Electronics > Projects, Designs, and Technical Stuff

Overvoltage Protection Circuit : Damaging buck converter

<< < (3/4) > >>

Saboteur94:

--- Quote from: MasterTech on November 27, 2021, 08:57:11 am ---
--- Quote from: Saboteur94 on November 27, 2021, 08:54:43 am ---To replicate the inductor behaviour,can I use a dummy load instead of hooking the converter back up?,just to avoid me blowing up regulators unnecessarily.

--- End quote ---
Yes you can do that.
Also I'm curious the ferrite filter BLM21 is for small currents, but the converter is a 4A one, what is your average output current? It is customary to add a resistor is parallel to reduce frequency ringing in these types of filters:
look at figures 23 and 25 https://www.analog.com/media/en/technical-documentation/application-notes/an-1368.pdf

--- End quote ---

Average output current from the converter is 3A at 5V. However,the input voltage to this the system will be 11V at a minimum so the current through BLM21 will be less than 2A majority of the time owing the "buck" effect.

magic:

--- Quote from: MasterTech on November 27, 2021, 08:26:42 am ---All I can say is that L3 seems fishy, why would you put an inductor like that in a path that can be cut at any time, it should be with C37+ a diode, if necessary at all.
That's why maybe you don't see any overvoltages when you isolate the converter, there a no currents to replicate the inductor behaviour, which could be generating a spike going thru the transistor as avalanche.

L3 should not be there, whether it is the culprit or not.

--- End quote ---
Yes, you have made yourself a nice boost converter. When the protection PMOS turns on, capacitors short the inductor to ground and current ramps up. Then voltage ramps up too and when the capacitors are fully charged, the energy stored in the inductor's magnetic field keeps pushing current forward and overchares the capacitors.

BTW, is there any problem with sensing the output voltage (drain side) of the protection circuit rather than its input?
edit
Nevermind, the protection would probably turn into a crude linear regulator under such arrangement - not good for the MOSFET.

Ian.M:
Yes, Magic has described the problem.   Its going to ring to over twice Vpwr_in as you've chosen a non-polar, so presumably ceramic, Vin decoupling capacitor (C37: 4.7uF 50V).   If its value was stable with respect to voltage, it would ring to a bit under 72V (or whatever the LM61460 breaks down at), but as a high-K capacitor will typically loose two thirds of its value by the time it reaches its rated voltage, if you test with a HV current sink load in place of the LM61460, I'd expect either it will peak at over 150V or (far more likely) the capacitor will fail.   

Note that even if you delete L3, you may get some ringing due to loop inductance, so adding an electrolytic in parallel to C37 to provide damping due to its ESR may be advisable. 

Kleinstein:
The capacitance at the buck converter input is relatively small, so less energy is needed to cause spikes in the voltage. This also cause the current through the P-MOSFET to still have quite some ripple which is not so good for the efficiency.
Besides too high a voltage, though could worst case also be the voltage going negative from the buck converter trying to suck out  the last bit of charge, before it turns off.

L3 could be a problem, especially if the OVP circuit part start oscillating in an unintended mode. There are some circuit parts to act against this, but it may not be enough.

PartialDischarge:

--- Quote from: Saboteur94 on November 27, 2021, 09:20:37 am ---Average output current from the converter is 3A at 5V. However,the input voltage to this the system will be 11V at a minimum so the current through BLM21 will be less than 2A majority of the time owing the "buck" effect.

--- End quote ---

Ok, but at 1.5A or more passing thru that BLM21 its inductance is going to drop quite a lot nulling its filtering effect. You should find a 3 or 4A rated ferrite if you want to maintain a its HF filtering specs. This is described also in the PDF I linked.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version