|
I have a NanoVNA-H4
Using it I am trying to measure the capacitance of a large variable Cap.
I have read about as much as I can find, and watched dozens of You-Tube videos.
My Test lead is one of those 6" long pieces of coax that comes with the unit.
It ends in a very short like 1" long leads with micro alligator clips on the ends.
Now i do the whole calibrate thing and all that,
and simply connect the leads on each section of the cap as close together as possible.
I have two caps, Manufacturer specs say
250 to 40 pf
and
170 to 20 pf
Using the VNA when I measure them the max "C" values are dead on, 249 and 169 pf
But the min's are double what they are supposed to be, the 20 measures at 40
and the 40 measures as 80.
Not sure why. Or if I am doing something wrong when measuring these caps.
Anyone?
Joe WB9SBD
|
On 11/22/21 7:12 AM, Joe WB9SBD wrote: ?I have a NanoVNA-H4
Using it I am trying to measure the capacitance of a large variable Cap.
I have read about as much as I can find, and watched dozens of You-Tube videos.
My Test lead is one of those 6" long pieces of coax that comes with the unit.
It ends in a very short like 1" long leads with micro alligator clips on the ends.
Now i do the whole calibrate thing and all that,
and simply connect the leads on each section of the cap as close together as possible. when you calibrate, was it "at the clipleads" or did you calibrate "at the connector on the VNA". If the latter, then you've got the capacitance of the coax and the "fixture" (i.e. the clipleads). That will make a bigger difference when measuring small capacitances than big ones.
I have two caps, Manufacturer specs say
250 to 40 pf
and
170 to 20 pf
Using the VNA when I measure them the max "C" values are dead on, 249 and 169 pf
But the min's are double what they are supposed to be, the 20 measures at 40
and the 40 measures as 80.
Fixturing error? - although being dead on at the high values and too high at the low values is a bit weird.? Where are the test leads and surroundings relative to the plates? When you're at minimum C, the plates are "unmeshed" - maybe they're closer to something in that configuration. Is your capacitor sitting on a table or suspended in air well away from everything?
Not sure why. Or if I am doing something wrong when measuring these caps.
Anyone?
Joe WB9SBD
|
Hi Jim,
Thanks for the answer.
Yes someone else mentioned about the added "C" from that short piece of coax.
Is there a way to Null out cancel out, calibrate out, whatever so it is just measuring what is between the tiny alligator clips?
when measuring I kept everything away from the plates area. and tested on a wooden table.
Joe WB9SBD
toggle quoted message
Show quoted text
On 11/22/2021 9:20 AM, Jim Lux wrote: On 11/22/21 7:12 AM, Joe WB9SBD wrote:
?I have a NanoVNA-H4
Using it I am trying to measure the capacitance of a large variable Cap.
I have read about as much as I can find, and watched dozens of You-Tube videos.
My Test lead is one of those 6" long pieces of coax that comes with the unit.
It ends in a very short like 1" long leads with micro alligator clips on the ends.
Now i do the whole calibrate thing and all that,
and simply connect the leads on each section of the cap as close together as possible. when you calibrate, was it "at the clipleads" or did you calibrate "at the connector on the VNA".
If the latter, then you've got the capacitance of the coax and the "fixture" (i.e. the clipleads).
That will make a bigger difference when measuring small capacitances than big ones.
I have two caps, Manufacturer specs say
250 to 40 pf
and
170 to 20 pf
Using the VNA when I measure them the max "C" values are dead on, 249 and 169 pf
But the min's are double what they are supposed to be, the 20 measures at 40
and the 40 measures as 80.
Fixturing error? - although being dead on at the high values and too high at the low values is a bit weird.? Where are the test leads and surroundings relative to the plates?
When you're at minimum C, the plates are "unmeshed" - maybe they're closer to something in that configuration.
Is your capacitor sitting on a table or suspended in air well away from everything?
Not sure why. Or if I am doing something wrong when measuring these caps.
Anyone?
Joe WB9SBD
|
Are there trimmer caps on the sides of the unit. If soback out the screws.
This will reduce the minimum capacitance.
*Clyde K. Spencer*
toggle quoted message
Show quoted text
On Mon, Nov 22, 2021 at 10:12 AM Joe WB9SBD <nss@...> wrote: I have a NanoVNA-H4
Using it I am trying to measure the capacitance of a large variable Cap.
I have read about as much as I can find, and watched dozens of You-Tube
videos.
