EMRFD Message Archive 9174

Message Date From Subject
9174 2013-09-27 09:06:36 Ford Peterson Vector Network Analyzer

For years I’ve had interest in working with filters of various description.  My interest is once again focused on filters and this time do a roofing filter.  In all previous endeavors, I have not had access to a Network Analyzer and would find one to be invaluable tool at the bench.  I currently use a zener diode spanked by a 555 timer and mixed with my signal generator to form a pattern of noise I can see on my spectrum analyzer.  The process works great for tuning but I’m left in the dark about things like return loss, impedance matching, etc.  I think the VNA would helpful.

 

VNA’s are HUGE dough.  State of the art can be as high as six figures.  Age worn units continue to approach 5 figures.  Junque status even falls into the multiple 4 figures.  These babies are well outside the scope of my budget for tinkering.

 

N2PK did a series for QEX many years ago.  http://n2pk.com/

 

TAPR did a design that flourished for a while.  http://www.tapr.org/kits_vna.html  no longer available

 

TAPR claims ten tec picked up the device but has since been abandoned as I don’t see it on their website.

 

DG8SAQ has a SDR kit http://sdr-kits.net/VNWA3_Description.html and runs $735

 

VE3IVM has the N2PK version boards available on his website for $36.  http://www.makarov.ca/vna.htm  Software availability is questionable and the project is daunting assembly of trick components, which may prove difficult to procure (unobtainium).  I haven’t evaluated the cost of procuring components in Q=1 volume but I’m guessing a complete build would approach $500 quickly. 

 

Sigh…

 

My interest is basically HF filters.  Measuring the DUT for insertion loss, matching, etc., is my goal.  What is a good approach to this?  Characterizing crystals, comparing responses, evaluating filters is not ‘new’ by any means.  I’m hoping for finding a cognac solution to my brandy ideas within reach of my beer budget.

 

Ford-N0FP

 

 

 

 

 

9175 2013-09-27 09:25:44 William Carver Re: Vector Network Analyzer
Windows software for the N2PK, “MyVNA” by Dave Roberts G8KBB, is great.
That software also will run the DG8SAQ.

The N2PK has somewhat more dynamic range, useful for looking at filter
stopbands below 100 dB, but without optional frequency shifting hardware
it operates only to 60 MHz. I tested mine against NIST standards at a
university laboratory and was within 1%, and at 30 MHz it happened to be
a little closer to the NIST values than a $30K hp network analyzer on
their bench.

The DG8SAQ goes up to UHF frequencies, dynamic range drops as you go
higher and higher, but not dramatically so. I've had software hangups,
my own problem of not concentrating on it long enough to slog through
the software setup, so have not used my DG8SAQ unit.

W7AAZ
9176 2013-09-27 09:46:47 Ivan Makarov Re: Vector Network Analyzer
Hi Ford,

There is no unobtanium parts for the N2PK VNA, the parts list is documented and all parts can be purchased from two sources: Digikey and Minicitcuits. It will probably cost you around $250, not 500.

The software is also can be obtained from two sources: G8KBB (site temporary down) or DG8SAQ

Regards,
Ivan
VE3iVM

9177 2013-09-27 10:16:59 Dan Rae Re: Vector Network Analyzer
9178 2013-09-27 11:13:23 Ashhar Farhan Re: Vector Network Analyzer
ford,

do you really need a VECTOR analyzer? A scalar analyzer is far less complicated and it could work as well. I made one last year. It is just an Si570 (as a signal source, mixed with 120 MHz crystal controlled oscillator to provide signals from DC onwards) and an AD8307 power meter. Both are controlled by an Arduino that is in turn controlled by a PC based program that does the scans. 

I design and measure by filters, Q of the unknown inductors, amplifier gain, etc. I can also measure the VSWR of any port/antenna using a bridge. To measure the reactive part of the an impedance, I use a little trick. I add 100 pf capacitance in series with the DUT and measure the impedance again. The two measurements can be used to extract the reactive part!

I would guess, even a homebrew receiver with fixed gain can be used for your purpose. A receiver meant for such work is called an Instrumentation receiver. It is a slower but more accurate than a spectrum analyzer. 

- farhan


9179 2013-09-27 11:47:07 Dale Hammer Re: Vector Network Analyzer

Ford,

 

You might want to check out the activity on the Yahoo Group, PHSNA, which stands for “Poor Ham's Scalar Network Analyzer”.  The url of the group is:

http://groups.yahoo.com/neo/groups/PHSNA/info

 

The project is being designed by Jim, N5IB, and Jerry, W5JH and built around an Arduino (Funduino) controller, the cheap AD9850 signal generator board found on Ebay for $5-6, and the W7ZOI power meter.  As the text on the first page of the overview found in the files section of the group says:

“The Poor Ham’s Scalar Network Analyzer

What can you do with an Italian microcomputer, a Chinese signal generator, and some ham ingenuity, for about 50 bucks?”

 

A pcb is being put together for around $10 to go on top of the controller to hold the AD9850 signal generator.

