Re: The clothespin component test jig
I had a similar idea. Just take an SMA to alligator clips cable, and replace the alligator clips with a single Kelvin clip.
Regards,
Jason
|
Re: The clothespin component test jig
Roger,
I like that test system with little standard-sized pins a lot! For situations where accurate testing on VHF and UHF is needed, and SMDs are available to sacrifice. Remember that SMDs shouldn't be de-soldered and re-used... At least not in circuits requiring reliability.
But if the the rule of the day is perfectionism, then I think that the 50? test load and the test short could be improved! Because as they are in that photo, both of them put the inductances of the 50? resistor and the short into the calibration. So these small inductance values, maybe 1 or 2nH, will be subtracted from actual measurements of components, causing a slight error.
This could be avoided by shortening the loop lengths of the 50? load and the short circuit device a little. The short could be made by bending the tips of the standard-length pins together and soldering them at their tips, while the 50? resistor could be soldered to slightly shortened pins.
I don't know if the improvement would be detectable, but let's be perfectionists today! :-)
|
Re: sma rf or rp: difference and exchangeble?
Hi Robert,
Voor de NanoVNA gebruik je de standaard sma connectors , geen rp of rf.
Succes
Jos
Op 7-2-2020 om 15:08 schreef PA7RG:
toggle quoted message
Show quoted text
Hi,
Would like to buy port savers and cables for a nanovna. Found out that there are sma rp and rf connectors. The first seems to be for wifi (?) and the latter for radio use (vna use)? Are the mechanical the same?
Which one should I buy for my vna (which is still in a far country waiting for shipping)?
Tnx,
Robert
|
Re: The clothespin component test jig
Manfred & Roger -
Those jigs are great! I don't expect to do anything above HF, certainly not above the 2 meter band, so they're just right - for me, anyway.
And I use my -H and -F for hobby work - no need for professional accuracy, either.
I had done something similar to Roger's BNC/binding post idea, but much cruder.
Thanks so much for the contributions. This kind of stuff makes this group such a great place.
--
Doug, K8RFT
|
Personally, I always use NanoVNA-App for firmware updates, rather than DfuSE... much friendlier.
Mike
------------------------------------------------------------------------
*From:* DougVL [mailto:K8RFTradio@...]
*Sent:* Wednesday, 20 January 2021, 12:43 pm
*Subject:* [nanovna-users] NanoVNA-H software update #nanovna-h
toggle quoted message
Show quoted text
Some say that we learn best by making mistakes - and then correcting them.
Don't worry too much - as Mike said, you can't hurt it!
Just accept the fact that you might have to try more than once.
I found that for me, the hardest part was clicking the right 'buttons' in the DfuseDemo upload program. (The worst that happens is that the file doesn't upload!)
|
Some say that we learn best by making mistakes - and then correcting them.
Don't worry too much - as Mike said, you can't hurt it!
Just accept the fact that you might have to try more than once.
I found that for me, the hardest part was clicking the right 'buttons' in the DfuseDemo upload program. (The worst that happens is that the file doesn't upload!)
--
Doug, K8RFT
|
Liam, I got the gist of what you were saying. The government comment was in jest. :)
I wonder if #hugen could make a modification to his firmware to include power levels in dBm?
|
Re: The clothespin component test jig
For more precision work and to go up to 250 MHz I use a jig based on an SMA connector and sockets with SMD loads. I have seen this in several places including Owen Duffy's blog. You can get some very accurate results as shown in the plots below. Impedances as low as 1 ohm and as high as 1K ohm are measured. The one that tests the effect of lead length on the measurement using a small 10 pF ceramic capacitor shows what the jig and NanoVNA can do for the experimenter that needs to measure components.
Roger
|
Re: The clothespin component test jig
On Tue, Jan 19, 2021 at 12:07 PM, Jim Lux wrote: For HF work, banana plugs and jacks might be a "fast to change"
fixturing thing. the 0.75" spacing and similar length is a tiny fraction
of a wavelength at 30MHz. My reference plane is not at the 0.75" jack terminals. I use double sided PCB material with both sides connected together and a small slit in the middle to complete the jig. The reference plane is at the PCB slit. It works OK up to about 150 MHz where the "lumpy" impedance is not calibrated out very well. As the load impedance goes up the effect of the jig and the NanoVNA error becomes evident. The plots below show what kind of results to expect.
|
NOTE: The exceptionally high values of CM resistance are likely due to
resonances below my measurement window which has run from 1 MHz through 50
MHz. In de-Qing with a shunt 4.7k carbon resistor, the curves become much
flater and values pretty much between 1k and 2k. I noted the K9YC's data
increased as he increased frequency. Mine did not. I asked why. De-Qing
is likely the reason.
