There are always these tiny nuggets of info i love hearing from experienced folk that I'll never hear otherwise like what you mentioned about the output capacitor on a regulator helping to stabilize the feedback portion to prevent (or encourage in the case of no or poor capacitance) oscillations. Please do more of these tiny tidbits or compile them into a video one day. Edit: you did it again! You mentioned how the resistor divider is located internally for a fixed regulator, but externally for a selectable one. Im sure these are things that will become obvious after the thinking just a little bit, but it's so illuminating when you point it out in the proper context. Thanks for all you do!🎉
I love this feedback, and I'm glad to hear you find these tidbits helpful. I agree it's the little things like these tidbits that can be hardest to figure out on your own. Another tidbit for you. For some regulators (linear or switching) you'll see a small capacitor placed in parallel with the top resistor of the resistor divider. This is called a feedforward cap, and it also helps with stabilizing the feedback loop on some designs. Thanks for watching and for commenting:)
In my design, I’m using a TPS61022 boost converter IC. Designed a test board using all the TI recommended components and values, used suggested layout, and it would self destruct every time, even at low current. Found out from others that it NEEDS the feed forward capacitor which is just an optional thing on the data sheet. That baffled me. Using their actual design leads to a non-functional circuit. Suppose the lesson I learned was to always double check with other’s experiences with a part before using it to see if it works as intended? Adding that feed forward capacitor fixed everything. :)
@@PredictableDesigns The data sheet did have a complex formula to calculate the value of the feed forward cap which I tried to work out, but ended up just throwing on a really small value and it's working great! I also learned about the feed forward capacitor and how it's used to stabilize the feedback loop. Really interesting stuff! Thanks for another amazing video!
You should minimise the area of the SW net whilst preserving its ability to carry the peak output current of the DCDC which is probably 2A max for this device. The smaller the area the less noise. So I would rotate the bootstrap cap by 90 degrees, shuffle L1 closer to the SW pin. Datasheets are typically a bad place to get the best layout for EMC performance. The output cap is less important than the input. The input has the highest switching currents (full on, full off) so you need to keep the loop through C1 and GND as small as possible. Thus I would rotate at least one of the C1s and this will reduce the loop from the IC GND back to VIN through this cap.
There are always these tiny nuggets of info i love hearing from experienced folk that I'll never hear otherwise like what you mentioned about the output capacitor on a regulator helping to stabilize the feedback portion to prevent (or encourage in the case of no or poor capacitance) oscillations.
Please do more of these tiny tidbits or compile them into a video one day.
Edit: you did it again!
You mentioned how the resistor divider is located internally for a fixed regulator, but externally for a selectable one. Im sure these are things that will become obvious after the thinking just a little bit, but it's so illuminating when you point it out in the proper context.
Thanks for all you do!🎉
I love this feedback, and I'm glad to hear you find these tidbits helpful. I agree it's the little things like these tidbits that can be hardest to figure out on your own.
Another tidbit for you. For some regulators (linear or switching) you'll see a small capacitor placed in parallel with the top resistor of the resistor divider. This is called a feedforward cap, and it also helps with stabilizing the feedback loop on some designs.
Thanks for watching and for commenting:)
In my design, I’m using a TPS61022 boost converter IC. Designed a test board using all the TI recommended components and values, used suggested layout, and it would self destruct every time, even at low current.
Found out from others that it NEEDS the feed forward capacitor which is just an optional thing on the data sheet. That baffled me. Using their actual design leads to a non-functional circuit.
Suppose the lesson I learned was to always double check with other’s experiences with a part before using it to see if it works as intended?
Adding that feed forward capacitor fixed everything. :)
Interesting. That feed forward cap helps with stabilizing the feedback loop, but like you said it's usually optional.
@@PredictableDesigns The data sheet did have a complex formula to calculate the value of the feed forward cap which I tried to work out, but ended up just throwing on a really small value and it's working great!
I also learned about the feed forward capacitor and how it's used to stabilize the feedback loop. Really interesting stuff!
Thanks for another amazing video!
Excellent content John. I am Hware academy member
Thanks Hamad! I know your name well and I'm happy to have you as a member.
@@PredictableDesigns The best thing in your content as always is the practical approach
You should minimise the area of the SW net whilst preserving its ability to carry the peak output current of the DCDC which is probably 2A max for this device. The smaller the area the less noise. So I would rotate the bootstrap cap by 90 degrees, shuffle L1 closer to the SW pin. Datasheets are typically a bad place to get the best layout for EMC performance.
The output cap is less important than the input. The input has the highest switching currents (full on, full off) so you need to keep the loop through C1 and GND as small as possible. Thus I would rotate at least one of the C1s and this will reduce the loop from the IC GND back to VIN through this cap.
Ideally, though, use a smaller cap footprint for C1B so that it better aligns with the GND and VIN pins of the IC.
Great video!
Thanks!
Please do a boost smps video also.
In my case 12V to 24V 2A
Like how their example layout does not connect to the EN or PG pins.
Vin can not be constant. For switching regulators it is expected Vin to be within Vout to Vmax range.
Vin can be constant or vary, it doesn't matter. Vout can be higher or lower than Vin for switchers.
Hi and thanks for the video do you mean the polygons by saying copper core? thanks :)
Yes, copper pours are drawn as a polygon then filled with copper.
@@PredictableDesigns thanks a lot :)