Voltage regulator questions - capacitor on the input, difference between different regulators

Hi Tomasz,

If you can get by with one Amp you may want to check out these Recom Convetors. Since they are switching regulators they do not waste as much energy in IR loss. Further there is only a 1.5 volt drop out and probably no need for heat sinking. With efficiency up to 94% more of your battery power will go to your application.

http://www.newark.com/recom-power/r-785-0-1-0/dc-dc-converter-5w-5v-1a-sip/dp/62X9542

Incidently if you need more than one amp these devices can be paralleled.

John

Thanks! That's really helpful, I'm learning a lot. I wasn't aware of the fact that the linear regulators have such low efficiency. It would drain my batteries in no time. I really need to buy some books and learn about this stuff...

Anyway, I guess I'll go with the switching regulator you've suggested. I'm pretty sure that I'm not going to drain more than 1A - I have to power up a small LCD screen that's supposed to draw 500mA, and a Raspberry Pi with no peripherals (I think it's supposed to draw around 300mA).

Worst case I'll connect them in parallel (one for the screen, one for the raspberry). But I think just one should be fine. I'll measure that later.

However, there still is a question of what capacitors to use... neither the spec nor the "book of knowledge" really specify that (or I can't find it again). Maybe it's some common knowledge that I'm lacking. Can you by any chance give me some advice on that?

Also in the spec it mentions to add a diode to prevent current going back through the regulator, but since I'm only powering a small LCD and a Pi I probably don't need that, right?

Hi Tom,

Now that there is a bit more information!! beware that you can't assume 1A is sufficient if you're using the Pi 3.

According to this link the current consumption (I expect with nothing attached) during a reboot reaches 720mA.

If you make the assumption 1A is enough for the Pi 3 I think this is just waiting to fail once your parts arrive.

If your LCD screen really does need 500mA as you say, then you definitely need a higher power DC-DC converter

Anyway it is unclear what Pi you're using.

Also, there are LiPo USB power bricks which would power the Pi and display (save on needing 6xAA batteries).

I'm using Rasperry Pi Zero - from what I've read, it's got a really low current draw, ~200mA. I've read on some blog how someone was testing it's draw and it peaked at around 450mA under really heavy load. So that's why I hope to fit in the 1A with both (worst case I can dim the backlight on the LCD screen by 50%, that's supposed to drop it's draw to around 370mA).

Hi Tomasz,

The Data sheets are not real clear on this I agree. The problem seems to be that the need is very application dependent. I would recommend a 4.7 uF to a 10 uF on the input of the Converter. While they say it is optional I would also put a 47 uF on the output as well as a 0.1 uF ceramic. This is in line with the recommendations from shabaz. The diode is also optional but if you have one why not afford protection if it is available.

John

Hi Tomasz,

For the fun of it I put 8 used AA batteries in a pack and powered a 5 V Recom Converter. Here is a picture of the set up. The little voltmeters each draw 20 mA.

I did not put any capacitors on the input or output in this case.

John

Hi Tomasz,

For the fun of it I put 8 used AA batteries in a pack and powered a 5 V Recom Converter. Here is a picture of the set up. The little voltmeters each draw 20 mA.

I did not put any capacitors on the input or output in this case.

John

Interesting. I guess the capacitors help with filtering and stuff? Which is something you would not see on a voltmeter?

Anyway, thanks a lot for the help! I'll put the the capacitors there anyway, because why not and the diode.

shabaz - thanks a lot for the help as well!

The larger the capacitor the lower the frequency it tends to filter efficiently

Many people question why most schematics show a 10 or 100uF capacitor in parallel with 10 to 100nF capacitors, they think simplistically about capacitors in parallel and series (Much like resistors in parallel and series) but they do not consider the parasitic impedance / reactance etc at different frequencies

For instance, a Large electrolytic at high frequencies can have a very high reactance (Think Resistance) due to the inductance of the capacitor (Its made up of a coil of metal after all) and is therefore useless at filtering the noise from all the digital switching that can be into the Mhz or even higher. Putting a smaller nF or pF capacitors in parallel to the electrolytic solves this problem, so they compliment each other

so looking at this view

you can see how this happens, the frequency increases, the capacitor impedance lowers but the internal inductor impedance increases, at a certain point the impedance from the inductor becomes significantly more than that of the capacitor

In a power regulator like the one we're talking about in this thread, the larger capacitor is for smoothing of the typical mains voltages and to reduce low frequency oscillations, the lower value caps reduce the high frequency noise from the load and other sources.