P cables are good. Gore cables are better. But it really depends on what you want to budget. It's easy to spend a ton of $$$.

Have you got a link to M&P? I was able to find Gore.

here ya go: https://messi.it/en/catalogue/50-ohm-rf-coaxial-cable/n-male-connectors/hyperflex5/solder.htm

HF5 should be fine in the frequencies you're using, but going up to 7 or 10 gets better performance. I've shown them, and Gore, in the other thread.

Ah yes Messi & Paoloni, I use their UF7 for my HF antenna and patch leads. You know what, I never considered looking at them for 5mm cable (5mm max really for bench top test work), so thanks for the heads up. M&P cables often have better characteristics than the likes of LMR.

[pdenisowski]

One thing to keep in mind about VNA cables is that while you can cal out many cable imperfections, you can't "cal out" a cable that is not phase or temperature stable.  In particular, phase stability (during flexure) is one of the things that you're paying for when you spend $$$ on VNA cables.

Or, you could buy a big roll of tape and secure the cable to the bench that way 

[KungFuJosh]

One thing to keep in mind about VNA cables is that while you can cal out many cable imperfections, you can't "cal out" a cable that is not phase or temperature stable.  In particular, phase stability (during flexure) is one of the things that you're paying for when you spend $$$ on VNA cables.

Yup. This is where the Gore cables excelled compared to all the others I've tried.

[Rubo]

Agreed. And that's why I keep cheking eBay and keep spending money on Gore 

[knudch]

I can not recommend HF5  (Hyperflex 5)

Loses are ok even at 1,3Ghz (which does not matter much for VNA use)
But the SRL / (Structual return Loss) is not meeting its specification 0,3-600MHz > 28 dB
I measure more something like 20-23dB (different lengths 2m - 6m)
Hyperflex 10 meets its specifications of 0,3-600MHz >30dB

Hyperflex 5 seems also sensitive to bending/flexing and don't be "hard" to it, it deforms easily.
HF10 seems also significant better in this respect but is a little un-handy for lab use.

I got some 0,5m cables made of RG142-PUR (5mm diameter) they are very good in respect til bending/flexing
But it is not widely available.

Knud
 

[tszaboo]

For GHz VNA measurements, I use RG402, sometimes it's called SS402. It also what comes with the LiteVNA for example.
It's semi-rigid, so it's not the easiest to work with, and you can absolutely damage it if you bend it the wrong way.
But it's reliable, low loss, and you can solder the connector or the cable yourself. Meaning that it doesn't require expensive blanking/crimping tools.
And you can buy pre-made cables, with high quality stainless connectors, that cost like 1/20th of the price of the typical name brand VNA cable.

[KungFuJosh]

I can not recommend HF5  (Hyperflex 5)

Loses are ok even at 1,3Ghz (which does not matter much for VNA use)
But the SRL / (Structual return Loss) is not meeting its specification 0,3-600MHz > 28 dB
I measure more something like 20-23dB (different lengths 2m - 6m)
Hyperflex 10 meets its specifications of 0,3-600MHz >30dB

Hyperflex 5 seems also sensitive to bending/flexing and don't be "hard" to it, it deforms easily.
HF10 seems also significant better in this respect but is a little un-handy for lab use.

I got some 0,5m cables made of RG142-PUR (5mm diameter) they are very good in respect til bending/flexing
But it is not widely available.

Knud

There's no question that HF5 is not as good as its older siblings, but if it's made well enough, it should meet its specs. Here's one I made:


...and I'm not claiming I did a very good job making it.

It is more sensitive to bends, especially if the connector installation quality is questionable...which I found out when I made questionable connections at the ends. 

Thanks,
Josh

[knudch]

Interesting
This one is 2m long with special N-connector for that cable type, crimped.
CAL-set is Siglent F503
Have another 2m piece ..look the same, also a 6m piece..same.

cables are approx 3 years old and has not been subject to violence.

Knud

[knudch]

This is a RG400 from ROSNOL 1m N-connector crimped.
Teflon, double shielded, inner and shield is steel/copper silverplated.

Much better and ruggedized

Knud

[KungFuJosh]

Interesting
This one is 2m long with special N-connector for that cable type, crimped.
CAL-set is Siglent F503
Have another 2m piece ..look the same, also a 6m piece..same.

cables are approx 3 years old and has not been subject to violence.

Knud

Hyperflex 5 doesn't have an official N crimp connector. That's what I like the least about it.

My cable is with their connectors: https://messi.it/en/catalogue/50-ohm-rf-coaxial-cable/n-male-connectors/hyperflex5/solder.htm

[knudch]

Other manufacturers make the same cable and connectors.

Soldering the center pin in a 5mm cable with foam isolation, is risky...very little heat melt the foam and offset the center conductor.

Crimping is always preferable unless we talk about teflon cables.

Maybe there is fluctuating quality...seen before amongst some European cable manufacturers in there competition for lowest attenuation.

They use low density foam (you can read that from the velocity of the cable, HY5 is 87%), low density makes it vulnerable to handling both in production and user handling.
Of course that gives lower looses.

Knud
 
 
 

[KungFuJosh]

What connectors fit it? HF5 is 5.4mm. I've not found anything that fits it.

Risky maybe, but it worked better than those 3rd party crimp connectors. Joking aside, it's a fast solder. It might take some practice, but really not bad.

[knudch]

Ecoflex 5 from SSB electronic (Germany)
Belden H155 (Dutch)
LMR240 will also go in this category, maybe with a different outer ferrule
 
Connectors which fits these cables will also fits HY5
Same design multi core inner conductor, braid + foil shield, foam dielectric

It is jungle to navigate in which connector and crimp tool goes with which connector.
Some times is the connector housing the same but different center pin and different outer ferrule.
Luckily I have a supplier close to me which has a great variety of parts/cables and knowlegde in this area.

But it is interesting that you have clearly better results from  HY5 than I have
I could be tempted to order a few meters and connectors...not very expensive here..out of curiosity

Knud   

[KungFuJosh]

That's funny. Despite my better results, I still want to try the crimp connectors. 🤦‍♂️

[knudch]

The connectors is definitely not the difference over a such wide frequency range.
The cable is the guilty
So be happy that the HY5 works you

May I got the last meters they produced before the weekend


But in the low price segment the competition is hard....focus on low attenuation is a highlight "mine is faster than yours"
For VNA use this is secondary

Knud   

[KungFuJosh]

I disagree. I did a terrible job originally with the connectors, and didn't get those results until replacing them.

The cable meets spec for me, and not for you. If it was the cable, it would not meet spec for either of us.

[knudch]

I fully understand your deduction
You made crappy connection got poor results, change to good connections get good results

Vice versa I get poor results...I must have made crappy connections

What are against that conclusion:
I made 6 identical bad connections ....I am very good

I have seen more poor cables than poor connections

All my other crimped connections (close to 100) are ok

I believe it is more likely that I have got a length of "monday" cable


Knud