|
Hello,
I have made (several) CMC-filters by wrapping RG58 on a FT240 mix 31 toroid, the result is very good. I have a special instrument to check the current on the coax outside shield and that goes to zero when I connect the choke on the transmitter side.
When I started to read K9YC "A hams guide to RFI, Ferrites, Baluns .. etc" I wanted to measure the impedance, both resistance and reactance and not only the attenuation as "LOGMAG". I have used to measure LOGMAG via CH0 and CH1 and I get very probable results such as -35 dB at 5 MHz and -30 dB at 20 MHz. The connections is coax-shield on one side to CH0 and coax-shield on the other side to CH1 (to center conductors).
I thought that I would see the impedance of the filter by using another trace (CH1) and asign it the format Resistance to start with. The result however is around 50 ohms when I expected some kohms. Measuring a 560 ohm resistor also gives around 50 ohm. SOLT calibration for 2-30 MHz is performed. What is wrong with my thinking?
If I measure the choke (of course always only the shield) only by CH0 I get a credible result of e.g. 1,3 kohm at 5 MHz, but the resistance is getting lower by increasing frequency, so this is probably not the right way to measure the filters resistance.
Why can I not measure the impedance correct when the CMC-filter is connected between CH0 and CH1?
I have tried with both a NanoVNA-H4 and a SAA2-N with similar results.
73/Torbjorn/SM6AYM
|
Obtain the transmission Zt from a s21 measure requires a calculation.
Zt = Zo * 2(1-s21)/(s21)
Zo at 50 ohm.
Then find the real and imaginary parts.
|
What you have constructed is not quite a common mode choke (CMC). It is a current choke or current "balun". Technically, it isn't even a true current "balun". It works by dening current flow on the outside of the coaxial braid. A true balun would accomplish this function in addition to assuring the currents on the two conductors - the inside of the coax braid and the outside of the inner conductor - are of equal amplitude and opposite phase. But the second requirement is pretty much assured using coax cable as what occurs on the inside of the coax is not (greatly - depending on the integrity of the braid) influenced by what goes on outside the cable. I, as well, have constructed and measured several (maybe as many as 15 or so) true CMCs on 31, 43, and 75 material. These consist of, for the most part, 10 to 18 turns of bifilar wound heavy stranded and insulated copper conductor (#14 and the last on 43 material, #10). The last one on 43 material uses two stacked 3" OD cores. The others are wound on two stacked 2.3" OD cores. The single 31 material is wound on 5 stacked cores of 2.3" OD. Since I can run the full legal limit to my parallel conductor feeders (no coax), I don't want to sense any or absolutely minimal heating in either the cores or the conductors. I'll attach the results of my measurements. I cheated and used the HP 8753C for the measurements as it has most of the required conversions built in. Note the 31 material shines on 160 and good on 75. However, the last one, (BRN) which is wound of #10 stranded (11-turns, bifilar) and insulated wire on two stacked 3" OD cores of 43 material is pretty much the winner for general use on 75 through 10-meters. With a house filled with new appliances the chokes not only transform CM to DM for my parallel conductor feeders, but keeps the SMPS noises from the appliances (with love, fromChina) out of the feedline and antenna. Dave -W?LEV On Thu, Dec 31, 2020 at 3:19 PM Torbj?rn Toreson <torbjorn.toreson@...> wrote: Hello,
I have made (several) CMC-filters by wrapping RG58 on a FT240 mix 31
toroid, the result is very good. I have a special instrument to check the
current on the coax outside shield and that goes to zero when I connect the
choke on the transmitter side.
When I started to read K9YC "A hams guide to RFI, Ferrites, Baluns .. etc"
I wanted to measure the impedance, both resistance and reactance and not
only the attenuation as "LOGMAG". I have used to measure LOGMAG via CH0 and
CH1 and I get very probable results such as -35 dB at 5 MHz and -30 dB at
20 MHz. The connections is coax-shield on one side to CH0 and coax-shield
on the other side to CH1 (to center conductors).
I thought that I would see the impedance of the filter by using another
trace (CH1) and asign it the format Resistance to start with. The result
however is around 50 ohms when I expected some kohms. Measuring a 560 ohm
resistor also gives around 50 ohm. SOLT calibration for 2-30 MHz is
performed. What is wrong with my thinking?
If I measure the choke (of course always only the shield) only by CH0 I
get a credible result of e.g. 1,3 kohm at 5 MHz, but the resistance is
getting lower by increasing frequency, so this is probably not the right
way to measure the filters resistance.
