Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Problem in measuring Capacitor

Status
Not open for further replies.

imijoon

Newbie level 5
Joined
Jul 1, 2013
Messages
9
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
70
Hi everyone,

I have a strange problem in measuring NPO MLCC capacitor. I want to measure a NPO capacitor and according to datasheet I should measure the capacitance and DF in 1Mhz,1Vrms condition. but My lcr-meter is GW-instek LCR-819 and it can measure up to 100khz. when I measure C and D in low frequencies i.e 1khz, D value is far from datasheet maximum limit but it is a positive value. when I increase the test frequency up to 40khz, D decreased and is closed to Zero. With further increase of frequency, strangely D value becomes negative.

What is the reason for this phenomenon? how can I measure the right C and D with this LCR (GW-instek LCR-819 100khz)?

thanks
 

My guess is that long test leads are resonating with the capacitance
Could only happen if the instrument isn't used with suitable test cables and open/short calibration.
 

I use Kelvin clip to measuring. Is it possible that problem is from that? Is it solved if I use 2 wire clip?
 

Reason: the dielectric is lossy down at 1KHz.

MLCC's, especially High-Q versions (typically Barium titanate ceramics) are sintered to have crystal lattices which are optimized for RF frequencies (1MHz ~10GHz, smaller wave lengths). This only leaves the electrode 'skin effect' and material of the electrode to have the biggest contribution to losses. The crystal lattices are large enough so the flux and build up and collapse freely without coupling with the atomic structure of the ceramic (reducing heat and loss). Increasing the wavelength may resonate the atomic structure 'more' which causes more losses. The dielectric losses are magnitudes more than what is seen from the skin effect.

I have attached ATC's appnote which goes into more detail.
 

Attachments

  • ATC_esrlosses_appnote_resonant_line.pdf
    107.4 KB · Views: 35

Reason: the dielectric is lossy down at 1KHz.
No. NP0 is known to have constant permittivity and very low losses over the kHz to MHz frequency range. While capacitors < 1 nF are characterized at a standard frequency of 1 Mhz, the measurement frequency is 1 kHz for capacitors > 1 nF. This (arbitrary) decision already clarifies that the properties don't actually change much over the 1 kHz - 1MHz range, it's apparently a concession to allow operation of LCR meters in an optimal impedance range.

Unfortunately we don't know the actual capacitance of the device under test. If it's too low, the instrument may be simply operated outside it's specified accuracy range, or as already guessed, with incorrect open/short calibration.

A small negative D value within in the specified instrument tolerance would be normal operation, of course.
 

Reason: the dielectric is lossy down at 1KHz.

MLCC's, especially High-Q versions (typically Barium titanate ceramics) are sintered to have crystal lattices which are optimized for RF frequencies (1MHz ~10GHz, smaller wave lengths). This only leaves the electrode 'skin effect' and material of the electrode to have the biggest contribution to losses. The crystal lattices are large enough so the flux and build up and collapse freely without coupling with the atomic structure of the ceramic (reducing heat and loss). Increasing the wavelength may resonate the atomic structure 'more' which causes more losses. The dielectric losses are magnitudes more than what is seen from the skin effect.

The dielectric loss will peak near the resonance and will show typical dispersion behaviour. If the material is Barium Titanate based, they will have absorption close to IR and mm wavelength regions. It is very unlikely that the dielectric is lossy at 1kHz and is fine at 1MHz/1GHz. Compared to 1GHz, 1kHz is almost DC!
 

Thanks all,
My capacitor is Yageo 22pf NPO. So how can I measure it? It's too important for me because I want to validate if it's original or fake.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top