An oddity I noticed in passing on the Arduino Due schematic, is the Atmel device requires a 4.7 μF capacitor on VDDOUT, but I only see 10 μF capacitors with 100 nF capacitors. How does this work?

  • \$\begingroup\$ Please ask the first part of your question on Electrical Engineering :) \$\endgroup\$ Apr 13 '13 at 9:55
  • 1
    \$\begingroup\$ @Manishearth Isn't it all EE, really? I'm not complaining about the questions, but something about capacitance like this doesn't sound very Arduino-specific. \$\endgroup\$
    – Polar
    Apr 13 '13 at 11:47
  • \$\begingroup\$ @Polar Manishearth edited out the pure EE.SE type portion of the question. This remaining bit is arguably Arduino-centric enough. \$\endgroup\$ Apr 13 '13 at 11:59
  • \$\begingroup\$ Could you add the references to the question ? The schematic as well as Atmel's requirements. \$\endgroup\$
    – asheeshr
    Apr 13 '13 at 12:14
  • 1
    \$\begingroup\$ @AnindoGhosh I saw that, just thought this was still EE.SE \$\endgroup\$
    – Polar
    Apr 13 '13 at 12:18

The datasheet for the Atmel ATSAM3X8E Microcontroller used in the Arduino Due specifies the following about bypassing and stabilizing VDDOUT:

To ensure stability, an external 4.7 µF output capacitor, CDOUT must be connected between the VDDOUT and the closest GND pin of the device. The ESR (Equivalent Series Resistance) of the capacitor must be in the range 0.5 to 10 Ohms.

Solid tantalum, and multilayer ceramic capacitors are all suitable as output capacitor.

A 100 nF bypass capacitor between VDDOUT and the closest GND pin of the device decreases output noise and improves the load transient response.

From Table 46-3 of the "SAM3X/SAM3A Series Complete" datasheet

The capacitor value specified for stabilizing / decoupling a power input or output pin in datasheets is typically a minimum, indicative value - If there is also a maximum acceptable value, that may conceivably be separately specified.

Electronics designers often use a higher value capacitor within the same order of magnitude, if they foresee a high load on the power rail - The higher the capacitance, the better it is able to weather out load-related ripples on the power rail.

The upper practical limit on this capacitor value is set by the initial power-on current that would be drawn from the power pin to charge this stabilizing capacitor - Too high, and the supply may be damaged.

The (optional) 100 nF capacitor specified for bypassing, is for the purpose of providing a short-circuit path to ground for any high frequency noise on the power rail. This is placed in parallel with the stabilizing capacitor, both as close as possible to the power pin concerned.

That is why you see the 10 µF and 100 nF capacitors in the Arduino Due reference schematic.

  • \$\begingroup\$ I am familiar with the reasons to add capacitors, but since this cap is part of the charge pump circuit, then changing the value from the recommended should be done with caution, this isn't just a bulk cap. I'm not sure I've ever seen a maximum value listed on a data sheet for a bulk cap. There are generally standard recommendation for bypass and bulk caps. \$\endgroup\$
    – naven87
    Apr 15 '13 at 14:59
  • \$\begingroup\$ @naven87 Fair point... However, my reading didn't indicate it was being used for a charge pump, it specifically mentions "To ensure stability": Perhaps I was expecting a charge pump purpose to be mentioned in the footnote, but if you're seeing something in some documentation saying this is so, please share so I can review and edit my answer, thanks. \$\endgroup\$ Apr 15 '13 at 15:54
  • \$\begingroup\$ @naven87 I've gone through the Atmel ATSAM3X8E datasheet, the "complete" version, and there is no mention of a charge pump. Also, the Electrical Characteristics states that the device contains a voltage regulator that can supply up to 150 mA to external devices, sourced from the Vcc of 1.8 to 3.6 Volts, and can be software controlled to shut down / start up. That apparently isn't a charge pump there, and the parameters are in the table 46-3 mentioned in the answer. \$\endgroup\$ Apr 15 '13 at 18:56
  • \$\begingroup\$ After re-reading, I can't find what I swore I read, my best guess is I took the comment about cap value being selected for stability to indicating some form of switching regulator rather than a linear regulator which it seems like it is. I will start experimenting, my end goal is to harness the 1.8V supply from the Due board to power other circuitry even though it isn't directly pinned out. \$\endgroup\$
    – naven87
    Apr 16 '13 at 6:36
  • \$\begingroup\$ @naven87 :-) No matter, it pushed me into a detailed study of the datasheet, which I had been putting off for a while. And... yes, you can use the 1.8 Volt output safely for a fair bit of current before the bulk cap is overwhelmed, I would estimate 50 to 100 mA as safe. Do keep watch on the temperature of the MCU while experimenting, and let us know what the experiments reveal. \$\endgroup\$ Apr 16 '13 at 6:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.