0
\$\begingroup\$

My goal is to suppress transient voltage spikes from damaging the 5V, 600mA Step-Down Voltage Regulator D36V6F5. I believe the transients are a result of high inductance in the wires between the power supply and the circuit.

enter image description here

There are a couple of solutions listed in Pololu's guide to transients, but only adding a series resistor or a large electrolytic cap is possible in my case.

The resistor would result in power loss and a cap is just physically very large.

So my idea was to use a TVS diode -- SMF48A-TP. Now ignoring the fact that the maximum breakdown voltage Vbr is 58.9V, which is already too much for the D36V6F5, whose limit is 50V, I'm still seeing spikes of up to 65V.

Why is that?

enter image description here

I don't have the circuit for the D36V6F5, but I do know that it has 5.3 uF and 100 nF capacitors on the input.

\$\endgroup\$
3
  • \$\begingroup\$ Where are these transients coming from? Also, realize that the voltage drop across a TVS diode in breakdown varies with current. More current = more voltage drop even if it's clamping. \$\endgroup\$
    – DKNguyen
    Sep 5, 2019 at 1:19
  • \$\begingroup\$ Experiment with capacitors. You may not need a resistor. An electrolytic capacitor acts as a damper for the transient. Just try whatever size you can fit comfortably and see if it works. \$\endgroup\$
    – user57037
    Sep 5, 2019 at 6:25
  • 1
    \$\begingroup\$ Also, if the input voltage is 50V, and the output is 5V, then the input current will only be around 70-75 mA. You can put a 10 Ohm resistor in series with 75 mA no problem. That is only 750 mV of drop in a 50V input. 10 Ohms plus the existing 5.3uF already on the D36V6F5 will kill any inrush transients. If you aren't comfortable with 10 Ohms at least try 4.7 or 2.2 Ohms. There has to be some amount of voltage you are willing to drop to get rid of this problem. \$\endgroup\$
    – user57037
    Sep 5, 2019 at 6:29

1 Answer 1

Reset to default
1
\$\begingroup\$

The datasheet specifies 58.9V at 1mA, the very beginnng of conduction, but voltage can be as high as 77.4V (maximum) at 2.6A. Maybe the energy of the spikes generates more than 1A transient current.

What other loads are connected to the 48V line?

A possible solution would be using a dedicated cabling from the power supply to the regulator, possibly including ground, so no heavy current changes circulate thru the cables.

Another one: add more external C and if possible some series inductance in the cable from the supply. The datasheet mentions "If you are connecting more than 28 V or your power leads or supply has high inductance, we recommend soldering a suitably rated 33 μF or larger electrolytic capacitor close to the regulator between VIN and GND". Do this little by little, sometimes adding an integrator like that will slow down the peak voltage but may increase the peak due to sub dampened "resonance".

Yet another hint: place the TVS from collector to base of a NPN high current transistor, 100 ohm from B to E, and connect its C end E in parallel with the power lines. Effective clamping voltage will be reduced because hfe times less current will pass thru the diode.

\$\endgroup\$

Your Answer

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

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