So, in practice, is there enough loss in the system, likely mostly earth
losses, to effectively de-Q the chokes? Neither of my wires are more than
35-feet off soil surface (decaying upper Permian mudstones).
Was the data presented by K9YC bifilar wound chokes or coaxial cable wound
on the ferrite toroids? I suspect they are with bifilar wound chokes. I
also have to wonder if K9YC de-Q'ed his chokes for measurement?
Also, after winding and testing the 31 materisl with solid #12 solid copper
wire, I am beginning to believe the losses in the stranded wire chokes I've
wound and tested are djue to the insulation. I'm going to tear down the
#10 stranded 43 material choke and rewind it with the solid #12 wire with
wider spaced turns, maybe no more than 8 or 10 turns on 2 stacked cores of
3" OD.
I've also ordered a couple of the 4" 31 material cores from KF78P
Metalworks which will be wound with the #12 solid copper wire. More
updates coming.
PROBLEM with HOME BREW and TEST EQUIPMENT to MEASURE what YOU'VE
BUIILT.......... . . . NEVER SATISFIED............ All the chokes do a
good job of reducing the CM noise, but measurement shows the devil in the
details.....................
Dave - W?LEV
|
Re: MORE CMC SINGLE CORE DATA
I forgot:
An excellent source of ferrites for CMCs and other 'ferrite intensive'
applications is KF7P Metalworks in Utah. I have no connection to that
business other than having ordered a number of ferrites from that source.
He has very competitively priced 31 material ferrite (and a lot more) in
any size you might desire.
Dave - W?LEV
toggle quoted message
Show quoted text
On Tue, Jan 19, 2021 at 8:33 PM AG6CX <edwmccann@...> wrote: Sorry Dave:
For clarity can you comment on your 450 foot doublet (225 ft per side) fed
with ??? Ohm Ladder line of length XXX feet, choke at end, into shack,
matching with 1-30 uH variable inductor. Any capacitor?
Out of matching unit to coax to rig? Of is there another CMC ?
Sorry to be so dense; maybe I missed it in the lengthy thread.
Thanks for your contribution to our great hobby.
73,
Ed McCann
AG6CX
--
*Dave - W?LEV*
*Just Let Darwin Work*
|
Re: MORE CMC SINGLE CORE DATA
I'll attach the measurements of my 450-foot long doublet, yes 2 by 225-foot
wires, which has a schematic of the setup at the end. The ladder line is
nominally 400-ohm. There is a section across the back of the garage which
consists of about 45-feet of AWG #6 bare wire strung an inch apart on
electric fence standoffs. Outside, there is roughly 20-feet of 450-ohm
window line to the feed point and inside the radio room (my wife calls it
my 'box) about 5 feet of window line. That connects to the CMC and into
the L-network matching network that consists of a shunt variable 30 ?H
inductor and a series variable capacitor of 80 to 1000 pF @ 4.5 kV whose
output is adjusted to 50 ¡À j0 ohms. The output at 50 ¡À j0 impedance (from
the variable capacitor) then feeds about 5-feet of RG-8 to the output of
the AL-82. The input of the AL-82 then feeds an Alpha Delta 2-way switch
with another coaxial jumper of about 3-feet to go between the Icom 7300 or
the 7610. All coax is RG-8.
All this started when we moved into this home about 6.5 years ago from
losing everything in the forest fire in 2012. The appliances in the old
mountain house had pre-Chinese switchers in them - FCC tested. The new
appliances in the new home all contain multiple Chinese SMPSs with no
attention paid to EMC and associated CM noise coupled onto about everything
that conducts. We recently had to replace a (only!!) 6-year old
refrigerator (Samsung) with a 'made in the USA' unit (Hotpoint). I had to
build and install a 'rather' aggressive CMC on its linecord to alleviate
the noise from its SMPSs (From China, with Love). It's been an uphill
battle to keep all that noise out of the antennas and feedlines. When I
finally (now who is dense) discovered the usefulness of the lowly CMC in
accomplishing that task, I went to town. I believe I am finally winning.
However, all radio astronomy antennas are located a fair distance from the
house out of necessity.
