Author Topic: How to build a do-it-yourself vector-network analyzer that’s actually useful  (Read 2578 times)

0 Members and 1 Guest are viewing this topic.

Offline ogden

  • Super Contributor
  • ***
  • Posts: 1128
  • Country: lv
Did you guys have like billion dollar university laboratories or something that you are complaining about layers of abstraction with this beast?

Why it is so hard to believe that many tech university (research) labs are packed with expensive equipment? Just example: http://www.myfab.se/Chalmers/Resources/Alltools.aspx

If you want it cheap and 6GHz then go go build this http://hforsten.com/cheap-homemade-30-mhz-6-ghz-vector-network-analyzer.html If you want it really easy to build then get the minicircuits design.

It barely works in 2..3 GHz range, that's it. You are advised to read conclusions chapter at the end.
 

Offline hendorog

  • Super Contributor
  • ***
  • Posts: 1234
  • Country: nz

If you want it cheap and 6GHz then go go build this http://hforsten.com/cheap-homemade-30-mhz-6-ghz-vector-network-analyzer.html If you want it really easy to build then get the minicircuits design.

It barely works in 2..3 GHz range, that's it. You are advised to read conclusions chapter at the end.

Perhaps some better advice would be to read about the latest version:
http://hforsten.com/improved-homemade-vna.html
 
The following users thanked this post: ogden

Offline ogden

  • Super Contributor
  • ***
  • Posts: 1128
  • Country: lv
Perhaps some better advice would be to read about the latest version:
http://hforsten.com/improved-homemade-vna.html

Very good advice indeed. Improved version looks much better. Thanx
 

Offline LaserSteve

  • Frequent Contributor
  • **
  • Posts: 722
  • Country: us
I work at a university. Mini-Circuits  can be quite generous to schools who can make a justification for a discount. Since they get bombarded with requests, they screen the requests very carefully.  Their founder/president/management want to ensure their program makes an impact.


I  have actual experience with Mini Circuits academic discount program, in my  case Doppler tracking using a low phase noise microwave  oscillator. Three words: "Thanks Mister Kaylie!" Your looking at the public, list, price.  The academic price will be less, but it is negotiable based on student impact.

 Free sample parts /massive hardware donations  for academics pretty much disappeared after tax law changes in the late 90s, early 2000s.  At best we  get 15,20, or 50% discounts on small quantities of items. Often  we actually pay higher then consumer prices for a variety of reasons related to liability, and the extra paperwork a vendor has to do to sell to the State. The exception is software, we get seat licenses for some really nice code for next to nothing.

I start doing massive  extra paperwork at the 1500$, 5000$, and over 20,000$ levels.  Instruments over 20,000$ may need a year and a half of approvals from outside agencies, and require bids.  So you want a very specific thermal camera that only is made by one vendor? Well a sole bidder is NOT allowed at that 20K$ level..  Leads to very long meetings, especially if there is no other bidder. My corporate friends just whip out a payment card for stuff like that, and place the order.


In the US, academic hardware funding is way down the past four years.  I'm doing more of "Make due with less" and repairing ancient stuff.
Nothing dead goes off the loading dock unless I,  or the machinist strip it for useable parts.

What a grad student needs and what goes on a hobbyist bench are usually  two DIFFERENT things. I have a habit of trying to go "cheap" as often as possible as funding is scarce and the competition for funding is massive. But often times we need peak performance and high resolution. Graduate students are very costly and expensive employees who work 60 hour weeks.   


Undergraduate instruments have to be built better then mil-spec and bad user interfaces or glitches are intolerable.  Often I have 60 students in a lab with just two TAs, so the gear has to be good, as a TA can usually only help one small group at a time.  Some of our undergrad labs run for 7  TWELVE HOUR days, and every minute of lab time is used. If they do not have good results in the first lab, they will have problems in the next lab in the series, and it will ripple through the whole semester.  (Chemical Engineering in my case)


Undergrad EE students, yeah. you can buy them cheap stuff off Amazon as it's a learning experience for them.  For undergrad  students in other majors, the instruments are  often thought of as  little more then a needed annoyance, (by the students) as much as I'd and the Professors would prefer they do deep learning about instruments and their limitations. We do have limited time in a semester, and nearly every waking moment is used for education if you have good instructors. Thus we need to make some tradeoffs in what is taught and what is experienced in the lab.


