0
\$\begingroup\$

I'm thinking of building and selling a product which uses a DC-DC isolator. I've been prototyping with Recom RI3 3W Isolated DC-DC Converter Through Hole, Vin 5 V dc, Vout 9V dc, I/O isolation 1kV. The latest datasheet suggests an EMC filter using a 10uF 100V MLCC capacitor to meet EN55032. These capacitors are expensive, and so might affect the viability of my product.

For my own purposes, the prototype works fine both without a filter, and with an aluminium capacitor in place of the MLCC one. Are there other types of capacitor I could use which might enable me to achieve FCC and CE compliance while keeping down the cost of components?

\$\endgroup\$
10
  • 2
    \$\begingroup\$ If the commercial viability of your product relies on the penny or so difference between an MLCC capacitor and other potential capacitors then you are probably in the wrong game. \$\endgroup\$
    – Andy aka
    Commented May 1, 2019 at 12:02
  • \$\begingroup\$ I'm seeing MLCC capacitors for several dollars each, aluminium for cents. Maybe I'm in the wrong shop... \$\endgroup\$
    – OutstandingBill
    Commented May 1, 2019 at 12:05
  • 2
    \$\begingroup\$ What specific capacitor value and voltage rating are you looking at? The datasheet lists a 10 uF, 100 V MLCC (which is big and expensive) at the input which seems odd for a module which has a Vin max of less than 30 V. In my opinion you could try low ESR alu. caps but I would recommend to add a smaller value MLCC cap in parallel with that to help with RF performance and avoid EMI issues. For example a 470nF, 50 V MLCC. Note that guaranteed FCC compliance means: using the suggested cap or doing something else and then measure it to check compliance. \$\endgroup\$
    – Bimpelrekkie
    Commented May 1, 2019 at 12:13
  • 1
    \$\begingroup\$ +1 for that @Bimpelrekkie. You could read the last sentence in 2 ways. I think the right interpretation is: Note that guaranteed FCC compliance means:using the suggested cap or doing something else. In either case you need to measure it to check compliance. \$\endgroup\$
    – Huisman
    Commented May 1, 2019 at 12:45
  • 2
    \$\begingroup\$ @Huisman I actually meant that you could (with acceptable risk) skip the FCC compliance and just do what the manufacturer suggests (so use the expensive cap). I'm saying that you don't need to check FCC compliance if you know your design has no issues. For example: the box around your product is a Faraday cage. But I agree, for any serious mass production (and resulting liability) you would NEED verified FCC compliance. \$\endgroup\$
    – Bimpelrekkie
    Commented May 1, 2019 at 12:55

3 Answers 3

Reset to default
2
\$\begingroup\$

You could use a solid polymer (low ESR) capacitor, a low-ESR aluminum capacitor or parallel MLCC capacitors of lower voltage and/or capacitance. Probably film caps with MLCC in parallel too, but they're not so viable these days.

Keep in mind that the voltage coefficient of capacitance is quite large (and negative) with MLCC caps so using them near the rated voltage will result in their capacitance value typically being much less than the nominal value with a small DC bias. Similarly, the inductor has to be chosen so that it's not close to saturation.

As an example, suppose you tried to use two CL31B475KBHNNNE Samsung 4.7uF/50V caps in parallel (about 7.5 cents USD each). The capacitance at 30V will be nowhere near 10uF....

enter image description here

That's probably why the datasheet specifies 100V rating.

There are DIY LISN designs out there if you're concerned about pre-compliance testing for conducted noise. Note that particular DC-DC has an unusually low switching frequency, which may be of benefit, depending on which standards you are trying to meet. Testing is a must in any case.

\$\endgroup\$
1
  • \$\begingroup\$ (1) If actual Q = 15% of CV rating, for Vc>10%Vmax that implies Q becomes fixed above 10% Vmax, meaning it cannot hold any more charge. Does it mean any excess Q injected is lost in ESR heat and radiated energy, since C drops with rising V and Q stays constant at 15% of rated Vmax * C max. (2) A better question might be, how does ESR * C * V affect cost for each chemistry and how to optimize cost. \$\endgroup\$
    – D.A.S.
    Commented May 1, 2019 at 14:09
1
\$\begingroup\$

You could consider aluminum solid polymer e.g. Vishay 184 CPNS series has 10µF 100V version and Panasonic SXV series even has a 15µF 100V version. I don't know the cost but would think they are considerably cheaper than a big MLCC (and probably easier to source).

\$\endgroup\$
0
\$\begingroup\$

You may have to learn the hard way about EMI, noise, RF filters and SMPS... Because doing it well is hard and we don't have near enough information to know what your layout and effect of cables between DC and boost regulator and other design info will be.

Or you may be lucky, and never have any problems because you probably are only going to use < 10% of the 333 mA rated boost current on the 5 to 9V converter for Op Amps.

The 10uF Alum Electrolytic Caps on the LC "Pi" filter act as current buffer for the 20 to 50kHz switching boost regulator as well as attenuating noise going out to your 5V source and radiating on all your cables like an antenna. So it is bi-directional dual purpose.

enter image description here

enter image description here

These Korean electrolytic caps might do the job as well. REF https://www.tme.eu/en/details/ce10_35-smd/85degc-smd-electrolytic-capacitors/

There are combinations of impedance and reactance that can work well and others than amplify the noise DC input cable inductance shown above. ( I picked a length random.)

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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