Inductor and Capacitor Calculation for Variable input and output buck converter

[sabu31]

Hi all,

I wanted to know what values of input voltage and output voltage should be taken for designing inductor and capacitor of buck converter. Input varying from (Vin_min to Vin_max) and Output varying from (Vout_min to Vout_max).
Typically I have seen application notes using Vin_max for fixed output Vo. But what should be condition for Vo when it varies from Vout_min to Vout_max. The power remaining fixed.
Thanks

 

[KlausST]

Hi,

I recommend to use the formulas given in the datasheet.

It should be noted in the datasheet wheter to use the resulting values as "minimum values" or as "maximum values".

Klaus

 

[sabu31]

Hi KlausST,

I meant the case of designing a buck converter based on discrete components instead of ICs. If the input voltage varies (Vin_min, Vin_max) and output voltage (Vo_min, Vo_max) varies. What are the values of input and output values of voltage to be used for calculating indcutance. The power level is same.

At higher values of Vin_max and Vo_max, the inductor value is higher and current is lower and at case of Vin_max and Vo_min , the inductor value is lower and current is higher.

Now if I use higher value of inductance and calculate turns and core for peak current (at lower voltage), the core becomes larger. So whats the optimum case approximately.

 

[KlausST]

Hi,

physics (and the forumuals) is the same for integrated and discrete SMPS.

Klaus

 

[schmitt trigger]

For a variable output voltage which still maintains constant output power, the highest output current will occur at its lowest output.

Having said this, a bit of advice:
if you are serious about learning how to design a SMPS using your own discrete, you must invest in a good reference book. The requirements for a good design are too many to simply list on a forum post. Choosing practical inductor and capacitor values are the simplest parts of a SMPS design. Choosing the proper main switch drive, and most importantly control-loop stabilization, will be the next and most significant challenges. Not counting thermal management, snubber design, noise, etc.

There are many good books, but to me Switching Power Supply Design by Abraham Pressman and Keith Billings.

Many people also like Switch mode power supplies by Christophe Basso.

 

[sabu31]

For a variable output voltage which still maintains constant output power, the highest output current will occur at its lowest output.

Having said this, a bit of advice:
if you are serious about learning how to design a SMPS using your own discrete, you must invest in a good reference book. The requirements for a good design are too many to simply list on a forum post. Choosing practical inductor and capacitor values are the simplest parts of a SMPS design. Choosing the proper main switch drive, and most importantly control-loop stabilization, will be the next and most significant challenges. Not counting thermal management, snubber design, noise, etc.

There are many good books, but to me Switching Power Supply Design by Abraham Pressman and Keith Billings.

Many people also like Switch mode power supplies by Christophe Basso.

Highest current at lowest voltage. Does this imply that I should choose this value of inductance for further selection of core and number of turns.

 

[FvM]

Did you calculate the effect of using highest or lowest inductor and capacitor values, regarding inductor current waveform and output voltage ripple? Doing so is the first step to answer the question yourself.

 

[Akanimo]

Use a spreadsheet, calculate L and C for different combinations of the minimum and maximum input and output voltages and take the highest of your solutions for L and for C.

 

[BradtheRad]

Even if you wind your own inductor, and lose count of the turns, and it results in a Henry value which is off-spec by 200 percent...,
there will still be a range of frequency and duty cycle for your inductor which delivers the performance you wish.

The important thing is that it does not saturate at the Ampere levels going through it, nor does it overheat.

Example, if you want 10A to go to your load, then the inductor needs to carry peaks of 11 to 20A, depending on the extent to which you operate it in CCM (continuous conduction mode).

- - - Updated - - -

The reason is because it is a triangular, or a sawtooth, waveform (re current) which occurs in the buck converter. The waveform rides a certain amount of DC component.

 

[sabu31]

I am getting highest inductance value for when Vin is Vin_max and Vout is Vout_max (case 1). But at this condition, the output current is lower. When I am using Vin_max and Vout_min(case2), the output current is higher as power is constant with inductance being lower.. However, if i design inductor with value inductance value obtained from case 1 and current from case 2, it would result in large core. So what is a judicious decision/compromise.

 

[FvM]

However, if i design inductor with value inductance value obtained from case 1 and current from case 2, it would result in large core.

Well spotted. It's the natural consequence of designing a converter with variable voltage and constant power. Check how much inductor current ripple you can accept.

 

[sabu31]

Is there any thumb rule / reference material for such designs. Thanks in advance

 

[BradtheRad]

Is there any thumb rule / reference material for such designs. Thanks in advance

SMPS designs I've seen at this forum tend to use inductors in a range of 50 to 500 uH.

Current ripple is reduced as you go further into CCM. It permits you to get by with a low saturation rating in your inductor. One article advised 10 percent as a reasonable minimum value for current ripple. That is why you would look for 11A saturation rating, if you desire to get 10A average throughput.

 

[treez]

Well many solutions will be acceptable and may depend on whether you have a “nominal” vin and vout which occurs most of the time so you want to serve that first.
Otherwise, just make sure when you are at the case where your inductor ripple current is lowest, you have enough delta current to give you a decent ramping current sense signal…..if the ramp is too small, then the ramp may get lost in the noise floor….(I assume you use current mode control)
Sometimes constant off time control can be good to handle wide ranging vout or vin….eg when the range is really wide and you end up with cases where your duty cycle would otherwise be ridiculously large if you stuck with fixed frequency control.
Of course, if you do have a case where the current sense ramp gets very low, then there’s nothing to stop you adding in some slope compensation....but then, dont add so much that you cant work at the opposite extreme.
Likewise you dont want to end up with a case where your inductor peak value is ridiculously high for some cases……
Its just a case of doing it empirically, and then looking at it, and then tweaking it….but until you actually do it, you wont see how to tweak it…I would put it on the sim and check your result there first…there’s no fixed formula