Why can I not measure the impedance correct when the CMC-filter is
connected between CH0 and CH1?
I have tried with both a NanoVNA-H4 and a SAA2-N with similar results.
73/Torbjorn/SM6AYM
-- *Dave - W?LEV* *Just Let Darwin Work*
|
Nice work Dave.
Mel, K6KBE
toggle quoted message
Show quoted text
On Thursday, December 31, 2020, 10:55:20 AM PST, David Eckhardt <davearea51a@...> wrote: What you have constructed is not quite a common mode choke (CMC).? It is a current choke or current "balun".? Technically, it isn't even a true current "balun".? It works by dening current flow on the outside of the coaxial braid.? A true balun would accomplish this function in addition to assuring the currents on the two conductors - the inside of the coax braid and the outside of the inner conductor - are of equal amplitude and opposite phase.? But the second requirement is pretty much assured using coax cable as what occurs on the inside of the coax is not (greatly - depending on the integrity of the braid) influenced by what goes on outside the cable. I, as well, have constructed and measured several (maybe as many as 15 or so) true CMCs on 31, 43, and 75 material.? These consist of, for the most part, 10 to 18 turns of bifilar wound heavy stranded and insulated copper conductor (#14 and the last on 43 material, #10).? The last one on 43 material uses two stacked 3" OD cores.? The others are wound on two stacked 2.3" OD cores.? The single 31 material is wound on 5 stacked cores of 2.3" OD.? Since I can run the full legal limit to my parallel conductor feeders (no coax), I don't want to sense any or absolutely minimal heating in either the cores or the conductors.? I'll attach the results of my measurements.? I cheated and used the HP 8753C for the measurements as it has most of the required conversions built in. Note the 31 material shines on 160 and good on 75.? However, the last one, (BRN) which is wound of #10 stranded (11-turns, bifilar) and insulated wire on two stacked 3" OD cores of 43 material is pretty much the winner for general use on 75 through 10-meters. With a house filled with new appliances the chokes not only transform CM to DM for my parallel conductor feeders, but keeps the SMPS noises from the appliances (with love, fromChina) out of the feedline and antenna. Dave -W?LEV On Thu, Dec 31, 2020 at 3:19 PM Torbj?rn Toreson <torbjorn.toreson@...> wrote: Hello,
I have made (several) CMC-filters by wrapping RG58 on a FT240 mix 31
toroid, the result is very good. I have a special instrument to check the
current on the coax outside shield and that goes to zero when I connect the
choke on the transmitter side.
When I started to read K9YC "A hams guide to RFI, Ferrites, Baluns .. etc"
I wanted to measure the impedance, both resistance and reactance and not
only the attenuation as "LOGMAG". I have used to measure LOGMAG via CH0 and
CH1 and I get very probable results such as -35 dB at 5 MHz and -30 dB at
20 MHz. The connections is coax-shield on one side to CH0 and coax-shield
on the other side to CH1 (to center conductors).
I thought that I would see the impedance of the filter by using another
trace (CH1) and asign it the format Resistance to start with. The result
however is around 50 ohms when I expected some kohms. Measuring a 560 ohm
resistor also gives around 50 ohm. SOLT calibration for 2-30 MHz is
performed. What is wrong with my thinking?
If I measure the choke (of course always only the shield) only by CH0 I
get a credible result of e.g. 1,3 kohm at 5 MHz, but the resistance is
getting lower by increasing frequency, so this is probably not the right
way to measure the filters resistance.
Why can I not measure the impedance correct when the CMC-filter is
connected between CH0 and CH1?
I have tried with both a NanoVNA-H4 and a SAA2-N with similar results.
73/Torbjorn/SM6AYM
-- *Dave - W?LEV* *Just Let Darwin Work*
|
I too wanted to measure the resistance of my CMCs with my nano. And came to exactly the same point:
LOGMAG via CH0 and CH1 giving me plots of -db vs freq.
So to get resistance we need some extra math, which is why there are spreadsheet to do the drill. I follow that. I think.
So if the funky math is just a linear transform of log mag into resistance, then can I take the -db readings a a perfectly good enough proxy for resistance? Is it enough to say that (for example) -30db rejection of CM current is really good, and -35db even better. Is there a minimum value of negative db values like -25db? Or -xyz, that we simple types could use as a "good enough" value for CMC suppression on our antenna feed lines?
If the transform is linear, then using logmag as a proxy for resistance seems reasonable? Oh please....
|
Download ZPLOTS from Dan, AC6LA. You run an S21, save the touchstone files. Then open the zplots spreadsheet, import the touchstone file and it will tell you all you want to know.