My Test lead is one of those 6" long pieces of coax that comes with the
unit.
It ends in a very short like 1" long leads with micro alligator clips on
the ends.
Now i do the whole calibrate thing and all that,
and simply connect the leads on each section of the cap as close
together as possible.
I have two caps, Manufacturer specs say
250 to 40 pf
and
170 to 20 pf
Using the VNA when I measure them the max "C" values are dead on, 249
and 169 pf
But the min's are double what they are supposed to be, the 20 measures at
40
and the 40 measures as 80.
Not sure why. Or if I am doing something wrong when measuring these caps.
Anyone?
Joe WB9SBD
|
On 11/22/21 7:26 AM, Joe WB9SBD wrote: Hi Jim,
Thanks for the answer.
Yes someone else mentioned about the added "C" from that short piece of coax.
Is there a way to Null out cancel out, calibrate out, whatever so it is just measuring what is between the tiny alligator clips? You need a short, open, load at the clipleads.? So that's fairly easy - open is just the clipleads in the configuration and position you'll use them. Short is, well, shorted. Load is the only tricky one, if you have a 50 ohm resistor, it's easy. If not, you can try and cliplead to your sma load. Fixturing on this kind of thing is tough.? You might want to have the capacitor close to, but not connected to, the clip leads (because the C from the lead to the frame of the capacitor).
when measuring I kept everything away from the plates area. and tested on a wooden table.
Joe WB9SBD
On 11/22/2021 9:20 AM, Jim Lux wrote:
On 11/22/21 7:12 AM, Joe WB9SBD wrote:
?I have a NanoVNA-H4
Using it I am trying to measure the capacitance of a large variable Cap.
I have read about as much as I can find, and watched dozens of You-Tube videos.
My Test lead is one of those 6" long pieces of coax that comes with the unit.
It ends in a very short like 1" long leads with micro alligator clips on the ends.
Now i do the whole calibrate thing and all that,
and simply connect the leads on each section of the cap as close together as possible. when you calibrate, was it "at the clipleads" or did you calibrate "at the connector on the VNA".
If the latter, then you've got the capacitance of the coax and the "fixture" (i.e. the clipleads).
That will make a bigger difference when measuring small capacitances than big ones.
I have two caps, Manufacturer specs say
250 to 40 pf
and
170 to 20 pf
Using the VNA when I measure them the max "C" values are dead on, 249 and 169 pf
But the min's are double what they are supposed to be, the 20 measures at 40
and the 40 measures as 80.
Fixturing error? - although being dead on at the high values and too high at the low values is a bit weird.? Where are the test leads and surroundings relative to the plates?
When you're at minimum C, the plates are "unmeshed" - maybe they're closer to something in that configuration.
Is your capacitor sitting on a table or suspended in air well away from everything?
Not sure why. Or if I am doing something wrong when measuring these caps.
Anyone?
Joe WB9SBD
|
I have had good success measuring C of variable capacitors doing what you describe, except I don¡¯t understand ¡°connecting the leads on each section as close together as possible.¡± Just to fill in a few things you didn¡¯t mention:
- I get the C value from the Smith chart trace.
- I use S11 (reflection) involving only CH0.
- I use a range of about 0.5 to 5 MHz, and believe the most accurate C is pretty low in that range, although C is quite stable and almost constant.
- Any range I use for any measurement is also the range I used for calibration.
One unlikely possibility is that you have approached self-resonance at the lower C so you are seeing a parasitic inductance that gets combined with C. For these sorts of measurements I always have one of the other traces tracking reactance X so I can see how close I am to X > 0¡
Andy
toggle quoted message
Show quoted text
On Nov 22, 2021, at 10:12 AM, Joe WB9SBD <nss@...> wrote:
I have a NanoVNA-H4
Using it I am trying to measure the capacitance of a large variable Cap.
I have read about as much as I can find, and watched dozens of You-Tube videos.
My Test lead is one of those 6" long pieces of coax that comes with the unit.
It ends in a very short like 1" long leads with micro alligator clips on the ends.
Now i do the whole calibrate thing and all that,
and simply connect the leads on each section of the cap as close together as possible.
I have two caps, Manufacturer specs say
250 to 40 pf
and
170 to 20 pf
Using the VNA when I measure them the max "C" values are dead on, 249 and 169 pf
But the min's are double what they are supposed to be, the 20 measures at 40
and the 40 measures as 80.
Not sure why. Or if I am doing something wrong when measuring these caps.
