Electronics > Projects, Designs, and Technical Stuff
Boost converter + coupled inductor = short circuit protection?
(1/1)
Buriedcode:
Hi,

Many of my projects (and products) use boost converters, and on the occasions I have needed an extra higher voltage, low current rail I've just used a charge pump on the switch node, like getting 9V from a 2v - 5V boost for example.  One could use a coupled inductor to get higher outputs for devices with limited switch node voltage, but neither solution provides short circuit protection.  In standard boosts, the input goes through the inductor and diode to output.  With a charge pump it goes through an extra two diodes.  Assuming these are all schottky you'll get ~ 3x 0.3 = ~0.9 drop, so even a single alkaline AA has enough current to take these out when the output is shorted for longer than a second.

Using a coupled inductor with a boost for a flyback could provide the same output as a single inductor but how would the converter react with its output shorted?  I guess that question should really be - how does a flyback boost converter react to having its output shorted?  Obviously with this there is no longer a direct path from input to output, I'm assuming the peak current limiting of the boost converter would determine the input current. Might be a relatively easy way of preventing damage for when the design can't prevent the output being shorted.


This isn't some specific project, just thinking of methods to add inherent short circuit protection to boost converters without relying on it being built into devices.  For example the MCP1640 is a synchronous boost and claims to have short circuit protection since the output MOSFET is controlled, but the datasheet doesn't provide details of what exactly happens.  Whether it resets, or just maxes out the peak inductor current or what. 

I've noticed that coupled inductors with different ratios than the standard 1:1 are becoming more available (and cheaper) so it looks like a more viable option when multiple voltages are required or isolation, or short circuit protection.
T3sl4co1l:
Yes, it's called SEPIC (or you can make it full isolated flyback, but if it's 1:1 and common ground, you might as well).  With a proper current mode controller, it is short proof.

Tim
Zero999:
Another option is to add another transistor on the positive side, forming a buck/boost converter.
https://en.wikipedia.org/wiki/Buck%E2%80%93boost_converter#Principles_of_operation_of_the_4-switch_topology
Buriedcode:

--- Quote from: blueskull on January 24, 2019, 07:31:34 pm ---..Therefore, with a peak current protected primary side and a shorted load, the load will see pretty high average current.

--- End quote ---

Yeah I've been toying with LTspice and as expected, its essentially a current source which is what the secondary of a flyback is.  Actual short circuit current isn't much of a worry - most of the applications I have in mind are pretty niche/simple affairs, like hand held zener/jfet testers that require >15V but avoiding PP3 batteries (I hate them).  As it will have some connections exposed to the outside world, wanted to make it semi-rugged, rather than just put a resistor in series with the output.


--- Quote from: T3sl4co1l on January 25, 2019, 02:22:02 am ---Yes, it's called SEPIC (or you can make it full isolated flyback, but if it's 1:1 and common ground, you might as well).  With a proper current mode controller, it is short proof.

Tim

--- End quote ---

Yeah, that's what I've seen is the most common solution, but I have never actually designed one.  I'll have to put some boost converter devices to the test (they're all current mode AFAIK).


--- Quote from: Zero999 on January 25, 2019, 10:07:09 am ---Another option is to add another transistor on the positive side, forming a buck/boost converter.
https://en.wikipedia.org/wiki/Buck%E2%80%93boost_converter#Principles_of_operation_of_the_4-switch_topology

--- End quote ---

I didn't think of this, at least not a switch controlled by the converter itself.  I guess it wouldn't hurt to test it out by adding a PMOS driven by the switch node in series with the input, and a diode.

All these options really only add a few extra components, sometimes just one, each with different advantages - I'll have to test them out some time. Cheers!
David Hess:
Ways to solve this:

1. Use a transformer coupled topology like a forward, flyback, or bridge converter.
2. Use a capacitively coupled topology like SEPIC.
3. Add a series transistor to the output so that the bucked voltage has to increase higher than the input voltage as shown below.
Navigation
Message Index
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod