Do I misunderstand ESR measurement? Frequency dependence of ESR

[joeqsmith]

Measured a 1n, 10n & 1uF with BK and cheap VNA using a home made test fixture.  Parts are physically much smaller that what was originally asked about in the other thread. 

1nF   BK@10kHz 1.035nF,  6 ohm ESR, 3000 Q,   VNA 1.08nF, resonance 33MHz,  0.366 ohm ESR @ res
10nF  BK@10kHz 9.858, 2.46, 650 Q,  VNA 9.85nF, resonance 11.4MHz, 0.193 ohm ESR @ res

I did not recal for the 1uF part...
1uF  BK@10kHz 1.024uF, 0.144, 109 Q,  VNA 953nF, resonance 1.4MHz, 0.174 ohm ESR @ res

cap is of course open at DC, but shown impedance zoomed in.

[joeqsmith]

Retesting the 1uF after performing a calibration over a range of 10k to 5MHz,  we can see it resonates at 1.28MHz with an ESR of 24mOhms.

[gamalot]

The 1 µF capacitor you have is most likely a metallized polyester type. I tested a few 1 µF film capacitors I had on hand, and the Kemet MMK shows data very similar to yours, while other metallized polypropylene capacitors perform significantly better.

The following data reflects results at 1 kHz and 10 kHz respectively.

Kemet MMK 1.005/0.996µF Q=247/118

Kemet R75 1.017/1.017µF Q=5147/2655

Panasonic ECWFD 1.002/1.002µF Q=4352/926

Würth FTBP 0.973/0.973µF Q=4976/1226

Unknow Chinese Brand CBB22 1.005/1.004µF Q=2708/1303

[joeqsmith]

I don't have many film capacitors on-hand.  I have a Polycarbonate 1uF that has a Q of 130 and ESR of 0.120 @ 10kHz, when measured with the  BK.   

[Zero999]

What sort of wave form do the capacitor testers use? Is it sine or square? I wonder if it's the latter, in which case it should be easier to measure the ESR, even with low test frequencies, by looking at the voltage or current, when the waveform changes from one voltage to another.

[TimFox]

ESR is a parameter defined for a sinusoidal waveform, and it varies with frequency.
The reactance, obviously, is much different at the third and fifth harmonics.
The result measured with a square wave will differ from the true ESR, but might be useful as a diagnostic for larger capacitance caps.

[Doctorandus_P]

I feel like I deeply misunderstood some basic concept. 😳

...

In my mind it’s a function of frequency.

ESR stands for Equivalent Series Resistance. So it is a resistor (and therefore frequency independent) that is used as a model as a first order approximation of a real physical capacitor. You can make models for a capacitor that more closely follow the frequency dependent imperfections of a real capacitor, but you you'd have to model it for a specific type (and possibly size) of capacitor.

So there is a difference between the simply modeled "ESR", and real capacitor behavior, and when simulating (where you can use "Ideal" passives and then add parasitics to more closely follow real life parts) is always a compromise between accuracy, simulation speed and time spend to build / verify the simulation results.

[TimFox]

In fact, the ESR in the model for a physical capacitor can be measured readily at different frequencies and is a function of frequency.

[joeqsmith]

The 1 µF capacitor you have is most likely a metallized polyester type.

I checked and indeed, it is PE.  Looking around at work, we have some old PS parts.  These should have a low dissipation factor.

[gamalot]

The 1 µF capacitor you have is most likely a metallized polyester type.

I checked and indeed, it is PE.  Looking around at work, we have some old PS parts.  These should have a low dissipation factor.

PE as polyethylene? It's really uncommon.

[TimFox]

The 1 µF capacitor you have is most likely a metallized polyester type.

I checked and indeed, it is PE.  Looking around at work, we have some old PS parts.  These should have a low dissipation factor.

PE as polyethylene? It's really uncommon.

PE is polyester, i.e. polyethylene terephtalate or PET, commonly known as Mylar^TM.  It is really, really common.  As a capacitor dielectric, it has mediocre loss but allows smaller physical size than, for example, polypropylene, which has a lower dielectric constant that requires more area.

[gamalot]

PE is polyester, i.e. polyethylene terephtalate or PET, commonly known as Mylar^TM.  It is really, really common.  As a capacitor dielectric, it has mediocre loss but allows smaller physical size than, for example, polypropylene, which has a lower dielectric constant that requires more area.

These full English names are too hard for me to remember. I often mix them up. I can only remember some of the abbreviations.

For example, I know what a PET capacitor is, but I didn’t realize that PE means the same thing … 

[Bernd_2]

Hello,

As far as I know, PE is the usual abbreviation for polyethylene.

See also Wikipedia https://en.wikipedia.org/wiki/Polyethylene .

Best Regards
Bernd

[Zero999]

Zero999: strictly speaking I made a mistake of assuming it’s rectified 50 Hz. But the same would apply to anybody commenting there.

That wasn't my point.

Take a bridge rectifier, with a capacitor and constant current load. Look at the FFT of the current through the capacitor: there's still plenty of energy in the 1kHz range.

[TimFox]

Hello,

As far as I know, PE is the usual abbreviation for polyethylene.

See also Wikipedia https://en.wikipedia.org/wiki/Polyethylene .

Best Regards
Bernd

“Polyester” is a very generic term for polymers.
No one makes polyethylene capacitors.
PET capacitors are extremely common, and are often (carelessly) referred to as “polyester” by engineers.
“Mylar” is a registered trademark of DuPont, like “Nylon”, or Kleenex for Kimberley-Clark.
When first introduced as a substitute for paper-dielectric capacitors, ca. 1953,  PET capacitors were often called Mylar, but you won’t find that term now on Mouser’s search engine, but other vendors do not recognize the trademark.
For actual polyethylene, the longer-chain polymerization leads to higher density, hence the grades LDPE and HDPE.
If you ask a vendor for “PE” or polyester capacitors, he will certainly sell you PET.