Anyone?
Joe WB9SBD
|
Yes, it is important to do the calibration at the alligator clips. (By the way, most people here are aghast at the use of alligator clips, and they describe very nice test fixtures, but they seem to be working with very high frequency and very small devices. For all my crystal and shortwave radio hobbying, I have been EXTREMELY satisfied with the alligator clips.) I did get a 50 ohm ¡°low inductance¡± resistor for calibration.
Andy
toggle quoted message
Show quoted text
On Nov 22, 2021, at 11:10 AM, Jim Lux <jim@...> wrote:
On 11/22/21 7:26 AM, Joe WB9SBD wrote:
Hi Jim,
Thanks for the answer.
Yes someone else mentioned about the added "C" from that short piece of coax.
Is there a way to Null out cancel out, calibrate out, whatever so it is just measuring what is between the tiny alligator clips? You need a short, open, load at the clipleads. So that's fairly easy - open is just the clipleads in the configuration and position you'll use them. Short is, well, shorted. Load is the only tricky one, if you have a 50 ohm resistor, it's easy. If not, you can try and cliplead to your sma load.
Fixturing on this kind of thing is tough. You might want to have the capacitor close to, but not connected to, the clip leads (because the C from the lead to the frame of the capacitor).
when measuring I kept everything away from the plates area. and tested on a wooden table.
Joe WB9SBD
On 11/22/2021 9:20 AM, Jim Lux wrote:
On 11/22/21 7:12 AM, Joe WB9SBD wrote:
I have a NanoVNA-H4
Using it I am trying to measure the capacitance of a large variable Cap.
I have read about as much as I can find, and watched dozens of You-Tube videos.
My Test lead is one of those 6" long pieces of coax that comes with the unit.
It ends in a very short like 1" long leads with micro alligator clips on the ends.
Now i do the whole calibrate thing and all that,
and simply connect the leads on each section of the cap as close together as possible. when you calibrate, was it "at the clipleads" or did you calibrate "at the connector on the VNA".
If the latter, then you've got the capacitance of the coax and the "fixture" (i.e. the clipleads).
That will make a bigger difference when measuring small capacitances than big ones.
I have two caps, Manufacturer specs say
250 to 40 pf
and
170 to 20 pf
Using the VNA when I measure them the max "C" values are dead on, 249 and 169 pf
But the min's are double what they are supposed to be, the 20 measures at 40
and the 40 measures as 80.
Fixturing error? - although being dead on at the high values and too high at the low values is a bit weird. Where are the test leads and surroundings relative to the plates?
When you're at minimum C, the plates are "unmeshed" - maybe they're closer to something in that configuration.
Is your capacitor sitting on a table or suspended in air well away from everything?
Not sure why. Or if I am doing something wrong when measuring these caps.
Anyone?
Joe WB9SBD
|
On 11/22/21 8:36 AM, Andrew Kurtz via groups.io wrote: Yes, it is important to do the calibration at the alligator clips. (By the way, most people here are aghast at the use of alligator clips, and they describe very nice test fixtures, but they seem to be working with very high frequency and very small devices. For all my crystal and shortwave radio hobbying, I have been EXTREMELY satisfied with the alligator clips.) I did get a 50 ohm ¡°low inductance¡± resistor for calibration.
Andy I agree here - for measurements at 5 MHz, alligator clips (or EZ hooks) work fine. What I do is have a SMA to SMA jumper, a SMA to BNC jack adapter, and then I can use all manner of BNC Plug to leads (e.g. alligator, ezhook, banana plugs). I used a couple 100 ohm resistors in parallel (leads twisted) as my 50 ohm load.? A casual check of the resistor shows that the Z doesn't seem to change much with frequency. What *does* make a difference at the "picofarads" level is the lead layout. I've done things like tape them down so they don't move (and get closer or farther).
|
Use this type of clips see pix and calibrate to the end of the clips 50k to 5 mhz.
The lead is 50 ohm coax
If you are using standard alligator clips you probably are measuring the alligator capicatance and the var cap.
De K8HTB Joe Carron
|
Hopefully it is a LOW battery issue.
I started to try some of the suggestions,
First was the smaller Freq range.
I tried 0.5Mhz to 5Mhz
All the calibration seemed to go in just fine till I hit the done button. And the screen looks like the below.
Then i tried 5 to 10 mhz and same thing.
I noticed the battery indicator seemed a bit low, would that make the screen do this/
or what is happening?
