5
\$\begingroup\$

Background information:
I tried hard to get a Schmitt-trigger working, but it kept failing. By accident I figured out, that my AC-to-DC adapter produces a very irregular voltage output (I could hear the 50Hz very clearly). After putting a strong electrolyte capacitor (over 4000 microF) on the DC output, it started working.

However, I'm asking myself, if the DC output it stable enough that way, or if I could experience difficulties in the future when building (let's say) a radio.

What are common ways of stabilizing the DC output of such an adapter?

\$\endgroup\$
8
\$\begingroup\$

What is the voltage and current rating of the supply? Is there any other load on the supply? Is the supply a regulated or unregulated supply?

Most of these supplies require a minimum load before the output voltage is within a specified range. Other than the Schmitt trigger, was there any other load on the supply? If not, you are probably not loading the supply to the specified minimum current.

In most applications these supplies are followed with a linear regulator to provide a stable voltage. You still need a minimum load and some filtering capacitors. 4000 µF is excessive and possibly dangerous. Typical capacitors are in the 1-100 µF range.

The size and type of capacitor is dependent on the linear regulator. Follow the manufacturers recommendation exactly. Some devices have very specific capacitance and effective series resistance (ESR) requirements.

\$\endgroup\$
  • \$\begingroup\$ The supply is unregulated. 5V and 350mA from 230V AC. MY trigger was the only load. \$\endgroup\$ – java.is.for.desktop Jan 24 '10 at 9:33
  • 2
    \$\begingroup\$ You should lookup the specification on the supply. I would be very surprised if you have enough load on the supply for specified performance. Once you find the minimum load requirement add a resistor to draw the minimum current. \$\endgroup\$ – jluciani Jan 24 '10 at 14:21
  • \$\begingroup\$ if it's just a wall-wart, good luck. I don't think I've ever seen specs for those. \$\endgroup\$ – Connor Wolf May 13 '11 at 22:02
2
\$\begingroup\$

In my understanding, the more load you draw off a power supply, the less effective a particular-sized capacitor will be at stabilising the voltage. So, if you had a low load, the current capacitor might be fine, but if the "radio" had a decent amplifier and speaker incorporated the load might be enough to reduce the capacitor's effectiveness and cause noise to make its way into the circuit. The solution in that case is to increase the capacitor size.

If the DC output from the adapter is higher than what is needed in the circuit, another option that will help stabilise the output is to regulate the voltage down as required, after the capacitor, using a standard voltage regulator.

\$\endgroup\$
2
\$\begingroup\$

Make sure you use a power supply which is properly regulated for your application. (jluciani's comments are good ones) Ripple voltage should be one of the specs.

If it's not good enough for your purposes, get a different one -- you can't really fix a power supply in a way that's guaranteed to work in all cases.

\$\endgroup\$
1
\$\begingroup\$

Stabilizing the output voltage is also called "regulating" the output voltage.

I suspect you have at least one of the following problems: failing power adapter, unregulated power adapter, or insufficient bypass capacitors.

Failing power adapter: Some power adapters start to hum more and more loudly just before they fail. You may need to replace it soon. Are you the kind of person that: (a) waits until a light bulb fails, then runs around in a panic trying to find a replacement? Or (b) has a replacement light bulb in a closet so when the light bulb fails it can be replaced within minutes? Or (c) pre-emptively replaces lightbulbs when they are near the end of their expected lifetime?

On the other hand, a few kinds of power adapters always hum relatively loudly, even when they are working perfectly.

Unregulated power adapter: AC power in the UK and Europe is held to +-10% steady-state, and often has short-term surges and sags. The 74LS132 datasheet only guarantees it will work with a power supply of 5 V +- 5%, so already the +- 10% change in voltage from an unregulated power adapter is out-of-spec.

Many other Schmitt trigger chips work over a much larger voltage range. However, the power supply voltage affects the trip voltage -- so while the chip may technically may be inside spec and "work" in some sense, it may not do what you want.

The big capacitor helps smooth out short-term surges and sags, so it temporarily masked over your problem, but it does nothing for the longer-term wandering.

You need a regulated power supply, especially if you are working with audio and radio circuits. Either (a) post-regulate your unregulated power adapter with a voltage regulator such as a 7805 and some capacitors or a LM317 and some capacitors, or (b) return your unregulated power adapter, and replace it with a regulated power adapter that includes a voltage regulator built in.

Insufficient bypass capacitors: Best practice is to attach a 1 µF capacitor to the power pins of each chip, plus a larger capacitor somewhere on the board. (Before 2001, best practice used 0.1 µF capacitors). Digital chips often pull brief pulses of high current, which are easily supplied by a nearby capacitor. Without capacitors, the resistance and inductance of the long path between the chip and the voltage regulator often allows the voltage at the power pins of the chip to sag enough ("VCC droop" and "ground bounce") to reset chips with hysteresis such as Schmitt triggers and flip-flops and microcontrollers.

See question "What is a decoupling capacitor and how do I know if I need one?" for more details.

\$\endgroup\$
1
\$\begingroup\$

Sparkfun has a good tutorial on power supplies that covers these types of issues.

[dead link , was from 2008, now archived at (https://web.archive.org/web/20080802115655/http://www.sparkfun.com/commerce/tutorial_info.php?tutorials_id=57) ]

\$\endgroup\$
  • 2
    \$\begingroup\$ Link is dead. It's better to put relevant content in the answer so if the link dies, the answer is still relevant. \$\endgroup\$ – Erik Philips Oct 30 '13 at 9:11

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.