 

73,

Dale  K9NN

 

 

 

 

9180 2013-09-27 12:21:12 William Carver Re: Vector Network Analyzer
A complete description of an AC signal is frequency, amplitude and
phase. A SCALAR analyzer does not measure phase.

A VECTOR analyzer has amplitude AND phase. With phase you have a more
complete understanding of impedances. For example a scalar analyzer may
say that an impedance is "71 ohms" and that's all. A vector impedance
analyzer might say "50 + j50" (which also adds up to 71 ohms). But with
the vector knowledge you know that if you insert 50 ohms of capacitive
reactance in series you will cancel the "+j50" and the impedance will
become 50+j0, a perfect match.

W7AAZ
9181 2013-09-27 12:28:41 William Carver Re: Vector Network Analyzer
And....in the case of filters....knowing phase v.s. frequency the
software will show you time delay through the filter at each frequency.
So if you want to make so-called linear phase filters, that don't ring
regardless of width, a VNA is crucial.
W7AAZ
9182 2013-09-27 15:28:33 Dan Mills Re: Vector Network Analyzer
The AD8302 is made for a poor mans vna..

Use a resistive bridge at the input of the DUT with an 8302 connected
across it, this will with trivial maths give you impedance and phase
at the input, and a second 8302 across the DUT will give you the
transfer function and phase shift.

RF source can be any sweep generator that you happen to own, or a dds
of some sort.

There will probably be some small signal relays for scale factors and
calibration test values, and control could be whatever your favorite
micro is.
That log amp has only 60dB range, but if an external phase detector
was acceptable then much better log amps are available (~100dB range).

Would there be any interest in such a board (0603 size parts in all
probability)?


73, M0HCN.

9183 2013-09-27 15:40:16 radioaustralia123 Re: [emrfd] Vector Network Analyzer
Hi Farhan.
 
Can you please describe the method (trick) that you use to measure the reactance.
 
Regards,
 
Roderick, vk3yc.
 
9184 2013-09-27 15:47:24 Graham Re: Vector Network Analyzer
Dan,

Count me in, I would have an interest in such a board.

cheers, Graham ve3gtc



On 13-09-27 10:28 PM, Dan Mills wrote:
> The AD8302 is made for a poor mans vna..
>
> Use a resistive bridge at the input of the DUT with an 8302 connected
> across it, this will with trivial maths give you impedance and phase
> at the input, and a second 8302 across the DUT will give you the
> transfer function and phase shift.
>
> RF source can be any sweep generator that you happen to own, or a dds
> of some sort.
>
> There will probably be some small signal relays for scale factors and
> calibration test values, and control could be whatever your favorite
> micro is.
> That log amp has only 60dB range, but if an external phase detector
> was acceptable then much better log amps are available (~100dB range).
>
> Would there be any interest in such a board (0603 size parts in all
> probability)?
>
>
> 73, M0HCN.
>
>
9185 2013-09-27 15:48:53 Phil Sittner Re: Vector Network Analyzer
Ford-
 
Ditto all of the earlier comments. My first VNA was the original release of DG8SAQ's product and that set the hook. I was fortunate enough to also build the remarkable N2PK unit and found it to have lab quality dynamic range. If you have an interest in only HF then both units are wonderful; for inclusion of VHF/UHF and a myriad of other features you may choose the DG8SAQ product which powers itself straight from your USB port. In either case a whole new world awaits you with the VNA's.
 
Phil
kd6rm
----- Original Message -----
9186 2013-09-27 16:38:28 Dan Mills Re: Vector Network Analyzer
I am tempted to just do a board that amounts to the bridges with log
magnitude and phase outputs and leave the issue of how to source the
rf to whatever dds or sweep gen people want to use.

I could put a dds on there, but that opens a can of worms given the
very good vhf (and in some cases, low microwave) performance of these
log amps, to take full advantage requires an expensive dds that is
both a fiddle to mount (100 pin QFN), has difficult clock requirements
(At least 1GHz ideally) and is overkill for folk wanting to work in
the HF spectrum.

If I cook up the measurement bridge board would somone else be
interested in doing the signal source and processor support (I am
thinking arduino with an ethernet interface or possibly a beaglebone
or something), I am guessing the bridge will talk spi in all
probability for the AD converters and relay control?

I am off to re read the Agilent impedance measurement handbook.

Regards, Dan.

9187 2013-09-27 16:43:41 Dave Miller Re: Vector Network Analyzer
Dan,
Have a look at the STM32-SDR board I am selling. www.stm32-sdr.com
It has a 3.2" TFT and more than enough Processing power. ;-)
I can get the software looked after I am sure.
I am interested in a board even if you find abetter solution.
Dave
VE7PKE


9188 2013-09-27 18:26:21 Dave Re: Vector Network Analyzer
Dan:
  I did sell a AD9912 board.  The 9912 is solderable in an ssop package.  With a 1 GHz clock(you could use an Si530 - available as free samples)) you can go to 400 MHz.  If there is interest, I could have another round of boards made.

