Electronics > Projects, Designs, and Technical Stuff
equation in datasheet seems to be off by a factor of one million
Clear as mud:
I posted earlier when I started this LT7101 switching regulator design. Input is 36 to 59 volts, output is 5 volts at 1 amp. I set the operating frequency for 1540 kHz. I've got the main part of the schematic done, here's a screenshot. I had previously skipped over the input capacitor selection. I worked on that today, and I'm having problems with it.
The data sheet does not really go over maximum allowable ripple current levels on the output or the input. I think it assumes you have those specified already when you start the design. The "Cin Selection" section on page 21 starts by saying ESR requirements drive the capacitor selection, but then it doesn't give any guidance for how to find or make up some ESR requirements. For the capacitance selection, I did some basic calculations of duty cycle and on-time, put in 0.1 volts in for the maximum desired sag in the voltage during each cycle, and got a minimum Cin required value of 1 uF. The data sheet suggests 4.7 uF, so that's in the ballpark. But then it also says you need capability for a ripple current of 1/2 the maximum output current, or 0.555 A considering the maximum overcurrent level. There aren't any capacitors in this value range and voltage rating that have that much ripple current capability, but there are if you go with the suggested filtering tactics, which supposedly reduce the required ripple current by 5 times. Selected bulk capacitor: ESL106M100AE3AA, 10 uF, 100 V, 680 mOhm, 280 mA max. ripple current at 100 kHz.
But here's the problem. Filter inductor and capacitor selection is given by a couple of formulas. I didn't look at the second one yet, but the first one is 5/(2π√(Lf∙Cf)) should be less than the operating frequency. The other formulas in the data sheet make sense when considering the base unit (Farads, Amps) and you have to manually adjust for the SI prefix kilo-, milli- or micro-. But when I put in some numbers for this formula in Henrys and Farads and Hz, I get a ridiculously large value for either the inductor, the capacitor, or both. For example, if I use a 10 uF capacitor, then I need a 100 mH inductor. Which I can find, but not with the required saturation current rating (200 mA). Well, I know my output inductor has enough saturation current rating, so I put that one in (10 uH), but then I need a 68000 uF capacitor! And this after the data sheet says it should be a "a relatively small and inexpensive capacitor," which I take to mean relative to the bulk capacitor.
I've triple-checked my formulas. If I'm making a math error, I'll feel stupid. But I don't see any error in my math. I think it's time to go back to basic principles; where did that formula come from, and is it wrong?
Alternatively, they give an example circuit with input filtering, figure 17 on the bottom of page 35. It's almost completely different from their filter suggestion on page 21 that I've been talking about, but I suppose I could just copy the example.
Previous post is here: https://www.eevblog.com/forum/projects/buck-converter-any-significant-emi-difference-between-lm5010-and-lt7101/
T3sl4co1l:
Show your calculation.
I get:
https://www.google.com/search?q=1+%2F+%28%282+*+pi+*+1540+kHz+%2F+5%29%5E2+*+4.7+microfarad%29
which is rather small actually, but shows that the operating frequency is fairly high so that we won't need much inductance to achieve useful attenuation here.
Tim
Clear as mud:
Oh, I did make a stupid math error! When I put the formula into Excel, I forgot the square root. Since both L and C have the prefix -micro, yep, that would make the equation off by a factor of about one million!
I was just reading the "stupid mistakes" thread, and the first page has several people repeating "don't work while you're tired." I should have just gone to bed earlier instead of asking this question on here; I would have figured it out in the morning anyway!
Clear as mud:
So, this is what I have now:
Bulk capacitor as before, 10 μF with 680 mΩ ESR.
Filter inductor is Taiyo Yuden MAMK2520T1R0M, nominally 1 μH, I think I should be able to expect 0.7 to 0.8 μH under the actual operating conditions. ESR around 45 mΩ typical.
Filter capacitor nominally 2.2 μF, I still need to find one with actual specs for voltage de-rating, I can't have it go below about 0.6 and still meet the criteria.
I'm thinking the ESR provides enough damping without adding another resistor and capacitor for damping. But I guess I should do some calculations or simulation on that.
T3sl4co1l:
--- Quote from: Clear as mud on August 06, 2020, 11:35:18 am ---I was just reading the "stupid mistakes" thread, and the first page has several people repeating "don't work while you're tired." I should have just gone to bed earlier instead of asking this question on here; I would have figured it out in the morning anyway!
--- End quote ---
Often, composing the post itself proves to be adequate rubber ducking. It seems a little more was necessary in this case (or the sleep). ;)
--- Quote from: Clear as mud on August 06, 2020, 01:51:14 pm ---Bulk capacitor as before, 10 μF with 680 mΩ ESR.
Filter inductor is Taiyo Yuden MAMK2520T1R0M, nominally 1 μH, I think I should be able to expect 0.7 to 0.8 μH under the actual operating conditions. ESR around 45 mΩ typical.
Filter capacitor nominally 2.2 μF, I still need to find one with actual specs for voltage de-rating, I can't have it go below about 0.6 and still meet the criteria.
I'm thinking the ESR provides enough damping without adding another resistor and capacitor for damping. But I guess I should do some calculations or simulation on that.
--- End quote ---
Well, the series equivalent is around 1.8uF, 0.75 ohm and 1uH. sqrt(L/C) = 0.745 ohms, so it should be quite nicely dampened, as long as the source isn't a low impedance (that would effectively short out the majority ESR). This can be assured by using a few uH in front (equivalent to a few meters of cable), or putting another one of those 10uF's in parallel with the ceramic. You can always use more capacitance after all.
Tim
Navigation
[0] Message Index
[#] Next page
Go to full version