Switch Current Limiter 1

271

03:18:47

In DC-DC converters, a "peak switch current" is usually set to turn off the switching transistor when the current through it hits a certain peak value. This is done to prevent an overload or an initial overcurrent. Close the switch to see what happens without current limiting.

In this circuit, peak current limiting is achieved by saturating the switch (spoiler: it's not the best method. More about it later). The MOSFET is put into saturation by an "auxillary" BJT, which turns on and saturates the MOSFET when the drop across the current sense resistor reaches the base emitter voltage of the BJT, just like in this Constant-Current supply: http://everycircuit.com/circuit/5204466287247360
The peak current I = 0.7/Rse. In this circuit, Rse=0.33ohms. So I=0.7/0.33=2.12A.

Is this a good method? It's simple and does the job, right? Well, not really. Yes, it's simple and yes, it does the job, but it doesn't do the job in an efficient way. First of all, sending a huge current through a saturated MOSFET is very inefficient. Second, the MOSFET will overheat if not cooled properly. 

published 6 years ago

hurz

6 years ago

this current sensing is useful for many usecase, here a semi digital skipping methode for a boost converter from @nikisalli http://everycircuit.com/circuit/5449309255630848 there was latest a filtered PWM driver from @faceblast search for it must be on top for search.

hurz

6 years ago

here the thread from @faceblast and his question about filtering shunt voltage http://everycircuit.com/circuit/6541742338473984

PrathikP

6 years ago

I'm not quite sure I understood your comment. Your first comment has something to do with a digital version of current sensing (I'm going to upload something like that in a few weeks, or maybe earlier). Your second comment is about "leading edge blanking" I think, which I may not incorporate, but it was useful.

PrathikP

6 years ago

I think you think that I'm make a current mode DC-DC converter. That not what I'm doing. I'm adding this switch current sensing (like it's done in the 34063) to kinda get a soft start as well as output overcurrent protection.

PrathikP

6 years ago

And talking about the 34063, the datasheet has some formulas to calculate L, C and Peak switch current (towards the end of the datasheet). Are those formulas specific for that IC or can those formulas be used while designing any DC-DC converters? To me it looks like those formulas can be used to determine L,C, Ipk since the formulas use parameters common to all DC-DC converters.

kiani

6 years ago

This kind of current sensing is power wasting, and a Hot resistor. A tiny current transformer ( loop core) in series with xformer ( collector circuit) much efficient and offers a wide range. Signed.. Professional noob.

hurz

6 years ago

34063 you taking about the original motorola all in one dc/dc converter. I dont know what you want exactly so i just bring some thought which i remember from the near past here on EC. @kiani, make the shunt small in milliohm range and spend some amplification and filtering no waste of power.

PrathikP

6 years ago

@kiani I mentioned that it's power wasting in the description

PrathikP

6 years ago

@hurz the 34063 (original or not) datasheet has some formulas. I was just wondering if the formulas for calculating L, C and Ipk values are specific to that converter on not.

hurz

6 years ago

As you already said, its not efficient to bring the mosfet again abd again into a linear regulating mode and burn power. It must end in some strategies to still run the mosfet in switch mode to dissipate as low as possible power but skip for example fron time to time some pulses like @nikisalli is doing, or do some filtering like @faceblast is doing and then on software level in arduino to change the PWM ratio to lower to target average power. Thats all i want to let you know.

PrathikP

6 years ago

I understood that. What I presented in this circuit is one "possible" solution. By that I mean a solution that works in theory. Yes, I know that this kind of current limiting cannot be used in a switch mode converter, as it defeats the purpose of an efficient switch mode converter. Like I said, I'll be uploading a more "sensible" , less wasteful solution later (that won't need pulse skipping). I put sensible in quotes because it won't be very feasible.

add comment in editor