That board is designed for education, and guess what, some times we DELIBERATELY make compromises / faults/quirks  in the design so that learning occurs.  Some of our labs are impossible to complete by design, just to see how the final report reads, and to get students to ask questions/think.   >:D



MY day job is "Senior Technician, Instrumentation" for 137 grad students and about 300 undergrads...  We're not allowed to have financial roll-overs, and if we don't spend it, we may not get the same funding the next year.  So we really work to see that tax payer money is carefully spent.  My work lab is very sparse compared to my home shop.  We only buy parts we need, I have almost no stock parts or spares at work.
I measure heat shrink tubing in Dollars per Inch, for example.

I've more then once heard, "Steve, I need a little more Nyquist in your design, cut the sampling rate please". Only to turn around and hear "I need to measure fementofarads to track a droplet passing down a filter medium" in the next meeting.

BTW, on average fifty percent of the grant is taken for overhead by the institution, world wide. Get a  250,000$ grant, only to have just  125,000 to spend.

Steve
« Last Edit: July 20, 2018, 11:28:12 pm by LaserSteve »
"I've Never Heard of a Nuclear Meltdown Caused by a Buffer Overflow"  filssavi
 
The following users thanked this post: Sparky49, Kean, vk4ffab

Online bd139

  • Super Contributor
  • ***
  • Posts: 6625
  • Country: gb

If you want it cheap and 6GHz then go go build this http://hforsten.com/cheap-homemade-30-mhz-6-ghz-vector-network-analyzer.html If you want it really easy to build then get the minicircuits design.

It barely works in 2..3 GHz range, that's it. You are advised to read conclusions chapter at the end.

Perhaps some better advice would be to read about the latest version:
http://hforsten.com/improved-homemade-vna.html

Didn't even notice that. Thanks for the link!
 

Online Sparky49

  • Contributor
  • Posts: 42
Some good replies from people who understand the target market this is aimed at. Most people in a university lab won't have spaces crammed with billions of dollars worth of equipment, but things quickly add up. My research group is tiny in terms of the number of people, but over the years has built up an impressive set of test equipment including GTEMs, reverb and anechoic chambers, VNAs etc. What often happens is you are giving a fixed-size grant from some source to perform some work. This can range from 10's thousands to a couple of million. However, all that money _has_ to be spent, and unfortunately not on whiskey and beer. ;) So what happens is you end up with a thousand or two left over in the pot at the end of the project, which gets absorbed into uni accounts if you don't spend it. More incentive to buy _something_ which might be useful later in the line.

It might seem wasteful (I get frustrated with it myself), but if you don't spend it all, the uni will often reduce your funding in the next round because you were so 'resourceful' in the last. :P

One guy I know in another group bought a 4 port Keysight VNA to measure the impedance of something at 1MHz... I was so jealous...  :-DD
 

Online TheUnnamedNewbie

  • Frequent Contributor
  • **
  • Posts: 631
  • Country: be
  • Sending EM through plastic.
What often happens is you are giving a fixed-size grant from some source to perform some work. This can range from 10's thousands to a couple of million. However, all that money _has_ to be spent, and unfortunately not on whiskey and beer. ;) So what happens is you end up with a thousand or two left over in the pot at the end of the project, which gets absorbed into uni accounts if you don't spend it. More incentive to buy _something_ which might be useful later in the line.


That reminds me of something that happened in my group. Professor storms in: "Guys, we had to spend some money... I hope one of you can use a wafer thinning grinder?"

My university is lucky enough to have a lot of money coming in from industry cooperation projects. As a result, we have less issues with funding, and have a healthy cashflow.
« Last Edit: July 21, 2018, 03:25:00 am by TheUnnamedNewbie »
The best part about magic is when it stops being magic and becomes science instead
 
The following users thanked this post: Sparky49

Offline chris_leyson

  • Super Contributor
  • ***
  • Posts: 1026
  • Country: wales
@ogden I was was too harsh in my criticism and you're right I should have read about the improved version. Actually I managed to have a quick read today and I'm really impressed with the results, the performance is a hell of a lot better primarilly due to the use of resistive bridge couplers.
 

Offline CopperCone

  • Super Contributor
  • ***
  • Posts: 1415
  • Country: us
  • *knock knock*
Did you guys have like billion dollar university laboratories or something that you are complaining about layers of abstraction with this beast?

