Buck Converter Prototype

Thread Starter

Papabravo

Joined Feb 24, 2006
19,309
I recently discovered Power Electronics, by Daniel W. Hart as a free download from:

https://civildatas.com/download/power-electronics-by-daniel-hart and other sources

This book was published in 2014, so it is not a new work. I was drawn to the writing style and the examples that are presented in the text. In addition to covering the basic topology of many different SMPS types there is an emphasis on component selection. As an exercise in collaboration, I want to show the power stage for a buck converter and use it as a springboard for the staged development of a working converter.
1664999548907.png
This exercise shows a typical power stage for a buck converter along with the selection of the inductor and the capacitor. The simulation results show a pretty good match with the calculated values. I'm not entirely sure what the next step should be, but will doubtless include but not be limited to:
  1. Selection of a MOSFET switch
  2. Characterize the open loop response
  3. Decide on a control strategy
  4. Characterize the closed loop response
  5. Implement a loop compensation strategy
  6. Ensure an orderly startup phase
  7. Ensure stable and predictable operation
 

Attachments

eetech00

Joined Jun 8, 2013
3,418
Hi

So you have the frequency, DC, Vin, Vout, Iout, and ripple.
Next decide on operating mode, Voltage mode or Current mode, then choose a controller with and without an internal switch that supports your design spec and includes any extra features you may want (like shutdown or enable). Of course, cost should be considered. Since the output current is only 1A, the should be many devices to choose from. One that comes to mind is the LM2596 (but 150Khz). Also, in general, the higher the switching frequency, the smaller the inductor. Does the frequency have to be 20Khz? Or is that flexible? You see there will be trade-offs.

Hope I didn't misinterpret your post.
 

Thread Starter

Papabravo

Joined Feb 24, 2006
19,309
Hi

So you have the frequency, DC, Vin, Vout, Iout, and ripple.
Next decide on operating mode, Voltage mode or Current mode, then choose a controller with and without an internal switch that supports your design spec and includes any extra features you may want (like shutdown or enable). Of course, cost should be considered. Since the output current is only 1A, the should be many devices to choose from. One that comes to mind is the LM2596 (but 150Khz). Also, in general, the higher the switching frequency, the smaller the inductor. Does the frequency have to be 20Khz? Or is that flexible? You see there will be trade-offs.

Hope I didn't misinterpret your post.
None of the choices are fixed. The problem was taken from the mentioned text, as an example to show how to pick the inductor and the main filter capacitor from a set of specifications. We can change anything, but we are a long way from picking a chip because I want to examine the pieces inside the chip via simulation to promote understanding of the power conversion process.

Some of the choices in the problem may be less than practical. A possible contribution would be to offer a revised set of specifications for a more practical starting point for a device that would go beyond the simulation phase. There was a long-lasting thread on an EHT Power Supply if I recall correctly.

https://forum.allaboutcircuits.com/threads/eht-power-supply-design-and-construction.113504/

That thread has 2,146 posts so far. I should live so long.
 
Last edited:

Thread Starter

Papabravo

Joined Feb 24, 2006
19,309
So, I went looking for an external MOSFET switch by looking at the array of simulation models in LTspice and I started with a device that had:

\( V_{(BR)DSS}\;=\; 100 \text { V} \)
\( R_{DS(on)}\;\le\; 2\text{ m}\Omega \)

1665029633013.png

I found the IPT015N10N5ATMA1 at Digi-Key for the heartburn price of $8.14 for Quantity 1. How may in stock you ask? (You didn't really need to ask). Goose Egg, Nada, Zero. I suppose I could try to find a part in stock but by the time I need it who knows what the availability will be.

Would you now go for a device you can get, but may not have a simulation model or would go through the list of simulation models you have looking for one in stock?

Just an everyday engineering conundrum.
 
Last edited:

Thread Starter

Papabravo

Joined Feb 24, 2006
19,309
Why a 300A part when only a 3A part is needed?
I was only looking at Rds(on). It is actually somewhat difficult to select a part if you can't look at more than one thing at a time after the list has been sorted for all the parts with B(BR)DSS = 100 V. What 3A part did you have in mind?
 

eetech00

Joined Jun 8, 2013
3,418
I was only looking at Rds(on). It is actually somewhat difficult to select a part if you can't look at more than one thing at a time after the list has been sorted for all the parts with B(BR)DSS = 100 V. What 3A part did you have in mind?
Depends on the peak switching current supported by the controller.
 

Thread Starter

Papabravo

Joined Feb 24, 2006
19,309
So, I decided to search Digi-key for an N-channel MOSFET, with a V(BR)DSS of 100V and a maximum drain current of 3 to 3.3 Amperes that was in stock and less than $2.00 in single quantity. I found the FDT86106LZ from Onsemi.
https://www.onsemi.com/pdf/datasheet/fdt86106lz-d.pdf
Digi-key has 27,686 in stock @ $1.48 in single quantities.

Is there a simulation model -- you ask? Not with the standard LTspice distribution, and not in the extensive libraries of @Bordodynov but there is a pspice model on the Onsemi website, which I was able to use with LTspice without incident or incompatibility. This is a fairly common occurrence, but there are some exceptions.

The initial simulation results look OK for now and the switch losses at 1 A amount to 108 milliwatts at the maximum rDS(on).

1665073742040.png
 

Attachments

Thread Starter

Papabravo

Joined Feb 24, 2006
19,309
I tried really hard to characterize the open loop response of the prototype buck converter from the original problem in the book on Power Electronics by Daniel W. Hart which I mentioned in Post #1. It turns out that the open loop response at the low switching frequency is quite difficult to get a handle on because of the low switching frequency and the required values of the inductor and the capacitor. Raising the switching frequency to 100kHz and recomputing the values of L1 & C1 helps to make the open loop response more manageable. The MOSFET I introduced in post #8 seems to be just fine for this application.

1665113534678.png
 

Attachments

Thread Starter

Papabravo

Joined Feb 24, 2006
19,309
So here is the first shot at doing the open loop response of the 50V to 20V, 20 Watt Buck converter

1665251131910.png
I'm not entirely sure why things at low frequency appear to be so squirrelly. The peak between 2K & 3K is a bit of a surprise. The gain rolloff is much steeper than I expected, but I think I know why that is the case.
 

Attachments

Top