DC input filtering and protection

[mho_logos]

Hello All,

I am designing a product which will take a DC supply input (10-40V), powering a buck-converter with 5V 4A output. The product will be powered from externally (probably from a 24V switch mode), but in some installations, the DC power cable could be fairly long, so I need to design the DC input to handle various EMI related problems.

My old design uses an SMAJ58CA TVS on the DC input, followed by a Toko ECM85 common mode choke.

I am worried that the SMAJ58CA will not withstand surge testing at +/-2kV, and whether I need additional protection on my DC input. Should I use a varistor or other component, alone or in combination with a TVS.

I have spent a great amount of time searching, but I have not found much regarding the design of DC power inputs.

Does anyone have suggestions for a design that will provide adequate protection against surge and burst, and also limit conducted emissions.

Any pointers to circuits, suggestions of components, appnotes etc. will be greatly appreciated.

/Mikkel Holm Olsen

 

[mho_logos]

I have found some suggestions of surgectors, but I have trouble determining when to use surgectors, varistors and or TVS diodes.

Also, it seems I should add a fuse (high current rating slo-blow type) before the varistor, since if it fails, it will short circuit. Any comments on that (I suppose the same could happen for the TVS)?

/Mikkel Holm Olsen

 

[FvM]

A TVS diode is pretty good in handling surge events, but did you also design your power supply to handle the about 75 V present during the surge?

Adding a fuse is generally a good idea, but it doesn't solve this particular issue. It would only trip in case of a long enduring overvoltage, or e.g. input reversal.

 

[mho_logos]

Thanks for the replies.

I hadn't thought about the voltage present during the surge. My intention is to use a National Simple-Switcher to build the 5V 4A power supply. Many of these have a maximum input voltage rating of 42V. The external DC supply will be 12V or 24V, but allowing a slightly larger range would allow the use of a 24V unregulated supply as well. Still, I suppose it is reasonable to drop the clamping voltage somewhat from that of SMAJ58CA. E.g. if I specify an input working voltage of 10-32V, I could lower the clamping voltage, so the DC-DC will not suffer during the surge event.

The fuse was not meant to protect against the surge, but to protect the power supply if the TVS (or MOV) has failed short after excessive surges.

Regarding reverse polarity, the SMAJ58CA is bidirectional. The old design uses a full bridge rectifier after the ECM85 common mode choke. In the old design there was sufficient input capacitance to allow this to be run from an AC supply, but this application is specified for DC input only, so I want to get rid of the bridge rectifier and just use a single diode for the polarity protection, which will avoid the voltage drop in the ground connection (I am concerned about this in case someone decides to use the same DC input voltage to power my application, as well as some external equipment, which would result in differences in the signal ground - but even without the bridge, I will still get ground loops when doing this :sad.

 

[FvM]

Even reducing the specified input voltage to 32V maximum and using e.g. a 36 or 39V TVS can't strictly keep the 42V rating of standard Simple Switchers. There's no easy way. You can e.g. go for high voltage switchers or automotive types, that are designed to sustain a "load dump" surge. Or rely on some extra overvoltage margin of standard devices...

 

[haihun]

how much is the rated power of SMAJ58CA? the rated current for 8/20 waveform?
for the 2kV surge test, the coupling resister is?
i think you should use a vr with higher rated current.

 

[grid]

According IEC61000-4-5 level 3 test voltage 2kV coupling resistance 2 ohm:
A good starting point is high surge smd varistor 12J between (+) and (-) 24V DC input lines.
Can be found from Wurth catalog.

 

[mho_logos]

Thank you all for the replies.

I have seen a design that uses the common mode choke first, and a TVS afterwards. This way the common mode choke actually helps reduce the surge pulse before it reaches the TVS. Is this a good idea, or are there caveats?

Also, for this to be effective, I suppose the common mode capacitor on both sides of the common mode choke should be rated for the high voltage?

Also, I have decided to lower the TVS voltage and try to find a switching regulator which will accept the higher input voltage.

 

[haihun]

i think you should add a vr before the common mode choke,so the filtering and protection project is :VR+choke+TVS

 

[se06745]

I had the same problem recently, and knowing that this post is too old, if some one has the same doubt, I used the following circuit and it worked fine in the EMC laboratory.

For Surge and Burst protection:
(INPUT) + MOV + Inductance (Delay signal) + TVS +

For driven noise through the cables:
Capacitor + Common mode Choke + (LOAD).

For driven noise I had problems at 600KHz due to switching LED drivers. We solved with a common mode choke and a capacitor that was necessary to adjust the output impedance for the noise.

After that, we applied burst test and the TVS worked fine, but for surge it was not enought powerful. Then we put a MOV for surge test, and ferrite (or inductor) with the TVS in order to delay the signal to give time to the MOV to act,before the TVS.

I hope it helps.

Regards.