Capacitance Questions: Troubleshooting Parallel Plate Capacitors

[Marquize10]

Hi all,

I have mechanical background but I came across some electrical issues when I am doing my research. I bought (ams) pcap02 to measure capacitance behavior in parallel plate capacitors for certain material using a dielectric with 4.5 relative permittivity. As known, I used the upper plate side for potential and the bottom side plate for ground! When I start measuring no anything happened at all! But when I disconnected the ground it became very sensitive to my hand or any material that touches it. Is this wrong? If so how to correct it?

Thanks in advance

 

[sophiecentaur]

It could be worth while getting hold of a few capacitors of know value and see what your machine makes or them. Also, you could measure the resistance between the two terminals; they could be shorted. Once you have eliminated those two then someone on PF may think of some other thing to look at.

 

[Marquize10]

It could be worth while getting hold of a few capacitors of know value and see what your machine makes or them. Also, you could measure the resistance between the two terminals; they could be shorted. Once you have eliminated those two then someone on PF may think of some other thing to look at.

Before starting, no short at all. After starting no capacitance reading change ( it’s a ratio to a capacitance in the device). After disconnecting just the ground but the potential connected it became very sensitive. How does this come? I am confused, when I measure known capacitors the device read the amount of capacitance designed in it!Any suggestion?

 

[sophiecentaur]

Sounds like it could be an open circuit then (or vey low). What value are you expecting? What is the area and what is the spacing between plates? This link can help you if you haven't got the information.
What is the lowest C that your machine claims to be able to measure?
Sorry if I'm telling you how to suck eggs but these questions are logical.

 

[Marquize10]

Sounds like it could be an open circuit then (or vey low). What value are you expecting? What is the area and what is the spacing between plates? This link can help you if you haven't got the information.
What is the lowest C that your machine claims to be able to measure?
Sorry if I'm telling you how to suck eggs but these questions are logical.

This is my area and spacing and capacitance below, the device claims that is reads up to 100 nano farad

 

[Marquize10]

When I connect both terminals, the device gives me the same number 141 pF but when I apply any force to it doesn’t react or change the capacitance just 141 pF. When I disconnect the ground it gives me the number 47 pF which is the capacitor inside the device but any force applied to my sensor the 47 pF increases by around 10 pF

 

[sophiecentaur]

This is a mighty small capacitance you’re talking about in the presence of ‘normal’ small capacitances. Your machine has a lower(?) C value? You told me it’s upper limit.

 

[Marquize10]

This is a mighty small capacitance you’re talking about in the presence of ‘normal’ small capacitances. Your machine has a lower(?) C value? You told me it’s upper limit.

What do you mean by by lower C value?

The device has a capacitor in it that has 47 pF, and I connect my capacitor to it it gives me the ratio between device capacitance and mine when I connect two terminals.

When I connect one terminal(potential) of my capacitor to the device it gives me just 47 pF ( the device capacitor) plus any reaction comes from my capacitor.

What does this mean?

 

[anorlunda]

When I connect one terminal(potential) of my capacitor to the device it gives me just 4

The device must have two leads. You need to use both for meaningful readings.

 

[Marquize10]

The device must have two leads. You need to use both for meaningful readings.

Yeah that’s true, it gave me my capacitance correct. I mean what does mean when I disconnected one lead and gave me reading for any reaction on capacitor

 

[sophiecentaur]

Yeah that’s true, it gave me my capacitance correct. I mean what does mean when I disconnected one lead and gave me reading for any reaction on capacitor

You are just measuring some parasitic capacitance. As I don't know the actual layout, I can't account for what you are getting.
Your capacitor is a very low value; anything below a pF or two is pretty hard to measure. This is why I wanted to know the lower limit to the pcap02 measurement capability. The details of the way the machine is connected can add many parasitic pF's. What is the connection layout? The capacitance between two lengths of connecting wire can be 20pF or more.

