HTML-код
- Опубликовано: 17 сен 2022
- In this episode Shahriar discusses the engineering challenges associated with making good RF & mm-Wave cables. The significance of the connector, cable frequency range, loss, phase/amplitude stability & temperate are all examined from a theoretical & practical aspects.
The stability of Junkosha cables are demonstrated up to 50GHz through S11 & phase measurements. The Junkosha MWX7 series cables offer optimized PTFE dielectrics as well as some of the best phase/amplitude/temperature stable operation:
www.junkosha.com/en/products/...
www.TheSignalPath.com
/ thesignalpath
www.Patreon.com/TheSignalPath 
Наука
A characterization of common aliexpress SMA cables would be very nice. Many hobbyists uses them for RF assemblies.
I would say that Aliexpress cables are a giant step from my own homemade cables.
I've had good experiences with $5-$15 SS405 cables. IIRC their specs only go up to 18GHz, but I've only used them at
I second that - it would be very nice to have some experiments showing what happens to components when you exceed their ratings. For example, if I have a BNC connector that's rated to 3 GHz, how bad will my signal be if I use it at 5 GHz? Will it just be some attenuation and maybe a few ps of phase shift, or will the signal be unintelligible?
@@charlesmcanany6806 Signal could be perfect at one calble length with that connector but change the length a half inch it could drop into the noise! Reflections from mismatch can be trecherous!
Nice review - I didn't know how good those Junkosha cables were - impressive.
Nice video. I trick I do to test return loss stability is Data->mem then display Data-Mem, the cable will be flat across the band at the system noise floor. Then bend the cable and you see the true residual. You can also see the relaxation this way by bending and hitting data->mem again and watch S11-Mem move as the PTFE inside settles into it's new shape. In your video, you see the rise in S11 on the trace initially (which should be at the noise floor directly after cal) likely due to a slight cable movement/relaxation when you did the OSL cal, maybe due to moving slightly the cable after applying the load. Or just a bit of temperature drift. You can test the phase stability of a cable by putting a short on the cable, Data->Mem and Data/Mem and then you can see the two-way phase change. But, to quote myself: cables are like dogs, either they are bad, they've been bad or they are going to be bad, and if they are good, they only stay good with great care.
Shahriar, I always love the stuff that you produce but this episode in particular was worth it’s weight in gold for me. Thank you so much for doing it!
I didn't think I would be interested in this topic, I got hooked. The cable testing was fascinating. Thanks!
Excellent video, thank you! This kind of information is not common to stumble across, so thank you for the opportunity to learn some new things. Cables, connectors, switches, all that stuff has obscure but important characteristics.
Great video! Would definitely be worth sharing to coworkers .
I've noticed this effect before, the shifting of S11 values as the cable moves. Thus far I've been working under the assumption of that the shielding is not perfect, and as the cable moves the shielding provided differed and thus affecting the amount leaking out of the cable, which, in terms of EMI, bothered me to no end. Nice to know that it could be due to phase difference... I wish I had cables can take such levels of abuse...
You have a truly great gift for explaining complex things with incredible clarity. Thanks for making this video--I'll probably never use a cable like this in my little hobby lab but it's cool to learn about it here.
Excelent content. Very instructive.
When I was a newbie graduate, I have had the opportunity working with VNA up to 50GHz in UWB flat antennas. I used torque controlled key for connectors and very expensive coaxial cables. I was warned that I couldn't bend them below a certain bend radius because once bended beyond this limit, the cable could have an unrecoverable impedance distortion, so every morning I was extremelly careful doing my setup lab, I could spend several hours doing the setup and considering the good temperature for the lab too.
The "reward" of doing this, was the repeteability, accuracy, precision and extremely high similarity between EM simulations and measurements for S parameters inclunding phase too, which is one of the magnitudes that many researchers omit in many articles because they are not able to reproduce them, they even present measurements with parasitic resonances that do not have in EM simulations and they do not ask why this discrepancy. I remember that once I had some parasitic resonances around 7 GHz with a connector that were not in the FTDT simulation and when I removed the arms of the SMA connector which was in transition with a microstrip, the resonances disappeared. So what I learned is that before drawing conclusions from a measurement, you have to double check or triple check you lab setup. The same holds for câbles and transitions, in fact they tend to be the root cause if measurement discrepancies due to multiple reflections, insertion loss and impedance mismatches.
Indeed. A connector that is not properly seated can cause all sorts of things, including faster-than-light neutrinos. ;-)
@@EdwinSteiner LOL!
This was a great walk-through! I went ahead and bought some Junkosha cables for my VNA setup after watching.
Eye opening!!! Thank you, Shahriar!!
Very interesting to see the amount of engineering put into making these wires...thanks a lot for the video!!!
Cool video!! For my little basement lab I use a lot of RG-402 semi-rigid and SMA connectors. It's extremely easy to solder new assemblies and it has pretty good performance for sub-6G and short runs, and it's cheap in bulk. I also have some LMR-600 with N connectors going up to the discone on the roof.
I'd really love to see a tour of where microwave connectors are machined. The precision that goes into higher end instrument and metrology grade connectors (including precision slotless) is really fascinating.
I wonder if that level of production is fully automated or if there is a tool room with gray beards to maintain that level of precision. Maybe KS or another company that sells connectors could give you a tour that can be recorded and watched by us as well?
@aaaaa Nice! What kinds of connectors are those? Would they ever consider uploading a video tour of the manufacturing process for marketing? That would be a first; there are no videos on this subject from what I can gleam online.
I might get a chance to visit Rosenberger at some point, would be cool to see this stuff
If you want to get an impression about how those things are machined, search on YT for the channel "Machining and Microwaves".
@@gammarayflash1170 I've been watching his channel for a year or so now. I've actually commented about this exact subject (precision connectors) and he noted that the precision involved has traditionally required manual machining. He also noted that he has never seen a shop making these connectors... which makes me even more curious!
Find the microwaves and machining RUclips channel. You’ll love it.
Excellent video as always.
A great video. I cal spectrum analyzers and power sensors at a US Army Reference Lab. We have Gore cables up to 50GHz. I learned a lot from this video that I didn’t know, despite doing the work for many years. Thanks!
I can personally attest to Junkosha's quality as well. Fantastic product and superb customer service.
Really nice demonstrations, that Junkosha cable is excellent. What you are measuring are the so called stability and recovery of the cable, both for magnitude and phase. For evaluating these parameters, I prefer connecting a short at the end of the cable under test.
Me: Hey engineer, (literally any question)
Engineer: Hmmm, well it depends.
Indeed an insightful video!
Informative as always. Thank you very much.
Very useful video for me to learn.Thank you very much!
Thank you, learned something new.
great video! was searching for this type of video explaining how to choose the right cables.😀
impressive....good insight into selection of cables..that helps..!!
Very interesting stuff...cheers.
Impressive stuff. Thanks.
Great stuff
Thanks for the overview, Shahriar! While I've known for a while that bending high-bandwidth cables is a no-no, seeing it illustrated so starkly was enlightening. As someone who's accustomed to overblown claims about cable performance (looking at you, audio industry!), I might have skipped clear over a brand like Junkosha on account of suspiciously high cost.
Awesome. Thanks.
What are the other constants in the formula at 5:44, K2, Kb, etc? I'm curious what this formula looks like plotted over ranges, there's probably a paper or something someone can link in?
Many years ago I upgraded all of my bench test cables to "RF Orange" series MegaPhase brand. The RF Orange cables are rated to 110GHz. Absolute overkill for the type of radios and repeaters I work on. Most top out at 70cm amateur band with the occasional 23cm band radio. I didn't upgrade to 110GHz cables and 18GHz connectors because I needed phase stable cables but because I wanted something that was ultra durable with repeatable results. After several years of use they perform as good as the day I bought them. I can't begin to imagine how many thousands and thousands of times they been flexed and they still perform like the day I bought them. Had I used "amateur radio" grade cables they would be worn out long ago.
I lucked out a couple years ago and got about a half a pallet of new never used MegaPhase Warrior series cables surplus at a military auction for pennies on the dollar. I use these for my amateur radio use on my 2.4GHz and 5.6GHz radios. I can't begin to imagine how much uncle Sam paid for these crush proof cables.
Mike KC3OSD
OMG can I buy some from you for my new portable shack setup!?!
I was going to ask if Shahriar had any MegaPhase cables...I understand that every centimeter of RF cable used in the LIGO project is made by them.
In my self studies of HF electronics some 20 years ago i read that the electric current in a coax with higher frequencies starts to dissociate from the wire and the signal itself mostly travels inside the dielectric as a wave. Therefor kinks and twist are to be avoided to not introduce changes in diameter to the dielectric medium.
I personally seldomly meddle with anything above 100MHz so i am a pedestrian in comparison to you ;-)
But i still get a lot out of your videos. I mean you are more or less the only channel i've encountered so far that deals with this kind of stuff.
My understanding has always been that RF cannot travel through copper, and that it travels through the dialectric.
I wonder if we're misunderstanding? The voltage develops across the copper, but the wave is electromagnetic and travels in the dialectric - Something like that?
@@digitalradiohacker Yes, although this happens afaik only in RF above 300+MHz.
Below that and above 30 MHz, you'd observe more of a skin effect, where charge transport mostly happens on the outer 1% of the inner conductor. (That would be why you see cables for that specific frequency range coated with i.e. silver. Does not make sense for shortwave or microwave.)
That being said, my knowledge of that stuff is almost 25 years old and from books referencing who knows what sources (DARC/ARRL training material).
what about F and BNC and other types for under 20GHz?. I would be interested in seeing that as well..
Thanks, learned new manufacturer today. I wonder if there are RF cables with non-round conductor geometry? For example in high current (tens of amps at tens kHz) AC metrology some devices use arrays of flat ribbons to reduce inductance and losses.
The cable materials are generally similar (unless you need high stability cables). Low density ptfe dielectric, woven shielding with possible steel outer armoring.
Lower frequency cables can be larger in diameter due to physics, and most F and BNC connectors also use ptfe in the connector as opposed to 3.5, 2.4, 1.85, etc which are air dielectric connectors.
There are precision variations of these "lower frequency" connectiors such as F, N, BNC, which are machined to tighter tolerances, use better materials and platings, and can even be air dielectric. However, the geometry of the connector is still the prime factor in determining the impedance vs. frequency.
@@xDevscom_EE Illya it’s funny that you commented, I was wondering as I watched this whether you had ever used them. I’m hopping that the prices are somewhat less expensive than the Gore phaseflex cables.
I was kinda surprised to see that you looking at an amplitude of S11 in order to assess a phase stability of a cable. Wouldn't be arg(S11) i.e. phase of the reflection coefficient be a more appropriate metric?
Love this video. What about twisted pair? How do they behave?
Excellent, i would have stressed that another key parameter is the connector itself from one manufacturer to another one the connector VSWR can have a big impact. Often user forget to look at the max frequency of the connector, usually they look at the coax frequency.
We also ended up buying a bunch for our lab instead of Gore cause of this video ^^ we already had Gore so its not like anyone's missing out but yeah, the Junkoshas are a lot more flexible it seems.
Isn't a type of sensor that measures deformation, and the sensor itself is a cable? I cannot recall. I wonder if it uses something like this. I am not talking about strain cells.
Very interesting video, the kind of things one never thinks about, but when present it clicks as a logical thing.
Thank for this video, for me it's one of the more interesting ones you have put out in a while (not that the others aren't 😀). I wouldn't have minded this video being a bit longer in fact.
One question what remains for me is: given the fact that higher bandwidth cables give a higher loss at the same frequency, why would they ever make cables that are higher bandwidth than the connectors they are fitted with (like at 10:56)?
Sometimes it's for mechanical reasons, you might use a thinner cable for improved flexibility (even if you don't need the higher BW) as long as the higher loss is acceptable.
The solder-in probes I'm developing have SMPM connectors to mate with the probe tip but SMA at the instrument side, but are using Koaxis KF047 cable that's rated to 110 GHz. SMPM is good to 40 GHz, SMA to 18 GHz, and my probes top out around 8.5 GHz. But I need an extremely thin, flexible cable to avoid putting excessive forces on the fragile solder joints and am willing to sacrifice insertion loss (the 3-foot cable has about 3.6 dB of insertion loss at 10 GHz) to get that. You can correct for the cable loss with a combination of equalization in the probe head and DSP de-embedding on the acquired waveforms.
