2
\$\begingroup\$

I have a device with a supply voltage range requirement 12-24AC. The B80S bridge rectifier I am using recommends a 3.5 Ohm series resistor to limit inrush to acceptable levels. Initially I decided to go for 2 x 2.7 Ohm in 0805 (metric) package. While testing the prototype it rarely happened that one of those resistors failed when using 24VAC.

I am somewhat unsure about the single pulse peak power rating that is required for the resistors and would like some feedback.

About 400uF of bulk capacitance (calculated with respect to my load requirements) are used to smooth out the AC before hitting a 5V switching regulator.

A quick estimation, ignoring parasitics like AC transformer resistance, inductance and capacitor ESR:

At 24VAC and an assumed no load factor of 1.3 for the transformer I expect a peak voltage of

U = 24V * SQRT(2) * 1.3 = 44V

With 2 x 2,7 = 5.4 Ohm and completely discharged capacitors this leads to an inrush current of about

I = 44V x 5.4 = 8.15A.

So the dissipated peak power on the resistor would be

P = 44V / 2 * 8.15 = 180W

Having 400uF load capacitance, the pulse duration would be about

t = 5 * R *C = 5 * 5.4 Ohm * 400uF = 10ms

So I would need a 2.7 Ohm resistor with a maximum pulse dissipation of 180W for 10ms.

Maybe this is a somewhat unrealistic estimation since the parasitics will work in my favor, so I performed a simulation:

enter image description here

enter image description here

While this looks "better", this would still lead to a pulse load requirement of about 90W for 3ms. I don't know if 500uH as transformer inductance are realistic though. Since the customers will used totally different supplies this will vary anyway.

Looking at some Pulse Proof Thick Film Chip Resistor data sheets:

enter image description here

I would have to select at least a 2010 resistor type.

Does all this sound/look reasonable or am I way off?

\$\endgroup\$
  • \$\begingroup\$ For extra realism you might want to add the ESR of your caps \$\endgroup\$ – PlasmaHH Feb 14 '17 at 10:39
  • \$\begingroup\$ @PlasmaHH: Cap models have specified 0.34 Ohm which is pretty similar to the caps I am using. \$\endgroup\$ – Rev1.0 Feb 14 '17 at 10:41
  • \$\begingroup\$ excellent. I often model the transformer side of things with an actual ltspice coupled inductor set and values from just some transformer I can find on digikey or somewhere that seems roughly fitting. One could also go full berserk with ltwiki.org/index.php5?title=Transformers but that rarely adds much detail \$\endgroup\$ – PlasmaHH Feb 14 '17 at 10:51
  • \$\begingroup\$ @PlasmaHH: I haven't used LTSpice so I tried to keep it simple. However I would dig into it further if required. But judging from your comment my results seem to be reasonable? \$\endgroup\$ – Rev1.0 Feb 14 '17 at 11:02
  • \$\begingroup\$ But then you bypass those resistors with relay/switch or what? You can also put resistors on primary transformer side, the ratio and inductance would give you a help. \$\endgroup\$ – Marko Buršič Feb 14 '17 at 11:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.