Joe
|
I did a test on a 365 pF variable capacitor salvaged from an old broadcast band receiver. The "test jig" was a coaxial cable with some banana plugs on the end. The OSL calibration was done at the alligator clips using a non-inductive carbon resistor for the load, the clips open and then connected together. Calibration range was 1 to 10 MHz. NanoVNA was connected to the PC and NanoVNA app used to make plots of "effective capacitance" vs frequency.
In open position capacitance was 8 to 8.3 pF over the frequency range. Closed was 363 to 371 pF.
Results attached - Roger
|
I'm trying to understand a few things in your pictures, the one where you are at max "C" makes sense to me, the other looks more like a dead short, the curve for the smith is either not there or very very short.
and what is curcles in the red box about?
Joe WB9SBD
|
The close together is to minimize any inductance?
The clips are on the stud for the stators and the frame of the rotors, they are like 1/4" apart from each other. the ceramic in this photo is 3/16" thick.
Joe
WB9SBD
toggle quoted message
Show quoted text
On 11/22/2021 10:25 AM, Andrew Kurtz via groups.io wrote: I have had good success measuring C of variable capacitors doing what you describe, except I don¡¯t understand ¡°connecting the leads on each section as close together as possible.¡± Just to fill in a few things you didn¡¯t mention:
- I get the C value from the Smith chart trace.
- I use S11 (reflection) involving only CH0.
- I use a range of about 0.5 to 5 MHz, and believe the most accurate C is pretty low in that range, although C is quite stable and almost constant.
- Any range I use for any measurement is also the range I used for calibration.
One unlikely possibility is that you have approached self-resonance at the lower C so you are seeing a parasitic inductance that gets combined with C. For these sorts of measurements I always have one of the other traces tracking reactance X so I can see how close I am to X > 0¡
Andy
On Nov 22, 2021, at 10:12 AM, Joe WB9SBD<nss@...> wrote:
I have a NanoVNA-H4
Using it I am trying to measure the capacitance of a large variable Cap.
I have read about as much as I can find, and watched dozens of You-Tube videos.
My Test lead is one of those 6" long pieces of coax that comes with the unit.
It ends in a very short like 1" long leads with micro alligator clips on the ends.
Now i do the whole calibrate thing and all that,
and simply connect the leads on each section of the cap as close together as possible.
I have two caps, Manufacturer specs say
250 to 40 pf
and
170 to 20 pf
Using the VNA when I measure them the max "C" values are dead on, 249 and 169 pf
But the min's are double what they are supposed to be, the 20 measures at 40
and the 40 measures as 80.
Not sure why. Or if I am doing something wrong when measuring these caps.
Anyone?
Joe WB9SBD
|
On Mon, Nov 22, 2021 at 11:53 AM, Joe WB9SBD wrote:
I'm trying to understand a few things in your pictures, the one where you are
at max "C" makes sense to me, the other looks more like a dead short, the
curve for the smith is either not there or very very short.
The NanoVNA measures the complex impedance R + jX. For a air variable R is very small and for a very small capacitance like 8 pF the reactance X is very high under 10 MHz. At 10 MHz. it is 1/(2*pi*8e-12*10e6) = 1,990 ohms and 19,900 ohms at 1 MHz. . So on the Smith chart when you plot from 1 to 10 MHz. it is just a fat dot at the right hand side. and what is circles in the red box about? The Smith Chart marker can be set to show the "effective capacitance". I put a red box around it to show the value of 8.3 pF with the capacitor open at 9.34 MHz. (same as NanoVNA app graph) and with the cap closed at 8.83 MHz. yielding 365 pF. The other red boxes are just the magnitude of the complex impedance |Z| which I was measuring for an experiment I was doing and can be ignored. Roger
|
On 11/22/21 12:55 PM, Joe WB9SBD wrote: The close together is to minimize any inductance?