--Shane
|
As a wide hand wave, yes. Looking at my measured data for those CMC I
measured both attenuation and resistance, there is a correlation between
loss and resistance. However, it is not linear. Bottom line is that
greater attenuation does indicate greater resistance, but not a simple
linear mapping. I believe there was a simple formula given in a previous
post within this thread which relates resistance based on s21.
Dave - W?LEV
toggle quoted message
Show quoted text
On Fri, Jan 1, 2021 at 3:02 PM Don - KM4UDX <dontay155@...> wrote: I too wanted to measure the resistance of my CMCs with my nano. And came
to exactly the same point:
LOGMAG via CH0 and CH1 giving me plots of -db vs freq.
So to get resistance we need some extra math, which is why there are
spreadsheet to do the drill. I follow that. I think.
So if the funky math is just a linear transform of log mag into
resistance, then can I take the -db readings a a perfectly good enough
proxy for resistance? Is it enough to say that (for example) -30db
rejection of CM current is really good, and -35db even better. Is there a
minimum value of negative db values like -25db? Or -xyz, that we simple
types could use as a "good enough" value for CMC suppression on our
antenna feed lines?
If the transform is linear, then using logmag as a proxy for resistance
seems reasonable? Oh please....
--
*Dave - W?LEV*
*Just Let Darwin Work*
|
Happy New Year Dave!
Do you have a link to one your attachments or presentations that can guide those of us (who can¡¯t find the HP 8753C in the shack under all the other stuff!) using our nanovnas and conversion formulae as required?
Thanks!
Ed McCann
AG6CX
|
Well, I've sent this out twice yesterday, but I'll attach the tabulated
data again. I probably need to take pictures and write this up to make it
more useful to those who don't have all the test equipment.
A month or so ago, I ordered 2.3" OD type 31 cores to do this
'experiment'. In the past, 43 material was the preferred core material,
but 31 material came along with considerably greater ?r that 43 material at
the lower HF bands. My measured data indicated that 31 material is,
indeed, better on 160-meters and marginally better on 75-meters than
equivalent bifilar windings on 43 material. However, 43 material offered
considerably better performance on 40-meters on upward. The measured data
on the 43 material cores formed a smooth convex curve from 1 through 50 MHz
with no resonances indicated over that range. Remember, these are
configured as CMCs with simple bifilar (no twists) windings on the cores,
no contorted or crossed windings in some form or another on the cores.
Yea, I need to write this up so it's a bit more useful to others.
Dave - W?LEV
toggle quoted message
Show quoted text
On Fri, Jan 1, 2021 at 4:35 PM AG6CX <edwmccann@...> wrote: Happy New Year Dave!
Do you have a link to one your attachments or presentations that can guide
those of us (who can¡¯t find the HP 8753C in the shack under all the other
stuff!) using our nanovnas and conversion formulae as required?
Thanks!
Ed McCann
AG6CX
--
*Dave - W?LEV*
*Just Let Darwin Work*
|
Sirs, you have thrown down the gauntlet, Sir, and I accept your challenge. In my previous life, I was a more than adequate spreadsheet wiz. So I should be able to handle this job. haha. First, the object. Build and validate a CMC based on this published and evidently respected source. Specifically the 11 turn on 2.4 #61. I think I know how to wind wire around a core and count to 11. So that is one. Picture attached. Now I want to validate that the CMC delivers (about? sort of?) the same Rs as k9yc published in his NM/THHN #12 on single 2.4 #31 Toroid table.
Using my beloved nano, I built a test jig (please don't laugh too hard...), tried to calibrate over 1-30, and saved the required s2p file. Then imported that file into ZPLOTS from AC6LA. The spreadsheet loads fine, and the initial plots appear fine. But. When I click "generate data" to get the resistance terms, well, it all goes to zeros. So I'm screwing something up here. I've attached the populated spreadsheet and the raw nanosaver output fine.
can anyone give me the slap in the face I need and set me straight...?
Thank you all soooo much.
Signed, a pitiful liberal arts major in Virginia.
km4udx
|
I am such an idiot. Where you see type 61 in my text, please replace it with type 31. Sorry!!
|
The way the picture shows you have connected the CMC, you are not measuring
CM rejection but rather DM which should be low. You need to tie both ends
of the choke together and feed CM and sense CM. Jumper the 'common' or
coax shield around the setup. Both ends going to CH 00 and CH 01 are tied
together and then go to the coax center conductor on each end. That way
you are exciting CM and sensing CM.
