# Electrolytic capacitor ESR at DC?

Discussion in 'General Electronics Discussion' started by eem2am, Oct 14, 2010.

414
0
Aug 3, 2009
2. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,448
2,809
Jan 21, 2010
Technically speaking, the ESR doesn't matter for DC because there will be no current flow.

However I assume you're interested in the charging of the capacitor when DC is applied. I would assume 0.141R since 120Hz is a really low frequency,

3. ### Resqueline

2,848
2
Jul 31, 2009
The capacitance, and ESR for that matter, is both undefinable and uninteresting at DC, since a capacitor doesn't pass DC currents. ESR = delta U / delta I ( = 0 / 0 )
There has to be 1 or more frequency components involved, and anything lower than mains won't be able to involve a power level high enough to bother much about ESR.

4. ### eem2am

414
0
Aug 3, 2009
the reason for asking is that this capacitor will be charged up and then used to actuate a latching solenoid valve by discharging it through the coil.

...so its essentially DC when its discharging.

But i fear that the Zero Hz ESR may be well high and poorly toleranced from part to part (or batch to batch?)

5. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,448
2,809
Jan 21, 2010
ESR tends to decrease with decreasing frequency, however it is undefined for DC.

You are probably interested in quite a high frequency as the rising edge of the waveform as the switch is turned on is NOT DC.

Just assume it's the rated value. It will be close enough.

6. ### Resqueline

2,848
2
Jul 31, 2009
There you are, the frequency when discharging is not zero. Don't confuse its state before discharging with the discharge itself.
A solenoid has an inductance, and as such the current through it will start at zero and rise more or less linearly when a square voltage is applied to it.
So, the current drawn will be triangular, or rather sawtooth, -shaped. A triangle-wave has certain frequency components, definitely not zero.

I think steve mixed up the ESR/frequency relationship, I've always measured (and seen in datasheets) an increase in ESR with a decrease in frequency.
The ESR has a loose tolerance to begin with and is not to be relied upon to remain very stable. It varies with temperature & age as well as several other parameters.
If for some reason you need a well-defined ESR then I suggest you get the lowest-ESR cap possible, and then add a real resistor in series with this.
Organic Semiconductor caps have an ultra-low ESR, and the Nichicon HZ series have the lowest ESR of the standard electrolytics.

7. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,448
2,809
Jan 21, 2010
Quite possibly...

I was relying on my understanding that inductive effects caused an increase in the apparent ESR as frequency increased.

I might have to set something up and measure it!

Here are 2 useful references: here and here.

Interesting to note that the minimum ESR occurs around 100kHz.

You learn something every day.

p.s. I also note that tactful way that Resqueline told me I was completely wrong

Last edited: Oct 15, 2010