Abbreviation meanings differ from application and profession.  In medicine, “PE” means “pulmonary embolism”.

[joeqsmith]

The 1 µF capacitor you have is most likely a metallized polyester type.

I checked and indeed, it is PE.  Looking around at work, we have some old PS parts.  These should have a low dissipation factor.

PE as polyethylene? It's really uncommon.

PE as in polyester.   Very common.

https://www.mscpl.com/general_technical_information.html


Article on the end of Polystyrene capacitors.   
https://www.electrocube.com/pages/white-paper-what-happened-to-polystyrene-capacitors-data-sheet

[joeqsmith]

Back when polystyrene was a thing

https://www.edn.com/wp-content/uploads/2016/11/1982-10-13_EDN_Pease-capacitor-soakage.pdf

[TimFox]

In the Pease article, he shows a model for a "Mylar" capacitor, with mediocre soakage.
The main problem with polystyrene is its low service temperature, which makes it difficult to use with modern through-hole PCB fabrication.
(I don't believe PS was ever made in surface-mount:  surface-mount film capacitors use plastics with high temperature rating.)

[joeqsmith]

In the Pease article, he shows a model for a "Mylar" capacitor, with mediocre soakage.
The main problem with polystyrene is its low service temperature, which makes it difficult to use with modern through-hole PCB fabrication.
(I don't believe PS was ever made in surface-mount:  surface-mount film capacitors use plastics with high temperature rating.)

He talks about polystyrene's limited temperature in that article.   "... although polystyrene can’t be used at temperatures greater than 80°C"  We used it for sample and hold in automotive (-40 -  85C).  Back then, we were using all TH, wave solder. 

Somewhere around, I have one of the very first surface mount boards we ever did.  Hybrid of TH and SMT.  Hand placed and soldered.

[TimFox]

Similarly, in a multiplexer and ADC board that I designed at the beginning of my career, we used through-hole components and wave soldering, but specified hand soldering for the four polystyrene capacitors used in sample-hold circuits.

[joeqsmith]

I think the main problem with using polystyrene is we have not made those parts in decades. 

100pF caps measured with BK @ 10kHz, CD-ceramic disc  PS-polystyrene   

Type, val, ESR, Q
Mica, 100.5, 200, 1300
CD, 97.5, 1640, 100
PS, 102.8, 77, 2200
Air  98.9,  Q hunts, Dissipation factor 0 

500pF (PS 480pF)  measured with BK @ 10kHz
Mica,  520.1, 23, 1300
CD, 481.2, 540, 64,2
PS, 481.6, 10.5, 3500
Air 486,  Q hunts, Dissipation factor 0 

***
Uncalibrated, looking at S11 impedance of the 100pF PS vs air capacitors

[gamalot]

I'm not a native English speaker, so I really dug deep to figure this out. I did tons of searching, used translation tools, and even asked a few different AIs for help.

In the end I found that PET and PE are totally different materials (just like I used to think), PET is widely used in making film capacitors, while PE is more for things like packaging and insulation.

[Doctorandus_P]

(Attachment Link)

If this model is considered, in my mind ESR is a function of frequency. But this is not the picture I get from the other thread.

I do not understand the argumentation in your mind.
The rectangular thing representing the ESR is a resistor, and frequency independent.
The capacitor on the right represents the capacitor itself.

In the end it also depends on the definition of ESR you use.
If you consider it to be an "idealized parasitic" as a part of modelling a real life capacitor, then it is a simple resistor, and thus frequency independent.

If you use a practical approach and define it as "a bunch of the non-idealities of the capacitor thrown together" then it may very well be a frequency dependent part.

Capacitors have a lot of non-ideal behavior. ESR, leakage, inductance, dielectric absorption, voltage dependent capacity change (especially with ceramics) just to name a few.

Also, if you are measuring ESR on a "real life" capacitor, you are not only measuring the "idealized resistor", but you are unavoidably measuring a mixture of other non-ideal behavior too. And that's one reason the measurement method and frequency are mentioned.

[joeqsmith]

I'm not a native English speaker, so I really dug deep to figure this out. I did tons of searching, used translation tools, and even asked a few different AIs for help.

In the end I found that PET and PE are totally different materials (just like I used to think), PET is widely used in making film capacitors, while PE is more for things like packaging and insulation.

I am not sure there is much to figure out.  I called out the dielectric, and provided you with a link of the nomenclature for a reference:
https://www.mscpl.com/general_technical_information.html

The old Illinois Capacitor books from the 70's and 80's were a few inches thick and loaded with details.   They may provided you with some of the history.  I have attached a paper from 1984.   You will notice they only show the same basic four that were listed in the link I had provided.  Historically speaking, marketing, sales & EEs could only pack so much into two letters but you can see why we used PE,PP,PS & PC to identify them.  Over time, this confusion may be why we now have an IEC standard where dielectrics begin with "K" and one letter for the material.  For example, KP is Polypropylene.    Of course, you have things like Polyphenylene sulfide which is KI rather than KS..... 

[mtwieg]

Overall it's unclear what question this thread is trying to address... seems to have turned into a dumping ground for everyone's personal capacitor anecdotes. Many of these anecdotes are interesting and worth discussion, but probably not in the same thread...

IMO, ESR does not have a "true" definition. If you put a gun to my head and forced me to choose, it would probably "the real component of the measured complex impedance, with some AC measurement frequency/amplitude/bias". Of course ESR can also be measured in the time domain (as is standard for supercapacitors). The proper method of characterization depends on your application.