Hello,
I am evaluating which resistance values can be right to balance two pairs of capacitors placed in series. An empirical formula hovered in an old forum thread, but to be more analytical from a capacitor manufacturer's guide they took into account the leakage current of C, that is:
Rba ≤ Rla (Vr)/10
where: Rba= balancing resistor value for the capacitor Ca and Rla= leakage resistance for Ca (at the rated voltage of C)
Now, I took a look varius brands of capacitors and from datasheet I noticed that for h.v. DC bus capacitors (that is, in my specific case 100uF/450V) the value for the leakage current is in the range from 100 to 600uA. And the present value is relative to some minutes from the applied voltage.
In particular, from i.e. Nichicon caps, there is a formula valid for a h.v. capacitors or: I(L) ≤ 3√(C*V) (in uA). With this assumption I found a I(l)=600uA, so Rba= 450/0.0006*10= 75k.
That seems to me a quite low value leading to have some heat dissipation. So I wonder if going to 100k or better 150k does no harm , since I don't know the I(L) for my brand of capacitors...Thanks
I am evaluating which resistance values can be right to balance two pairs of capacitors placed in series. An empirical formula hovered in an old forum thread, but to be more analytical from a capacitor manufacturer's guide they took into account the leakage current of C, that is:
Rba ≤ Rla (Vr)/10
where: Rba= balancing resistor value for the capacitor Ca and Rla= leakage resistance for Ca (at the rated voltage of C)
Now, I took a look varius brands of capacitors and from datasheet I noticed that for h.v. DC bus capacitors (that is, in my specific case 100uF/450V) the value for the leakage current is in the range from 100 to 600uA. And the present value is relative to some minutes from the applied voltage.
In particular, from i.e. Nichicon caps, there is a formula valid for a h.v. capacitors or: I(L) ≤ 3√(C*V) (in uA). With this assumption I found a I(l)=600uA, so Rba= 450/0.0006*10= 75k.
That seems to me a quite low value leading to have some heat dissipation. So I wonder if going to 100k or better 150k does no harm , since I don't know the I(L) for my brand of capacitors...Thanks
When I series connect capacitors, I generally use 300k across each cap. This is for 600+ volts using a pair of 350V or higher caps.
Ok, you use an approximate value that fits for the job. My doubt is with the second series, two 250V rated caps that must bear about 450V together. I'm thinking to be in the limit, so the balancing resistor should be rather low...They are Nichicon 220uF/250V PW(M) tolerance 20%
Wire them up and check them... Most of mine are within 10 volts or so. I did have one situation where one cap had 40V and the other had over 200V on it - that one was a bad cap.