How to determine ESR and ESL in practice

[goldsmith]

Dear All
Hi
Thank you for your time .
Suppose that i have a capacitor ( an electrolytic ) , how can i measure it's ESR and ESL in practice ? ( without datasheet )
Thanks in advance
Goldsmith

 

[Kripton2035]

lots of infos and measuring methods about esr here : **broken link removed**

 

[BradtheRad]

This is my collection of links about ESR and ways to measure.

There are homebrew methods and there are meters available for money.

http://ludens.cl/Electron/esr/esr.html



http://members.shaw.ca/swstuff/esrmeter.html

http://www.anatekcorp.com/ttg/tiptrick.htm#Scope ESR

http://www.siliconchip.com.au/cms/A_103243/article.html

http://web.archive.org/web/20101103042900/http://octopus.freeyellow.com/esr.html

http://web.archive.org/web/20101023201524/http://octopus.freeyellow.com/99.html

- - - Updated - - -

I have a Youtube video portraying behavior of a capacitor with various amounts of ESR, with different waveforms applied.

www.youtube.com/watch?v=vm8wf7y94rM

 

[erikl]

If you have a (sinus) signal generator and you can measure ac currents (e.g. with a scope like in the above link), you can separate the R and L measurement: Usually the RC (lower) -3dB cut-off-frequency f_low and the RL (upper) -3dB cut-off-frequency f_upp are sufficiently far from each other, so you can get R_ESR and L_ESL with good approximation via

R_ESR = 1 / (2πf_low×C) -- and
L_ESL = R_ESR / (2πf_upp)

from the -3dB (1/√2 = 0.707) cut-off points, s. the example below:


With f_low ≈ 160kHz and f_upp ≈ 16MHz you get the R_ESR and L_ESL values used in the simulation (R1 , L1).

 

[goldsmith]

Hi Dear All
Thank you for all of your helpful answers . your answers really helped me . :grin:
Best Regards
Goldsmith

 

[D.A.(Tony)Stewart]

ESR and ESL are one of most critical parameters to learn for any "Component Engineer" or designer to understand how to select the optimum component for the budget and space. It is one of my key Quality factors for any part, be it a MOSFET Rs, IGBT switch, 3Watt LED, a connector contact, a buffer drive impedance, or a capacitor.

My rule of thumb is.. "if the spec does not show ESR or equivalent V-I curve or table of values for V-I" do not use the part until you a) get the specs from the OEM or b) choose another part that shows it

One case today, I found a user trying to replace a 22uF 150Vac plastic cap from EPCOS for a SMPS that was buzzing and found it cost $40 from some disti. This part had very low ESR in a Polyester film cap with self-healing properties and could handle large current spikes.. but still failed. I found out by research today, that EPCOS improved that series of old parts and eliminated that big size all together from their catalog (perhaps due to extreme stress levels inside of the film)

But this user was advised to beware of substitutions, but when he asked "like does ESR not matter", he ought to have been advised YES, but was not. ( So beware of expert answers that have high ESR ) He wanted to know if a cheap AC fan or Motor cap with similar material, size and value would work for five bucks. Since the part had no ESR spec, I suspect that was why no expert said YES.

The cheap plastic motor cap. had an ESR of 5Ω which is fine for a shaded pole motor winding, but not for a SMPS demanding a cap with 30mΩ ESR absorbing 5A current pulses.

So the answer is yes ESR is always important in high current pulse applications for any part.

Now back in the 70's high ESR was a design feature for improving the MTBF of your part, because the forces of nature would increase heat dissipation with high surge currents thru the Rs series resistance (ESR). It is actually a reliability scaling factor in MIL-STD-HDBK-883 that drastically drops MTBF when your design ESR is lowered, meaning adding a resistor makes it more reliable than the resistance inside the part. This is still true if the resistance is high and the part cannot handle the load current that a design can deliver. So beware a part does not exceed the Absolute Maximum Rated current and getting close to that will compromise reliability.

STMicroelectronics has a fairly new LDO regulator in micropower that is "designed" for low ESR capacitors. This is owing to the fact that closed loop stability is compromised is the ESR of a cap is too low and that affects the loop gain margin, causing overshoot in other products.

They correctly give designers a range of acceptable Cap. ESR values in a chart for high and low voltage use of the regulator..
Lower means less ripple in the voltage but as more ripple current absorbs the load pulses. As long as the structure and heat is within reason, the part can handle it's rated current. From the graph below for this high impedance low current LDO regulator a capacitor with 10mΩ ESR would be too low. Whereas for a big old 3 terminal regulator with more drive current, lower ESR is even better for SMPS if you want low noise ripple.

For high reliability design, efficiency is traded-off choosing lower stress factors on the part and use other techniques of filtering the noise such as distributed parallel capacitors or a special circuit board substrate between power and ground. Much more expensive but most effective.