 

[Marquize10]

You are just measuring some parasitic capacitance. As I don't know the actual layout, I can't account for what you are getting.
Your capacitor is a very low value; anything below a pF or two is pretty hard to measure. This is why I wanted to know the lower limit to the pcap02 measurement capability. The details of the way the machine is connected can add many parasitic pF's. What is the connection layout? The capacitance between two lengths of connecting wire can be 20pF or more.

The device can read from 0 fF to 752 pF

 

[sophiecentaur]

The device can read from 0 fF to 752 pF

Any measuring device has a limit of resolution, though. A surveyor's chain will measure from 0 to 22 yards but you wouldn't / couldn't use it to measure the thickness of a sheet of paper. So the "0fF" is a bit hopeful. From what you describe, it looks as though you are trying to work below the resolution limit.
But you need to post a diagram of your arrangement; the details are vital when dealing with less than 1pF. I looked briefly at the pcap documents but I couldn't find an 'Application Note', which could be useful. The connection method will be very critical.

 

[sysprog]

Any measuring device has a limit of resolution, though. A surveyor's chain will measure from 0 to 22 yards but you wouldn't / couldn't use it to measure the thickness of a sheet of paper. So the "0fF" is a bit hopeful. From what you describe, it looks as though you are trying to work below the resolution limit.
But you need to post a diagram of your arrangement; the details are vital when dealing with less than 1pF. I looked briefly at the pcap documents but I couldn't find an 'Application Note', which could be useful. The connection method will be very critical.

Oh gorsh Sophie, it makes me shed a tear, I wish Old Jim were still here to explain, but oh heck, you explained well.

 

[sophiecentaur]

Jim had much more good Will than I can ever sum up. That (plus his endless enthusiasm) was his secret.

 

[sophiecentaur]

I have been looking at more of the documentation about the pcap devices. They seem to be used in large quantities for many applications so they must be fairly low cost(?).
So why don't we see little boxes for sale on eBay and elsewhere which will measure Capacitors in the home electronics lab? Or are they in fact used in upmarket DMMs?
Edit: Apparently, the device has a resolution of around 10fF, which beats most component bridges by a mile - under special conditions, I would assume.

 

[Tom.G]

they must be fairly low cost(?).

Yup. $8 in singles, $4 in thousands. Development/evaluation kit $300. at Mouser.com
Datasheet, User Guide, etc. at AMS.com

Measurement is by charge/discharge time of unknown capacitor compared to time with a 'known' capacitor. 4 floating channels or 8 common (grounded) channels, with one used for the 'known'. The chip has a selection of resistors for ranging/sample-time and a builtin CPU to do the calcs and averaging. Not sure, but I seem to recall that the CPU is user re-programmable, that may help explain its popularity.

The 47pF reading is the pre-installed capacitor on the Evaluation Kit.

If I understood the OPs question correctly, he is asking why there are bad readings with one lead of the 'unknown' capacitor disconnected. That would be stray capacitance to his hand and induced electrical noise.

The probable reason for not getting a reading is because the capacitor he is trying to measure has a small value relative to the default 47pF pre-installed ones. The device compares the charge/discharge time of the unknown capacitor to that of the capacitor on channel one. To get reasonable readings, the pre-installed capacitors must be removed and the capacitor on channel 1 must have a value close to that of the 'unknown' capacitor. It is likely that the appropriate on-chip timing resistor will have to be selected and perhaps other chip options selected in software.

There is much documentation on the AMS.com and the OP will have to understand how to operate the chip.

Cheers,
Tom
edit: a couple typos for clarity

 

[sophiecentaur]

So much documentation and I was looking for a noddy page like the above. Thanks.

 

[essenmein]

Yeah since talking very low capacitance values, as soon as you remove the "ground" and just measure between two plates in space, you also have to consider the capacitance into free space, moving your hand near it, or touching it will affect this significantly.

Classic case of capacitance to free space is the toroid or sphere on top of a tesla coil to complete the HV resonant circuit.

