ESR of film capacitor?.... used as SEPIC capacitor

O

Attachments

• 7.3 KB Views: 3
• 28.2 KB Views: 6

barry

Advanced Member level 5
If the manufacturer doesn’t specify a parameter, you can be sure it’s bad, not controlled or both. if it’s a requirement for you, then find a different device that DOES specify the parameter.

Or, you could contact the manufacturer who would probably have a better idea of their device’s parameters than a bunch of strangers on the internet.

opampsmoker

O
points: 2

d123

Advanced Member level 5
Hi opampsmoker,

This AVX pdf about X5R dielectric types has a load of great capacitor formulas on page 51, one is to calculate ESR . Hope it helps.

Attachments

• 264.8 KB Views: 2
opampsmoker

O
points: 2

Easy peasy

Advanced Member level 5
The calc is easily done data sheet specifies 1% DF, so Vrms on cap x Irms through cap x DF = watts in cap

in this case the volts are ~ 24Vrms say ( actually a bit lower ) the current is 1.7A so the dissipation is 400mW - typical for a polyester cap

you might get away with it - but it will run pretty hot - 2 in parallel will give 200mW per cap - much better ...

opampsmoker

O
points: 2

KlausST

Super Moderator
Staff member
Hi,

Two caps in parallel...
Isn't the current for each capacitor the half than before?
And the dissipated power just 1/4?

Klaus

schmitt trigger

Advanced Member level 5
What Barry said.
Additionally ESR does have some frequency dependence.
If your operating frequency is significantly different from the specification’s test frequency, then you will have to actually validate it in your circuit.

Easy peasy

Advanced Member level 5
short answer = no. the dissipation is affected by total VA not just the current - esp for polyester.

KlausST

Super Moderator
Staff member
short answer = no. the dissipation is affected by total VA not just the current - esp for polyester.
Do you agree that the current is 1/2? I can't think of anything else.
Then the voltage across the capacitor stays the same? .. when two capacitors are in parallel? I don't understand how this can be.

Klaus

Easy peasy

Advanced Member level 5
Oh yes, sorry, the cap voltage is of course Vc = I . Xc, so if the cap is doubled the Vc must halve, the current is of course half, so the VA is 1/4 of prev - VA x DF = watts in cap,

another point is that the cap volts will in fact be a bit smaller than I said above I/C = dv/dt so dv = 1.7 x 3.33uS / 4.7uF = 1.2 volts pk-pk

thus the VA is more correctly calc as 1.2 x 0.707 x 1.7 = 1.45 x DF of 1% = 15mW

So the original cap will be fine ...

Last edited:

FvM

Super Moderator
Staff member
Referring to the full datasheet https://content.kemet.com/datasheets/KEM_F3103_R60.pdf instead of the short "specsheet" gives some insights. Although it doesn't specify ESR, it has V and I versus f ratings. About 8 A at 100 kHz for the chosen capacitor which relates to a guessed ESR in the 15 to 25 mOhm range.

Capacitor ESR is effectively meaningless for post #1 circuit, because losses of the CRC combi are dominated by the 2R7 resistor.

O

opampsmoker

Guest
Though surely the losses in the 2R7 are minimal?....i mean....the 2R7_in_series _with_C4 is effectively bypassed by the C2?

C4,R3 is just there to stop the dreaded instability inherent with the devious SEPIC with all its L's and C's

Easy peasy

Advanced Member level 5
current mode control will take away the need for any RC dampers, as will having an Fsw 10 times higher than the LC res freq of the power bits ( in volt mode ) - a tightly coupled sepic choke is not really that optimum ( larger transient currents - including start-up and shorted load ) ideally at least 10% DM along with the coupled choke ...

opampsmoker

O
points: 2

FvM

Super Moderator
Staff member
The posted simulation circuit has heayy oscillation during startup due to unsuitable power supply sequence, respectively missing soft start feature.

Doesn't look like a finished design.

mtwieg

Advanced Member level 5
I generally don't bother with DF ratings, unless I'm using it under the same conditions as which DF is measured. Looks like Kemet doesn't provide more detailed data on ESR, so I wouldn't bother.