2
\$\begingroup\$

I'm part of a project in which we are developing a PWM controller for a 24V DC motor.

Our current problem is choosing a capacitor to replace C1, which keeps exploding. We've been running tests at 50Amps which is around the current that will be passing through the motor in normal conditions.

How do we choose the right capacitor in terms of voltage, capacitance, ripple current and ESR?

enter image description here

The capacitor we currently use is the SAMXON KM series 47uF 63V, with 146mA of ripple current and it is clearly under dimensioned. Here is the the simulation using a PWM signal of 4kHz, 50% duty cycle.

The current exceeds by a lot the maximum the capacitor can take but if I change it for another capacitor with higher capacitance and higher ripple current the problem keeps happening, but with higher currents, which makes sense I think.

enter image description here

Finally, one last question: Is the voltage change across the terminals of the batteries we see on the graph bad for them? These are lead-acid AGM batteries from Yuasa: https://www.yuasa.co.uk/batteries/industrial/rec-vrla-cyclic-use/rec36-12.html.

Some of the components we use:


Here is the PCB (older version), PCB layout (current version) and PCB render (current version).

PCB

PCB layout

PCB render


Version with multiple caps, same current passing through each one:

enter image description here


Battery and motor setup, during testing. Both motor and batteries positive poles connected to the middle contact in the controller, motor negative connected to the right, battery negative pole connected to the left. The other motor is connected to a couple of high power resistors to generate load.

enter image description here

\$\endgroup\$
17
  • \$\begingroup\$ What is the whole part number for the 47uF cap? Also you have in the sim everything connected to ground, in the real world you will have inductance and resistance for wires so this simulation is not accurate. \$\endgroup\$
    – Voltage Spike
    May 23 at 16:51
  • 1
    \$\begingroup\$ Can I ask what is the purpose of the cap? \$\endgroup\$
    – user253751
    May 23 at 17:12
  • 2
    \$\begingroup\$ Either you remove C1 at all or you use many polymer electrolyte caps in parallel, that can carry the high ripple current as a group \$\endgroup\$
    – Jens
    May 23 at 17:35
  • 1
    \$\begingroup\$ @Jens removing C1 will push the problem to the power supply, which is probably more expensive to replace. \$\endgroup\$
    – Aaron
    May 23 at 18:21
  • 2
    \$\begingroup\$ Adding more caps in parallel is the solution. Also, higher voltage caps. When the motor back EMF gets dumped to the cap, its voltage can go well above the supply rail. Also, these caps should be very close to the driving circuit (FETs?) \$\endgroup\$
    – Aaron
    May 23 at 18:24

1 Answer 1

2
\$\begingroup\$

As per the attached waveforms, peak-to-peak ripple current through the capacitor is in the range of 3-3.2A and the selected capacitor's ripple current rating is 146mA.

As per the datasheet, it is a general-purpose capacitor whose ESR value will be significant for the actual p-p ripple current. The explosion might be the result of I^2R losses on the capacitor.

As per calculation, the capacitor value should be more than 350uF for 4kHz and 50% duty cycle operation. Here we consider 2% voltage ripple.

Try increasing capacitor value and changing the capacitor type from electrolytic to MKP which provides more ruggedness to the system.

The ripple on the battery voltage may occur due to the parasitic inductance of the traces and it will increase if the distance is more. So place the caps close to reduce track inductance.

For further clarification, kindly contact us on Infineon Developer Community. https://community.infineon.com/

\$\endgroup\$
1
  • \$\begingroup\$ SuhasB888 - Hi, From the wording in your answer "kindly contact us..." it seems you might be representing Infineon. If so, please note the rule about promotion here. You must fully disclose any affiliation with anything you mention or link (i.e. anything you promote) in a post here. If this applies to you, please edit your post and add: "Disclosure: I work for Infineon" (or whatever explains your apparent promotion here). Please also read the tour and help center to see the main site rules. Thanks. \$\endgroup\$
    – SamGibson
    Sep 26 at 8:47

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.