Implementing a buck regulator typical schematic - the input capacitor.

Goal


Hello.

I am wanting to switch 48V to 5V with a 0.6A rating, and plan on using the TL2575HV (datasheet here). There is a trivial typical application schematic which I am attempting to implement (doesn't say too much for me then! :confused:):


Schematic

However I'm having a little bit of difficulty selecting the correct components. Currently I've done the calculations for the input capacitor and would REALLY appreciate somebody going over this with me...

CIn
I need a capacitor that has a voltage rating of over 48V, but adding a 20% safety margin gives a value of 57.6V, so anything over that is golden.

Datasheet for regulator states that capacitor RMS ripple current should be:

IC, RMS > 1.2 (ton / T) ILOAD
Where:
ton / T = VOUT / VIN

So I need a capacitor with a RMS ripple current rating bigger than:

1.2 x (5 / 48) x 0.6 = 75mA

I've found this capacitor which seems up to the task (datasheet here). 100uF capacitance and a voltage rating of 64V. All good so far... The capacitor series has a frequency correction factor for the ripple current of:

1kHz = 1.3
10kHz = 1.7

The regulator runs at 52kHz, which seems well above what this capacitor can handle (or at least what the datasheet shows), so I'm unsure if I could use this... Assuming I can I'll continue. The particular capacitor I'm using has:

Ripple Current (120Hz): 250mA RMS x worst case correction factor = 250 x 1.3 = 325mA

So I'm well over the 75mA RMS ripple of the regulator. I need to calculate the ESR value of the capacitor to make sure it would be suitable for the switching regulator, however what "low ESR" is I don't know, bit too much hand waving for my liking...

Tan Delta (120Hz): 0.12

I don't know if I multiply the tan delta value by the frequency correction factor as the tan delta is only given at 120Hz and not 52kHz. If I do, I get a tan delta value of:

Tan Delta (corrected): 0.12 x 1.3 = 0.156

I've used these sources to find the ESR value of the capacitor:

ESR = Xc x tandelta
Where Xc is: 1 / ( 2 x PI x f x C) = 1 / ( 2 x PI x 100000 x 0.0001) = 0.015915494

So ESR is 0.003 Ohms = 3 milliohms, which is nothing, so no real power will be dissipated over the input capacitor. But just to finish the exercise...

Your calculations for the Capacitor are basically correct and you are right that this capacitor will be fine.

Considering the level of detail that you went into in the analysis of the input capacitor, would it be safe to say that you have equally detailed questions about the diode, inductor and output cap?

Cheers,
Terry

BTW, This is a very low ESR for an electrolytic capacitor, you would be able to use a cheaper one that is not quite so good if cost is important to you.

Couple caveats - that diode, it's gotta be a high speed schottky diode, can't use a floor-sweepings diode for that. Also, are you aware of the significance of the layout in switching regulators? They are very sensitive to layout - gotta keep the traces between components as short as possible. Many datasheets even suggest a layout.

Thank you both for the replies.

Terry, yes I most likely will do. I'm just trying to gather as much together as I can about the three other components for the time being. Regarding the capacitor, I don't mind the cost too much as it will only be placed in a couple of dozen of devices. There weren't too many cheaper (without buying 1000+) that could actually fulfil the voltage (48V) and current requirements (0.6A).

DrAzzy. Thanks for the pointers, yes I understand the schottky diode, but I'm trying to study a little more about the reasons and how to come up with a suitable one. I had an idea that the layout was quite significant, and using short traces is something I've seen while doing a very brief scratch of the surface... I can't say I've seen any suggested layouts mind as yet... But there again, I have yet to go looking for that sort of thing.

Questions
Would either of you be able to look at the couple of questions in red. In particular I'm struggling to know if I could use the "frequency correction factor" if it is only given up to 10Khz rather than the 52Khz I'm using.

Also, do I scale the tan delta value by the "frequency correction factor", or is this only used for scaling the RMS ripple current?

Thank you.

Wawa:
DC 12V/24V/36V/48V(8V-58V) to 5V 1A DC-DC Converter Step Down Buck Module | eBay

Thank you, although I wouldn't trust one as far as I could throw it! :slight_smile: Would much rather soldier on through implementing my own with branded components for a much lower price and build up a good working knowledge of them in the process.... Thank you anyways though.