Why it is so hard to believe that many tech university (research) labs are packed with expensive equipment? Just example: http://www.myfab.se/Chalmers/Resources/Alltools.aspx

If you want it cheap and 6GHz then go go build this http://hforsten.com/cheap-homemade-30-mhz-6-ghz-vector-network-analyzer.html If you want it really easy to build then get the minicircuits design.

It barely works in 2..3 GHz range, that's it. You are advised to read conclusions chapter at the end.

Yea my university had seriously hardcore shit for researchers that don't need to learn how to make a VNA. Students got jack and shit that left the physics department in 1965. I felt like a peon going to visit french royalty in those laboratories. That list you have has a sputter machine. I think anyone that touched a sputter machine where I went to school would just go missing. People setup independent guarding of their expensive equipment with cameras. You needed to be a graduate student assisting the professor to do anything with that stuff. This VNA thing can be taught 2-3rd year as part of a wave theory class and actually be relevant. The one where you need to do calculations on things like stubs, admittance, etc.

My idea would be to make a wave theory class (typically taught) and teach this thing as a lab class that is recommended to be taken in parallel or afterwards. I'm not sure if the frequencies are high enough but having a bit of 'art' in the department to do things like cut out your own filters would a exacto blade would be a breath of fresh air, even if its really simple stuff.. otherwise you just memorize it for a test and forget it afterwards thinking that you will never see it.

I always felt like the physics labs that I took were more fun then the electrical labs, because you had to use IC's or manufactured transistors and I never really felt the same magic as you get playing with lens, vacuum pumps, big open coils, etc.

If you don't know much about component level problems in electrical engineering like resistor drifts, power ratings, parasitics.. it becomes very sterile and almost beurocratic like programming. I think looking at stuff with a VNA from the start could help people get more of a mind set that 'everything has its little quirks'. I think thats important for anyone to know.
« Last Edit: July 21, 2018, 06:11:30 am by CopperCone »
 

Offline LaserSteve

  • Frequent Contributor
  • **
  • Posts: 722
  • Country: us
Coppercone,

While we can't do it for everyone, come to my beloved  College where I work. Show some interest in our Pro's instruments and labs,  and you might be surprised. Such interest often leads to paid summer work.

Just a camera on the entrances and student work area. No where else.

 Plan on working next to a few high school students on summer internships as well. If the research works, expect to see your name on the paper, too. Please expect to use the instrument you drooled over. Expect to have a desk in the student office space for the duration of the stay.

Expect to learn about weekly  group meetings. Then have to defend your work.

In the US, involving ugrad and high school students opens a Prof up for outreach funding, and a better chance of grants. 

When I was a student I drooled over the laser lab thru a four inch window.
Now these days I would be invited in.
In fact I've been in, like 20 years later.

Did I mention the Rocketry, Robotics, Baja Racing, Mini Formula One teams?
All of whom get some decent gear.

Also expect to do one mandatory semester long, paid, corporate internship, your Junior year. Expect to probably travel out of state for the experience.  Then have the option for a second one Senior year.  Things have changed for the better in some respects.

Steve
« Last Edit: July 21, 2018, 01:06:18 pm by LaserSteve »
"I've Never Heard of a Nuclear Meltdown Caused by a Buffer Overflow"  filssavi
 

Offline Beamin

  • Super Contributor
  • ***
  • Posts: 1081
  • Country: us
  • Now with 50% more estrogen! Testosterone Free!
That's not DIY.  :bullshit:

Plus I think my 5'3" ass could beat up both those guys. Who is the target market DIY'er?  :-//
What happened to ODE TO SPOT by Data? Seriously I don't remember changing it.
 

Offline hendorog

  • Super Contributor
  • ***
  • Posts: 1234
  • Country: nz
That's not DIY.  :bullshit:

Plus I think my 5'3" ass could beat up both those guys. Who is the target market DIY'er?  :-//

Already answered.

Quote
DIY in this case refers to DIY implementation of the VNA equations. I.e. work out the code to implement the equations which you have derived in your study.

They mention Matlab (or was it Labview) and Python for doing so.

Here is a link for the simple one port stuff:
http://www.vnahelp.com/tip20.html

And this one goes a bit further:
http://emlab.uiuc.edu/ece451/appnotes/Rytting_NAModels.pdf

The implication of all this is that it isn't _that_ hard to design a decent VNA, because the maths can correct for many of the hardware issues.
And of course there is now open source software available with all of the equations implemented.
 