The clips are on the stud for the stators and the frame of the rotors, they are like 1/4" apart from each other. the ceramic in this photo is 3/16" thick. It's a trade, isn't it?? Close together minimizes inductance (area of the loop). Close together increases the capacitance (spacing of the wires). What I would be careful about is spacing of the wires from the body (and plates) of the capacitor. If I'm interpreting the picture correctly, the green wire goes to the frame and rotating plates the red to the fixed plates. When the plates are meshed, the first fixed plate "sees" the rightmost rotating plate, which partly shields the green wire. When the plaes are not meshed, the fixed plate sees the frame and the clip lead, which in the picture are aligned with the frame. Connecting the green cliplead farther down the frame, or somehow bringing it off at right angles (to the right, in the picture) would reduce that effect. This is one of those things that you just have to fiddle with.
|
I agree with Roger that the reactance is very high making it tough to measure the min C accurately.
|
On Mon, Nov 22, 2021 at 07:06 PM, WB2UAQ wrote: I agree with Roger that the reactance is very high making it tough to measure
the min C accurately. I think you misunderstood me. My comments about high reactance were an answer to the question posed by Joe WB9SB. I said... For a air variable R is very small and for a very small capacitance like 8 pF the reactance X is very high under 10 MHz. At >10 MHz. it is 1/(2*pi*8e-12*10e6) = 1,990 ohms and 19,900 ohms at 1 MHz. . So on the Smith chart when you plot from 1 >to 10 MHz. it is just a fat dot at the right hand side. I did not say that the NanoVNA could not measure the minimum C of the air variable cap. In fact the nanoVNA gives a good estimate as shown in the screenshot/plot below. The 8 pF is close to the 8.4 pF what I measured with my DE-5000 LCR meter. Roger
|
On Tue, Nov 23, 2021 at 10:37 AM, Roger Need wrote:
I did not say that the NanoVNA could not measure the minimum C of the air
variable cap. In fact the nanoVNA gives a good estimate as shown in the
screenshot/plot below. The 8 pF is close to the 8.4 pF what I measured with
my DE-5000 LCR meter.
Screenshot/graph of 8 pF capacitor was missing in last post. Roger
|
Just a general thought: We are not running a metrology lab. Sure, accuracy is worth chasing. But there is a practical limit to what amateurs can and should expect of affordable measurement instruments. Monday, yesterday, I was working with another PhD EE and PhD in Plasma Physics in designing a piece of equipment for a specific purpose. The comment was made that the latest NANOVNA - the full 3-GHz version - comes mighty close to challenging the equivalent HP (Agilent.....Keysight) equipment. We both have an 8753C with S-Parameter Test Sets and HP precision cal standards. And the cost ratio: maybe 1000:1 or greater. The NANOs are the best piece of test equipment to add to any ham shack or professional design effort since the advent of the DMM (Digital MultiMeter). BTW: The "OPEN" cal standard for RF systems and measurements has been and remains the most challenging of any "standard". A short or strictly resistive standard is relatively easy to fashion at RF and microwave frequencies. The "OPEN" will remain a challenge for metrology labs aimed at higher RF and microwave frequencies - even for amateurs. Frankly, the cal standards provided with most of the NANOs compare rather (embarrassingly) well against the HP (Agilent.....Keysight) standards to the frequency limits of the instruments. Dave - W?LEV On Tue, Nov 23, 2021 at 6:37 PM Roger Need via groups.io <sailtamarack= [email protected]> wrote: On Mon, Nov 22, 2021 at 07:06 PM, WB2UAQ wrote:
I agree with Roger that the reactance is very high making it tough to measure
the min C accurately. I think you misunderstood me. My comments about high reactance were an
answer to the question posed by Joe WB9SB. I said...
For a air variable R is very small and for a very small capacitance like 8 pF the reactance X is very high under 10 MHz. At >10 MHz. it is
1/(2*pi*8e-12*10e6) = 1,990 ohms and 19,900 ohms at 1 MHz. . So on the
Smith chart when you plot from 1 >to 10 MHz. it is just a fat dot at the
right hand side.
I did not say that the NanoVNA could not measure the minimum C of the air
variable cap. In fact the nanoVNA gives a good estimate as shown in the
screenshot/plot below. The 8 pF is close to the 8.4 pF what I measured
with my DE-5000 LCR meter.
Roger
-- *Dave - W?LEV* *Just Let Darwin Work*
|
Alternatively, I present my favorite C meter. The Ballantine 520 measures capacitance from 0.01 pF to 12 uF (not a typo) in 12 ranges.
It's an oldie, but a goodie. I can wave the probes near each other on the lowest scale and watch the meter vary accordingly.
The low end capability makes it valuable to measure PCB trace capacitance.
It is basically a 1 KHz sine wave generator with a precision resistive voltage divider followed by a VTVM calibrated in capacitance
Also, the probes don't respond to C to GND; only between the probe tips, so stray C is avoided. They are simply coax with UT-141 semi-rigid with the ends center conductor exposed as tips.
|
Roger Need