 

[Mark Harder]

Hi all,

I have mechanical background but I came across some electrical issues when I am doing my research. I bought (ams) pcap02 to measure capacitance behavior in parallel plate capacitors for certain material using a dielectric with 4.5 relative permittivity. As known, I used the upper plate side for potential and the bottom side plate for ground! When I start measuring no anything happened at all! But when I disconnected the ground it became very sensitive to my hand or any material that touches it. Is this wrong? If so how to correct it?

Thanks in advance

There are several things missing from your description of the experiment. Am I right in saying that the (ams) thing is a capacitance meter? What is the sensitivity of this device? Your theoretical calculation of the expected capacitance is 0.141 pico farads. That's an extremely small capacitance. I've never seen a practical circuit using a .14 pF capacitor. Are you sure your meter can detect much less measure 0.14 pF? Perhaps "nothing happened" means that there was a 0.00 reading on your meter? In which case, I'd say you are using the wrong instrument for the job. Good luck finding one that can measure capacitances that low.
As for the result you obtain after breaking the charging circuit by removing the ground connection, it's known that dielectrics have a sort of memory effect (I forget what it's called.) The molecules in the dielectric re-align themselves in the electric field of a charged cap but they often don't 'relax' as soon as the charge is drained. In fact, in some cases the voltage on the cap will jump up a bit after removing the charging current. Capacitors with an intense memory effect are useless in some electronic circuits. I'm not sure just what you mean by the ungrounded cap becoming 'sensitive' to your hand after disconnecting the ground. My best guess along the lines above is that your unknown dielectric has a strong memory effect and it's generating its own electric field after the ground circuit is broken and that your hand is displacing that field from the outside of your cap. Without more info, I'm afraid I can't be more helpful.
BTW, I recommend Scherz and Monk, "Practical Electronics for Inventors" as a good intermediate level electronics text. It covers theory and also practical matters like the properties of the various kinds of capacitors, such as the memory effect I mentioned. Unfortunately, it contains a lot of errors, too. If something doesn't make sense, it might be the authors, not you who are to blame. Despite these, I've learned a lot from it.

 

[sophiecentaur]

@Mark Harder I'm of the same opinion. The smallest chip capacitors I ever came across were, perhaps 0.5pF and they were only that value if they were soldered in just right. (I just took their word for it.)

 

[rbelli1]

Here is a selection of https://www.digikey.com/products/en/capacitors/ceramic-capacitors/60?FV=mu0.1pF|2049%2Cffe0003c&quantity=0&page=1&k=capacitor&pageSize=25&pkeyword=capacitor.

Notice that they range from ±50% to ±500% tolerance. They are probably not 0.1pF no matter how you solder them.

So it looks like a typical production measurement system for capacitors pretty much generates random numbers in that range.

BoB

 

[Tom.G]

0.1pF Air capacitor has plate separation 2.24 times the plate area, measured in Inches, resp. Sq. Inches

capacitance in picofarads = 0.224x(n-1)xKxA/d
n: No. of plates
K: relative Dielectric Constant (Air≅1)
A: Area of one side of one plate, Sq. Inches
d: distance between plates, Inches

formula from THE RADIO AMATEUR'S HANDBOOK, forty sixth edition, pg. 24
EDIT: specified dimensions as Inches as noted by @hutchphd and @sophiecentaur

 

[sophiecentaur]

They are probably not 0.1pF no matter how you solder them.

I was thinking that a blob of carelessly hand-applied solder could be of significant height compared with the track spacing.
@Tom.G gives a reference that could tell you just how much it could effect the capacitance between two tracks. Then there would be the problem of detecting the impedance change on a line. I guess Time Domain Reflectometers have come on a bit since I last used one but it would really constitute a tiny blip on the trace. I think you just have to believe that the computer got the values right for machine assembly.

 

[hutchphd]

0.1pF Air capacitor has plate separation 2.24 times the plate area

In cm I assume...?

 

[Tom.G]

In cm I assume...?

Irrelevant, if you measure in Cubits it still works. It's the ratio of the plate area to the separation distance. (A consequence of inverse-square field strength.)

EDIT:
That formula is for dimensions in Inches.
As pointed out by the kind folks here. I sure blew that one!
Original post has been corrected.
www.physicsforums.com/threads/capacitance-question.974834/post-6210177

 

[sysprog]

Any measuring device has a limit of resolution, though. A surveyor's chain will measure from 0 to 22 yards but you wouldn't / couldn't use it to measure the thickness of a sheet of paper. So the "0fF" is a bit hopeful. From what you describe, it looks as though you are trying to work below the resolution limit.
But you need to post a diagram of your arrangement; the details are vital when dealing with less than 1pF. I looked briefly at the pcap documents but I couldn't find an 'Application Note', which could be useful. The connection method will be very critical.

[emphasis added]

In 1991 I had a Minox 3 camera and it had a nicely-made chain thing and it had little beads or imbraidments [$\leftarrow$ made-up word] emplaced at differently-sized intervals (and I wondered why but at first I didn't check) and I didn't know until later that the chain was intended not only for retention of the device but also for measurement for the purpose of accuracy of focus $\dots$

 

[sophiecentaur]

In cm I assume...?

Woop Woop Woop dimension alert! Well spotted young man.

 

[sophiecentaur]

Irrelevant, if you measure in Cubits it still works. It's the ratio of the plate area to the separation distance. (A consequence of inverse-square field strength.)

Area is s² so Area/Distance (s²/s) is unit dependent, surely?

 

[sysprog]

Area is s² so Area/Distance (s²/s) is unit dependent, surely?

Please remember that 2019 is the year in which they/we screwed around with the units by assigning the speed of light/causation to be 1 -- oh and yeah, that speed is still somewhere around 286,000 miles per second -- where does that information go when we say that the speed of light is 1 something?.

 

[sophiecentaur]

we say that the speed of light is 1 something?.

In the 'natural system', c would to be dimensionless. The resulting values of distance and time would have suitably adjusted units so that the ratio would be 1. Once you choose the second (say, the units of distance would end up as 3e8 m so that c would measure as 1. (Other combinations are available)

 

[Averagesupernova]

Please remember that 2019 is the year in which they/we screwed around with the units by assigning the speed of light/causation to be 1 -- oh and yeah, that speed is still somewhere around 286,000 miles per second -- where does that information go when we say that the speed of light is 1 something?.

WOW! Someone really screwed around with something since the last I knew the speed of light in a vacuum was 186,282 miles per second.

 

[sophiecentaur]

that speed is still somewhere around 286,000 miles per second

I guess the mile can be defined in terms of the distance traveled by light in 1s but I think it is actually defined as 1,609.344 metres. The last time I used the 186k miles per second value in anger was in secondary School in the early 60s. The mile is a quaint unit but we still use it (UK) when driving around but the fuel consumption figures for new cars don't mention miles. I let the car information display tell me the mpg but does anyone really believe that?

 

[sysprog]

WOW! Someone really screwed around with something since the last I knew the speed of light in a vacuum was 186,282 miles per second.

You're right -- my mistake.

I guess the mile can be defined in terms of the distance traveled by light in 1s but I think it is actually defined as 1,609.344 metres. The last time I used the 186k miles per second value in anger was in secondary School in the early 60s. The mile is a quaint unit but we still use it (UK) when driving around but the fuel consumption figures for new cars don't mention miles. I let the car information display tell me the mpg but does anyone really believe that?

When I was in school the mile was defined as 5280 feet.

 

[sophiecentaur]

When I was in school the mile was defined as 5280 feet.

Not "defined" in the strict sense - there are ratios between all length measurements. In the end, you need to start with one single length - in this case, afaik, the metre is the prime unit of length and it is defined in terms of the second and c. All the others follow that.