Dave - WB6DHW


9189 2013-09-27 20:19:17 William Carver Re: Vector Network Analyzer
I don't have an immediate application but would get a couple of AD9912
boards, Dave.
W7AAZ


On Fri, 2013-09-27 at 18:26 -0700, Dave wrote:
>
> Dan:
> I did sell a AD9912 board. The 9912 is solderable in an ssop
> package. With a 1 GHz clock(you could use an Si530 - available as
> free samples)) you can go to 400 MHz. If there is interest, I could
> have another round of boards made.
>
> Dave - WB6DHW
> <http://wb6dhw.com>
9190 2013-09-28 02:59:37 ha5rxz Re: Vector Network Analyzer

If you're looking for a cheap VNA then please consider the MiniVNA from Mini Radio Solutions. The performance isn't spectacular as it only has 100dB range but it does cover from 1 MHz to 180 MHz and measures SWR, Phase, Return Loss and Q.


http://miniradiosolutions.com/


No connecti

9191 2013-09-28 03:07:51 Dan Mills Re: Vector Network Analyzer
I would note that unlike a conventional VNA there is no selectivity
inherent in this method, so the signal source must be really clean as
any spurs or harmonics will be measured as if they were part of the
fundamental response, further any non linearities in the dut
generating harmonics will confuse the detector as it is not frequency
sensitive.

Still it should be really simple to implement, and will probably be
good enough for many things.

73 M0HCN.


9193 2013-09-28 09:40:40 Ronald RiemVis Re: Vector Network Analyzer
Hi Ford Peterson,

I bought a VNWA3 and a new world went open for me.
It is fast and the price is right.
A nice peace of equipment.

Greetings,

Ronald


On 28 September 2013 00:06, Ford Peterson <ford@highmarks.com> wrote:
 

For years I’ve had interest in working with filters of various description.  My interest is once again focused on filters and this time do a roofing filter.  In all previous endeavors, I have not had access to a Network Analyzer and would find one to be invaluable tool at the bench.  I currently use a zener diode spanked by a 555 timer and mixed with my signal generator to form a pattern of noise I can see on my spectrum analyzer.  The process works great for tuning but I’m left in the dark about things like return loss, impedance matching, etc.  I think the VNA would helpful.

 

VNA’s are HUGE dough.  State of the art can be as high as six figures.  Age worn units continue to approach 5 figures.  Junque status even falls into the multiple 4 figures.  These babies are well outside the scope of my budget for tinkering.

 

N2PK did a series for QEX many years ago.  http://n2pk.com/

 

TAPR did a design that flourished for a while.  http://www.tapr.org/kits_vna.html  no longer available

 

TAPR claims ten tec picked up the device but has since been abandoned as I don’t see it on their website.

 

DG8SAQ has a SDR kit http://sdr-kits.net/VNWA3_Description.html and runs $735

 

VE3IVM has the N2PK version boards available on his website for $36.  http://www.makarov.ca/vna.htm  Software availability is questionable and the project is daunting assembly of trick components, which may prove difficult to procure (unobtainium).  I haven’t evaluated the cost of procuring components in Q=1 volume but I’m guessing a complete build would approach $500 quickly. 

 

Sigh…

 

My interest is basically HF filters.  Measuring the DUT for insertion loss, matching, etc., is my goal.  What is a good approach to this?  Characterizing crystals, comparing responses, evaluating filters is not ‘new’ by any means.  I’m hoping for finding a cognac solution to my brandy ideas within reach of my beer budget.

 

Ford-N0FP

 

 

 

 

 


9194 2013-09-28 10:28:13 Dan Mills Re: Vector Network Analyzer
Well I have been sketching things a little and studying some
datasheets, and there is a bit of a problem with the AD8302 in that
the phase output has an annoying ambiguity, and poor accuracy around
zero degrees, so something better is really needed.

What I have as an alternative is something where each detector has a
AD8309 log amp for magnitude (100dB dynamic range) with its limiter
output feeding the PFD in an ADF4002 to give an unambiguous phase
measurement up to 100MHz.

Thus the detector board effectively has two vector voltmeters and a
reference voltmeter to measure the power on the reference connection.

Such a board can be hooked to a return loss bridge to do s parameters
and can also be used as a wide range dual power meter, and I might see
about putting the return loss bridge on board, leaving one input
usable as a simple vector power meter if desired.

Time to sketch some circuitry I think.

Regards, Dan.


9195 2013-09-28 11:17:19 William Carver Re: Vector Network Analyzer
On Sat, 2013-09-28 at 18:28 +0100, Dan Mills wrote:
>
> Well I have been sketching things a little and studying some
> datasheets, and there is a bit of a problem with the AD8302 in that
> the phase output has an annoying ambiguity, and poor accuracy around
> zero degrees, so something better is really needed.

There's some kind of old saying that seems to apply: "So close, yet so
far away". I got excited as heck about the AD8301-as-VNA once. Still a
great chip but....your next idea may be better for a general-purpose
gadget.

W7AAZ