The following users thanked this post: Mechatrommer

Online TheUnnamedNewbie

  • Frequent Contributor
  • **
  • Posts: 631
  • Country: be
  • Sending EM through plastic.

The implication of all this is that it isn't _that_ hard to design a decent VNA, because the maths can correct for many of the hardware issues.
And of course there is now open source software available with all of the equations implemented.


For some definitions of ''decent'', yes ;) (I recall some people having a lot of issues with drift in one of the cheaper USB VNAs on the site...)


Yea my university had seriously hardcore shit for researchers that don't need to learn how to make a VNA. Students got jack and shit that left the physics department in 1965.

One of the reasons for this is because those tools are still very fragile if you don't know what you are doing. I actually just received an attenuator that was damaged because someone didn't know you should hold on to both the cable and the attenuator when you tighten down the nut on the SMA connector, and as a result one of the center contacts of the female connector on the attenuator got torn off when the connector started spinning. Sure, this attenuator is not quite as expensive as a full VNA, but still costs more than a entry-level Rigol scope, and is now useless.

Something like the minicircuits kit discussed here gives students more hands-on experience without risking damage to some very expensive kit. It being only up to a few GHz also means you don't need to have the very expensive and fragile connectors that most big VNAs come with.

And the real issue is not the cost of the actual damage - it is the lost time. All of our VNAs (and my university EE department has a good number) are pretty much constantly booked for measurements. Losing one for a week or two because it has to go out and be repaired would set back many measurements - likely more expensive than the repair of a connector. Oh and we have an old 50 GHz spectrum analyzer that gets it's first mixer blown out on a yearly basis because someone fails to see the "Max 0 V DC" warning label on the front...
We actually purchased a good bit of entry level kit for the students a while ago, but not yet a VNA... This might be interesting to look into. We give students their own kit because that way we don't need to risk planning issues, it just makes everything easier. And unfortunately VNAs are just a tad expensive for that.



Quote
My idea would be to make a wave theory class (typically taught) and teach this thing as a lab class that is recommended to be taken in parallel or afterwards. I'm not sure if the frequencies are high enough but having a bit of 'art' in the department to do things like cut out your own filters would a exacto blade would be a breath of fresh air, even if its really simple stuff.. otherwise you just memorize it for a test and forget it afterwards thinking that you will never see it.

(...)

If you don't know much about component level problems in electrical engineering like resistor drifts, power ratings, parasitics.. it becomes very sterile and almost beurocratic like programming. I think looking at stuff with a VNA from the start could help people get more of a mind set that 'everything has its little quirks'. I think thats important for anyone to know.

You can definitely do interesting microwave networks at the frequency this can measure. I have myself designed and tested some distributed filters at 2.4 GHz.

I can't comment on your approach of the labs since where I work the classes are different (we don't have many "lab" courses, most of our courses combine the two). But I can see how this could complement one of those courses very well. I would say that it might be more interesting to stick to the VNA part itself, and just building the math involved, as this is a good exercise in thinking in S-parameter blocks. Adding to it a number of actual circuit design classes likely takes too long, but only experience can tell for sure.
The best part about magic is when it stops being magic and becomes science instead
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 1128
  • Country: lv
Oh and we have an old 50 GHz spectrum analyzer that gets it's first mixer blown out on a yearly basis because someone fails to see the "Max 0 V DC" warning label on the front...

You can permanently put DC block on SA that is subject to careless use. Yes, there are some drawbacks but perhaps your application could be fine with them.
 

Online TheUnnamedNewbie

  • Frequent Contributor
  • **
  • Posts: 631
  • Country: be
  • Sending EM through plastic.
Oh and we have an old 50 GHz spectrum analyzer that gets it's first mixer blown out on a yearly basis because someone fails to see the "Max 0 V DC" warning label on the front...

You can permanently put DC block on SA that is subject to careless use. Yes, there are some drawbacks but perhaps your application could be fine with them.

Yeah, I don't know why they don't or what is going on there (it is not one of my group's SA's, so I only hear about it when yet another researcher blew up the front end)
The best part about magic is when it stops being magic and becomes science instead
 

Offline hendorog

  • Super Contributor
  • ***
  • Posts: 1234
  • Country: nz

The implication of all this is that it isn't _that_ hard to design a decent VNA, because the maths can correct for many of the hardware issues.
And of course there is now open source software available with all of the equations implemented.