Junkosha was a manufacturer of moulded plastic products before they started making cables. They were the first company to make meltable fluoropolymers, and are still one of the industry leaders in this product category.
When you say the phase and amplitude shifts can be caused by mechanical movement of the cables at "high frequency", can you define above what frequencies we have to be concerned??
I'm assuming the price is in the range of "If you have to ask, ya can't afford it".
Tho it probably wouldn't matter if you work at these frequencies and needed that level of performance.
Modest, off the shelf assemblies range from roughly 100 to 1000 USD, depending on length and type. Pasternack is one such vendor, and their website has prices, if you're curious.
Here's me messing around with coax splitters and phasing harnesses at 500MHz and below, and you come out with a video showing off 100GHz. 🤣🤣
دمت گرم مهندس ، استفاده کردیم
Спасибо.
Can you comment on the use of different diameter cables in series?
what is your experience with rigid solid copper cable ? i have to say i like those types of coaxials quite a bit
nice cables indeed
Hi Shahriar it might be interesting t do a video n HDSDI cables they have a double screen system that only joins at either bnc common ground end.a special BNC crimp is used but it still connects to standard BNC on the equipment used. a kind of double grounded screen ,Up to 6Ghz on 4k
The extra shield layer can be there to further mitigate electromagnetic interference or provide additional physical protection to the main coaxial cable inside for use in harsh environments, or it can be there to remove extra cash from uninformed prosumers. It does nothing for the signal transmission inside.
Why didn't you set the fieldfox sweep to Logarithmic?
today lot of manufacturer use “phase stable” but they all have different phase performance versus bending. W.L. Gore has some of their assembly with a phase performance other no. If you look at the W.L. Datasheet versus Junkosha regarding phase stability phase performance there will be surprises. Yes Junkosha are great but WL Gore to my eyes are the best.
We used hollow Heliax for all sorts of communications up in the Arctic (not confused with waveguides). The main commercial cable line was regular old 500. Surprised you didn't mention hollow cables.
How is a hollow cable not a waveguide?
@@stargazer7644 Skin effect.
@@schitlipz We call that waveguide.
@@stargazer7644 It's called lightening the load (mass) and expense. Hollow cables are not all waveguides. Every single "old school" TV cable strung from pole to pole is hollow, prior to the dropline. Hollow cables are exceedingly more common than waveguides.
@@schitlipz What you're describing is coaxial hardline. It is not hollow, it is "hard" coax with a solid metal outer conductor. It is constructed this way for lower loss. It has a center conductor, a dielectric and a solid aluminum or copper shield on the outside of it. If the cable is hollow (has no center conductor) then the only way it functions is as a waveguide. While most cables use a plastic or foam dielectric, some coax uses an air dielectric by having a spiral plastic support that keeps the center conductor centered in the cable. This is not a "hollow" cable. It has a center conductor. It is coaxial cable. And the "old school" TV cable hardline on my pole does not have an air dielectric. That stuff is ridiculously expensive. It has a plastic dielectric.
I'd be delighted to see how to actually fit an SMA connector to RG316, all the "tutorials" I have seen are hilariously bad. And proper manufacturer supplied guidance does not seem to be there.
It would also be interesting to see how the proper "machined center conductor as center pin" SMA connectors for .141 semirigid look like and work.
Fascinating. Now I understand why audiophiles spend so much money on cables! 😁
Heh, I use #14 ROMEX wire for my speaker cables! Solid copper conductors, "acoustically stable" with no noise introduced by the individual strands in a more flexible cable rubbing against each other during those awesome bass guitar solos 🙂 See, my golden ear is influenced by Science. Or so you can infer from the rapid hand-waving that accompanies my explanations.
Anyone else wish for part 2 of this video; maybe using some RG-58A/U cable, with the same tests? Maybe even adding in some thermal variation to see how it wiggles around. Just a run a coil of it and tape it down to the 3D printer bed and run the temperature up and down..
nice sarcasm.
In the comment section of Signalpath I assume this is sarcasm 😄
@@lmamakosromex is rubbish for speaker cables. I only use the highest quality straightened coat hanger wire here. You can really hear the difference!
