2

I know that USB PSU's are regulated but does that mean I really don't need anything at all in my circuit to have reliable operation?

Now my circuit will be running 24/7 and should be as reliable as possible. The most sensitive part and also central part is an ATMega MCU.

I'm still fairly a noob and would like know if there still could be any hazards coming from USB PSU's and what would you do to avoid them?

There are several questions about USB PSU's already but some of them have answers that give me some doubts about the general short "yes, this will work" answers.

For example I read in this answer that a charger might actually disconnect you at some point, is that true? If yes, can I do something to prevent it? https://electronics.stackexchange.com/a/99187/17371

timonsku
  • 1,368
  • 3
  • 14
  • 30
  • What AVR part, and what other circuitry will be surrounding it? – Matt Young Mar 23 '14 at 00:35
  • How much current do you need, and what are the voltage tolerances? – Ignacio Vazquez-Abrams Mar 23 '14 at 00:38
  • Everything in the circuit operates at 5v and the whole circuit will will draw at max around 200-300mAh but usually no more than 50mAh. The circuit can have components plugged into (simple sensors with not much current draw). The ATMega will be an Atmega328/1684/1284. Can my question only be answered for the specific case or can there be a general answer for this? – timonsku Mar 23 '14 at 00:48

1 Answers1

5

If you are running the AVR at 5v, then no, no extra regulation is needed. This is of course not talking about neccesary parts like decoupling caps. Any ic should have a ~0.1uf cap near its power and ground pins. If you are using the analog features of the AVR, also keep in mind the recommended board layout, decoupling caps and inductor.

Depending on how long your power cable is, you may want to add a 10 or 100uf cap at the input to your board.

Keep in mind that Usb allows ± 0.25v within spec. And some chargers are better than others. Some have fairly choppy/noisy regulation, so you might want check it on a scope to see how bad or good it is at keeping near 5v.

And that answer involves USB spec host behavior. USB chargers are not usb hosts. That's not an issue when using a USB charger as a simple 5v power supply.

Passerby
  • 72,580
  • 7
  • 90
  • 202
  • 1
    Be careful putting a large cap on the power rail. The USB inrush current is spec'd at 50uC, which is 10uF. Any larger, and you'll need to add some kind of current limiting. http://www.testusb.com/inrush_issue.htm It's also worth adding a ferrite bead and ESD clamp. – Matt Young Mar 23 '14 at 03:53
  • Thanks exactly what I was looking for. I see those tiny capacitors often used regarding power, what are they for? – timonsku Mar 23 '14 at 04:58
  • 1
    @PaulGreen think of them as tiny batteries. They fill in the power faster than the power supply can. This prevents noise (by filtering the power input at the IC), and they prevent brown outs, as an IC can pull large amounts of current faster than the power supply can keep up, dropping the voltage to brown-out levels. They are essentially necessary. – Passerby Mar 23 '14 at 15:40
  • Oh I see, I always thought they had a different purpose as they are so extremely small. Always thought you would need a 10uF~ cap for this sort of things. But isn't this rendundant if I also use the said 10uF cap for the pcb power input that you also mentioned? – timonsku Mar 23 '14 at 16:13
  • 1
    @PaulGreen redundant is good. And not really. The decoupling caps on the IC inputs are for the IC stabilization, while the 10uf at the pcb input is really for smoothing out the noise from the cable run. The 10uf is optional depending on the supply, but the decoupling 0.1uf is mandatory from any engineering standpoint. – Passerby Mar 23 '14 at 17:12
  • Thats good to know, I will read into decoupling a bit to understand this better. You were really helpful, thanks! – timonsku Mar 23 '14 at 21:27
  • @PaulGreen see the answer to this question as a good primer on decoupling http://electronics.stackexchange.com/questions/15135/decoupling-caps-pcb-layout – Passerby Mar 23 '14 at 21:49