Dave - W?LEV
toggle quoted message
Show quoted text
On Tue, Jan 19, 2021 at 8:33 PM AG6CX <edwmccann@...> wrote: Sorry Dave:
For clarity can you comment on your 450 foot doublet (225 ft per side) fed
with ??? Ohm Ladder line of length XXX feet, choke at end, into shack,
matching with 1-30 uH variable inductor. Any capacitor?
Out of matching unit to coax to rig? Of is there another CMC ?
Sorry to be so dense; maybe I missed it in the lengthy thread.
Thanks for your contribution to our great hobby.
73,
Ed McCann
AG6CX
--
*Dave - W?LEV*
*Just Let Darwin Work*
|
Re: MORE CMC SINGLE CORE DATA
Sorry Dave:
For clarity can you comment on your 450 foot doublet (225 ft per side) fed with ??? Ohm Ladder line of length XXX feet, choke at end, into shack, matching with 1-30 uH variable inductor. Any capacitor?
Out of matching unit to coax to rig? Of is there another CMC ?
Sorry to be so dense; maybe I missed it in the lengthy thread.
Thanks for your contribution to our great hobby.
73,
Ed McCann
AG6CX
|
Re: MORE CMC SINGLE CORE DATA
Are those catalogues, by any chance, available online?
Had no luck finding any info ...
I have attached a s2p file for a yoke from an old TV, I have
one (somewhere) from a CRT monitor. It would be interesting to
see if there are any major differences.
[image: image.png]
toggle quoted message
Show quoted text
On Tue, 19 Jan 2021 at 20:21, Manfred Mornhinweg <manfred@...> wrote: I would love to measure those yoke ferrites, but I have none. About 40
years ago I tried using such ferrites to wind common mode chokes for RFI
suppression, and had little success, but in those years I didn't yet
understand these matters very well. In hindsight I probably had far too
much stray capacitance in my chokes.
I have a lot of stuff in my junk box, but not even one of those yokes! And
nowadays when you go to TV repair shops you will find heaps of discarded
flat-screen TVs waiting for disposal, but hardly any CRT TV. They are
pretty much all gone...
Anyway, the shape of a yoke ferrite is very far from optimal for our
purposes, as it has a rather large difference between minimum and maximum
path length, and a poor ratio between core cross sectional area and turn
length.
My old Thomson Ferrinox catalog lists a range of yoke cores. They were all
made from either material B30 or B31. Both of those materials have ?i=1100
and are very similar in everything except bulk resistivity, which has a
factor of 25 difference between them. These are manganese-zinc ferrites,
and are rated by the manufacturer to 450kHz. TVs operate them with sawtooth
signals at 50 or 60Hz, and around 16kHz, so the strong spectral components
should roll off by roughly 200kHz or so. I would expect dimensional
resonance problems, high loss, and very reduced permeability, on HF.
Yokes from the last generations of CRT computer monitors might have been
made from more RF-suitable materials, since these ran the scan at
frequencies approaching 100kHz, so that the strong spectral components
reach into a range of a few MHz.
|
Re: The clothespin component test jig
On 1/19/21 9:58 AM, Roger Need via groups.io wrote: Very ingenious. A 50 ohm leaded load will work for lower frequency measurements but the lead inductance and spiral cuts in the metal film resistors will affect your results as you increase frequency.
Here is what I use. Works well up to 50 MHz for me. Small sockets can be added for leaded parts instead of soldering.
For HF work, banana plugs and jacks might be a "fast to change" fixturing thing. the 0.75" spacing and similar length is a tiny fraction of a wavelength at 30MHz.
|
Re: MORE CMC SINGLE CORE DATA
I would love to measure those yoke ferrites, but I have none. About 40 years ago I tried using such ferrites to wind common mode chokes for RFI suppression, and had little success, but in those years I didn't yet understand these matters very well. In hindsight I probably had far too much stray capacitance in my chokes.
I have a lot of stuff in my junk box, but not even one of those yokes! And nowadays when you go to TV repair shops you will find heaps of discarded flat-screen TVs waiting for disposal, but hardly any CRT TV. They are pretty much all gone...
Anyway, the shape of a yoke ferrite is very far from optimal for our purposes, as it has a rather large difference between minimum and maximum path length, and a poor ratio between core cross sectional area and turn length.
My old Thomson Ferrinox catalog lists a range of yoke cores. They were all made from either material B30 or B31. Both of those materials have ?i=1100 and are very similar in everything except bulk resistivity, which has a factor of 25 difference between them. These are manganese-zinc ferrites, and are rated by the manufacturer to 450kHz. TVs operate them with sawtooth signals at 50 or 60Hz, and around 16kHz, so the strong spectral components should roll off by roughly 200kHz or so. I would expect dimensional resonance problems, high loss, and very reduced permeability, on HF.