Fascinating video. Didn't ever think about temperature stability of the dielectric. Was this video sponsored?
No, the video is not sponsored.
@@Thesignalpath Thanks for clarification! Learned a lot from the video.
in practical terms, what sort of applications is equipment in these bandwidths even used for? telecom? network backbone? radar? this is beyond my scope of exposure in industry.
Do these Magic RF Cables can withstand the sheer output power of The Impossible Amplifier?
We are quickly approaching the regime where dielectric waveguides (i.e. optical 'fibers') might replace electrode cables. I keep expecting someone to release materials appropriate for that purpose. These will greatly reduce losses.
The keyword you're looking for is "terahertz gap". Humanity has yet to figure out what we can do between 200GHz (radio) and 40THz (long wave infrared). The terahertz gap is slowly closing (eg 80ghz radar is fairly mainstream today) and it might disappear in our lifetime.
Now I have to figure how to get a Field Fox! Maybe sell the house?
At 50GHz, every cable is a 3dB pad XD
Question: The old rule of thumb that if you're working with a particular wavelength, where you shouldn't be too concerned about mismatched impedance on scales of 1/16 or lower of that wavelength... is there truth to it? (It's been quite a while, so I'm not sure if I remembered that right, but it's along those lines.) Somebody set me straight on this one, please.
Depends on how bad the mismatch is, therefore: no that rule doesn't work well.
👍👍
Meanwhile in ham radio world: *still uses 1930s "M connector" / "UHF connector" (PL-259/SO-239)*
How about iron core, sleeve to contain the mag field
Iron core to contain/route/restrict the field
Those cables are quite an overkill for the L-band stuff I work with.
I thought only Audiophile guys can hear different in cable , now you say that cable have different character, RFFOOLE 🙂
Actually high end cables
The final transmission values of a coax connection can only be measured in the installation position of the cable. These values not only depend on the laying geometry, but also on any neighboring cables.
The installation of the cable has no effect on the maximum frequency. The cables are fully shielded. Furthermore, a cable whose maximum frequency of operation is a strong function of mechanical movement would be an pretty bad cable.
@@Thesignalpath If the load connected to the cable is matched and balanced to earth, the currents only flow inside the cable. Strictly speaking, this only applies to one frequency. However, the general case is that the TEM wave travels to the end of the cable at maximum frequencies and then partially reflected propagates between the outer conductor and other cables as a waveguide wave in an indeterminate mode. Many greetings.
Black magic wizardry, leave me at dc
Now I know high frequency measurements are impossible because they don’t like bending, moving and heating. While the technology evolves over time, we are still encountering the same problem
But what is the price of their various cable series? I just could not find anything on their resellers websites. It's so fucking annoying.
Dear viewer, If you need to ask the price, this Frequency is greater than your bandwidth.
nah, when someone asks what cables you use, you troll them and say gold plated phono cables.
Yuck. Numbers. Hate them. There are three things physicists are bad at: Calculating with numbers and counting.
What is *THE BEST* RF cable? Also, I don't understand any technical jargon, I just want a Yes / No answer.
Best is one you have right now :) There is no universal best, everyone has own.
First! Because... it matters I guess...
This video hurt to watch.
Our lab engineer would throw anyone who gets caught treating a cable like this out of the window.
The lab is on the 6th floor.
Longer cable would test better not worse,just saying
Longer cables (which would be more lossy) would show a better _absolute_ S11 measurement. However, this is not what we are talking about here. We are talking about the _shift_ in S11 due to movement. A longer cable can have more bends and twists along its length and therefore experience more _variation_ as it moves. This is the metric of interest here.
Interesting theoretical presentation in the first section that is a very use-full and quick explanation. Unfortunately you call out the sponsor of this video about 1000x . Therefore it is hard to ignore the biased view of your presentation. What is true and what appeals to the specific characterises of your sponsors product?
What do you mean? This video isn't sponsored, he said that he bought all of these cables.
As I clearly stated in the video, I am not sponsored in any way by Junkosha.
Anything good makes you positively biased. Just as I am positively biased toward Jennifer Aniston - without being sponsored from her side :)
This is a wonderful review! Have you done a review on waveguides? If not, would you be so kind? Thank you.