Dave - W ?LEV
toggle quoted message
Show quoted text
On Sat, Jan 2, 2021 at 1:01 AM Don - KM4UDX <dontay155@...> wrote: Sirs, you have thrown down the gauntlet, Sir, and I accept your challenge.
In my previous life, I was a more than adequate spreadsheet wiz. So I
should be able to handle this job. haha. First, the object. Build and
validate a CMC based on this published and evidently respected source.
Specifically the 11 turn on 2.4 #61. I think I know how to wind wire around
a core and count to 11. So that is one. Picture attached. Now I want to
validate that the CMC delivers (about? sort of?) the same Rs as k9yc
published in his NM/THHN #12 on single 2.4 #31 Toroid table.
Using my beloved nano, I built a test jig (please don't laugh too
hard...), tried to calibrate over 1-30, and saved the required s2p file.
Then imported that file into ZPLOTS from AC6LA. The spreadsheet loads fine,
and the initial plots appear fine. But. When I click "generate data" to get
the resistance terms, well, it all goes to zeros. So I'm screwing something
up here. I've attached the populated spreadsheet and the raw nanosaver
output fine.
can anyone give me the slap in the face I need and set me straight...?
Thank you all soooo much.
Signed, a pitiful liberal arts major in Virginia.
km4udx
--
*Dave - W?LEV*
*Just Let Darwin Work*
|
Yes I have one side of the pair attached to center pin of CH00 and CH01. So center pin to center pin via one side of the pair, and the other side of the pair going to common on each side.
Instead, I short each pair end and connect them to the center pin each side of the nano? And jump the grounds of ch00 and ch01?
And what Format do i pick on the nano? What is the typical expected display plot if I have both the CMC and nano wired correctly?
This will teach me for taking english literature when I should have taken RF101 for Knuckleheads. hahah.
|
Hi all,
the S21 method for measuring very high impedance in general, including CMC
works quite OK.
I hope the feature will be added to nanovna-saver, then it would be really
easy:
For now I'm using and excel to convert the S21 readings into impedance.
An example is here
Even better results could be obtained using an external bridge which can
have a reference much different than 50 OHm:
73 Ady YO2NAA
toggle quoted message
Show quoted text
On Fri, Jan 1, 2021 at 7:04 PM David Eckhardt <davearea51a@...> wrote: Well, I've sent this out twice yesterday, but I'll attach the tabulated
data again. I probably need to take pictures and write this up to make it
more useful to those who don't have all the test equipment.
A month or so ago, I ordered 2.3" OD type 31 cores to do this
'experiment'. In the past, 43 material was the preferred core material,
but 31 material came along with considerably greater ?r that 43 material at
the lower HF bands. My measured data indicated that 31 material is,
indeed, better on 160-meters and marginally better on 75-meters than
equivalent bifilar windings on 43 material. However, 43 material offered
considerably better performance on 40-meters on upward. The measured data
on the 43 material cores formed a smooth convex curve from 1 through 50 MHz
with no resonances indicated over that range. Remember, these are
configured as CMCs with simple bifilar (no twists) windings on the cores,
no contorted or crossed windings in some form or another on the cores.
Yea, I need to write this up so it's a bit more useful to others.
Dave - W?LEV
On Fri, Jan 1, 2021 at 4:35 PM AG6CX <edwmccann@...> wrote:
Happy New Year Dave!
Do you have a link to one your attachments or presentations that can guide
those of us (who can¡¯t find the HP 8753C in the shack under all the other
stuff!) using our nanovnas and conversion formulae as required?
Thanks!
Ed McCann
AG6CX
--
*Dave - W?LEV*
*Just Let Darwin Work*
|
I never got a conclusive response to my initial query regarding why I could not measure impedance, resistance and reactance on my CMC-choke, i.e. why it can not be shown on the Nano itself although there are formats for it (see first post in this thread). However I found a post by Owen Duffy were he is using the QRP's "NanoVNA MOD v3" (), so finally I reached for a connection to the PC. Now I have measured my choke (16 turns of RG58 on a FT240 mix 31), and using the format "Series RLC" in the program I get what I wanted: R, X and Z. I even get the same result/curves if I do an S11 measurement as well as an S21 measurement. The Nano is a H4. The result is that the resonance is below 2 MHz and since my main interest is 2-8 MHz it should be at bit higher to get higher resistance in my span, but the choke works very well to choke currents on the outside of the coax as measured with my coaxiacableoutershieldmeasurement device. I will see if I get an even better recipe in K9YC Cookbook.
73/Torbjorn
|
Don - KM4UDX