Yokes from the last generations of CRT computer monitors might have been made from more RF-suitable materials, since these ran the scan at frequencies approaching 100kHz, so that the strong spectral components reach into a range of a few MHz.
|
Re: The clothespin component test jig
I wouldn't trust my leaded resistors in the UHF range. But I'm pretty confident that in the HF range they give a reasonable result. How much inductance could they have? 10nH each? Two in parallel would make 5nH. At 100MHz that's a reactance of 3?, causing some error but not a terrible one.
Also it's likely that when I clip that clothespin with the resistors to the board, stray capacitance partially compensates for this inductance.
Eventually I will buy some good 1% SMD resistors and make a better calibrator.
Your test setup looks nice, but I bet mine is quicker to use! :-)
|
Re: The clothespin component test jig
Very ingenious. A 50 ohm leaded load will work for lower frequency measurements but the lead inductance and spiral cuts in the metal film resistors will affect your results as you increase frequency.
Here is what I use. Works well up to 50 MHz for me. Small sockets can be added for leaded parts instead of soldering.
Roger
|
Re: MORE CMC SINGLE CORE DATA
I have forwarded this message to him so maybe he will jump into the
discussion.
Very nice and extensive experiment, I will save your page in my ferrite
notes.
Something a bit unusual for a test if you have some old CRTs around:
Toroids from deflection yokes in CRTs. I have attached an article from
RadCom
about it.
toggle quoted message
Show quoted text
On Tue, 19 Jan 2021 at 17:22, Manfred Mornhinweg <manfred@...> wrote: Dragan,
A practical example that some may find interesting:
Thanks for that! It's interesting to compare his results to mine.
As I understand that video, he measured the impedance of his CMC at small
signal level, and then used software to calculate the losses at a high
applied voltage, assuming that the behavior of the ferrite is completely
linear with power changes. This isn't entirely correct, so there is a
source of some error, but at 5mT flux density that error might still be
small.
A few months ago I did a series of loss measurements at high drive levels
on various ferrite types. I applied RF voltage from a transmitter, in
parallel with a dummy load, just like him, but I thermally insulated the
cores under test, measured the heat rise over a certain time, and
calculated the power loss from this. My method also incurs in some
inaccuracies, for example I only have an estimated value for the heat
capacity of ferrite. That's why it's interesting for me to compare my data
to his.
He applied 276V to 11 turns on an FT-240-31, at 7MHz. This results in
about 5mT flux density. He measures in gauss, calculating 50.7 gauss,
meaning 5.07mT. Good agreement there.
He calculated 18W loss for the core, under those conditions. That's a
volumetric loss of 735mW/cm?.
I don't have an FT-240-31 to test, so I tested two sizes of 31-material
beads. My results differ slightly for the two sizes, which may be due to
measurement error, or to material differences between batches. On my
FB-31-6873 I got a volumetric loss of 500mW/cm? for that flux density and
frequency, while for my FB-31-1020 I got 590mW.
At least my and his values are reasonably close. I wouldn't expect much
better agreement, given that both his and my tests aren't entirely up to
laboratory standards.
My measurements are here:
|
Slightly different to my results, I only calibrated at 2mA and 4mA with this setup:
TinySA:
30dB attenuator on input
¡®External amplifier¡¯ setting -30dB
280MHz centre frequency, 10MHz span
Sweep setting ¡®precise¡¯
Marker set to 280MHz
NanoVNA-H4:
280MHz CW
Power level 2mA got -8.1dB on the TinySA
Power level 4mA got -2.1dB on the TinySA
Accurate enough for government use? I wouldn¡¯t think so but it depends on the requirement. If it¡¯s for quickly checking local authority 2 way radios, then possibly yes. If it¡¯s for NASA, definitely not!
Liam
toggle quoted message
Show quoted text
On 19 Jan 2021, at 16:22, Brian Donaldson <brianb253@...> wrote:
?i loaded my Nano with Dislord's 1.0.45 and was able to adjust the output. I started out by using a 20dB attenuation to ensure I did not violate the maximum input power on the Low Port. So, using the TinySA after removing the attenuation these are the values I got.
Auto - +2 dBm
2mA - -9 dBm
4mA - -3 dBm
6mA - 0 dBm
8mA - +2 dBm
These should be accurate enough for government work, yes?
|
Roger Need