For some definitions of ''decent'', yes ;) (I recall some people having a lot of issues with drift in one of the cheaper USB VNAs on the site...)


Again, already answered in that thread. The drift is because those VNA's do not implement a reference receiver.

This is quite easy to do, and costs next to nothing - just have a look at the Siglent implementation.

Why did those cheaper USB guys not do it? Beats me, its easy _and_ cheap.  :-//

 

Online Sparky49

  • Contributor
  • Posts: 42
That's not DIY.  :bullshit:

Plus I think my 5'3" ass could beat up both those guys. Who is the target market DIY'er?  :-//

What on earth does your ability to beat up two people have to do with the VNA? Have you ever thrown a hissy-fit because when you buy some paint to decorate your living room (DIY) because the manufacturer didn't release the procedure to make the paint yourself?

The target market is probably someone who (amongst other things) can read through a page of a forum thread.
« Last Edit: July 22, 2018, 12:40:37 am by Sparky49 »
 

Offline wraper

  • Supporter
  • ****
  • Posts: 8256
  • Country: lv
Who is the target market DIY'er?  :-//
Someone did not watch or listen what was said in the video, nor read this thread.
Quote
Plus I think my 5'3" ass could beat up both those guys.
Plus I'm pretty sure police will beat up your ass with taser and put you in jail. One of the stupidest and most irrelevant comments I've read on this forum.
 

Offline joeqsmith

  • Super Contributor
  • ***
  • Posts: 4849
  • Country: us

If you want it cheap and 6GHz then go go build this http://hforsten.com/cheap-homemade-30-mhz-6-ghz-vector-network-analyzer.html If you want it really easy to build then get the minicircuits design.

It barely works in 2..3 GHz range, that's it. You are advised to read conclusions chapter at the end.

Perhaps some better advice would be to read about the latest version:
http://hforsten.com/improved-homemade-vna.html

Looks like some nice work.   

I thought the original video was decent.  Puts the basic blocks together and write the software for it.  I rolled my own software and agree that if you were a student learning RF, this would be very good exercise. 
How electrically robust is your meter?? http://www.youtube.com/channel/UCsK99WXk9VhcghnAauTBsbg
 

Offline Beamin

  • Super Contributor
  • ***
  • Posts: 1081
  • Country: us
  • Now with 50% more estrogen! Testosterone Free!
That's not DIY.  :bullshit:

Plus I think my 5'3" ass could beat up both those guys. Who is the target market DIY'er?  :-//

Already answered.

Quote

Thanks I found that explanation almost as helpful as "google it" but at least when some one says "google it" they are saying something that could be useful.
What happened to ODE TO SPOT by Data? Seriously I don't remember changing it.
 

Offline hendorog

  • Super Contributor
  • ***
  • Posts: 1234
  • Country: nz
That's not DIY.  :bullshit:

Plus I think my 5'3" ass could beat up both those guys. Who is the target market DIY'er?  :-//

Already answered.


Thanks I found that explanation almost as helpful as "google it" but at least when some one says "google it" they are saying something that could be useful.

Did you bother to read the quote immediately after that comment? No, you didn't, because you were too busy looking for something to get angry about.
Mine was the most helpful response you got. Everyone else just told you top pull your head in. So my mistake.

Pull your head in.
« Last Edit: July 23, 2018, 05:46:08 pm by hendorog »
 

Offline mmagin

  • Frequent Contributor
  • **
  • Posts: 612
  • Country: us
This is PURE PILPILISM,you are talking to the wind.....
However....
this project is aimed to teach,so if it's intended for that purpose then why it's not opensource?
why they are trying to pass this bullshit like an opensource project if it's not ....
lol.

What is pilpilism?  My dictionary does not have that word.
 

Online Mechatrommer

  • Super Contributor
  • ***
  • Posts: 8188
  • Country: my
  • reassessing directives...
if something can select, how cant it be intelligent? if something is intelligent, how cant it exist?
 
The following users thanked this post: Sparky49, mmagin

Offline CopperCone

  • Super Contributor
  • ***
  • Posts: 1415
  • Country: us
  • *knock knock*
 

Offline Free_WiFi

  • Regular Contributor
  • *
  • Posts: 105
  • Country: it
There is a guy,which told me something like this : "to make an VNA you need another VNA ......."
At this point if this is true or was,then how the first VNA was born ?
LOL
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf