# Help designing an AC to DC converter within specifications

Discussion in 'The Projects Forum' started by HunterDX77M, Sep 30, 2011.

1. ### HunterDX77M Thread Starter Active Member

Sep 28, 2011
104
2
Hello,

I am a few weeks in to my first course in Electronics. We have been assigned a homework project to design an AC to DC converter with the following specifications:
• The input AC voltage is 120 V (rms) at 60 Hz
• The output should be 4.7 VDC with a maximum 1% ripple
• It should reach the final DC output in 80ms or less
• The load resistance is a fixed 470Ω
• The capacitor can be no more that 1000 μF

I have some understanding of all but the third specification (relating to time). We have been doing something like this class so I know I start off with the input into a transformer, to a (diode) rectifier, to a filter, to a regulator and then finally to the load.

I am not however, terribly sure on the specifics here. I especially need help understanding and meeting the time restraint. I'd greatly appreciate any assistance.

2. ### #12 Expert

Nov 30, 2010
16,665
7,313
According to radical 2 C Er F = I
1000 uf will produce 1.25% ripple if the measurement is peak to peak.
If your teacher is speaking in RMS, just build it the usual way.
If the requirements are in p-p, you'll need a Pi filter.
That's 2 caps with a resistor between them.
That is where the time delay comes in.
Calculate radical2 C Er F = I for 500 uf = 2.5%
Then calculate what resistance will form a voltage divider with the other 500 uf to reduce ripple from 2.5% to 1.0%
That is the minimum resistance.
It should work.

Dec 26, 2010
2,147
300
Yes, a 1000μF reservoir capacitor alone won't achieve these figures for peak to peak ripple, but two capacitors each about half that value separated by a resistance can indeed work.

You have been given an estimate for the minimum resistance, but don't be tempted to think that more is necessarily better. How large a resistor is acceptable between the caps will depend on the interpretation of the voltage "reaching its full value in 80ms" Strictly, the voltage never does reach a final value, but you might take R*C = 80ms/5 as a ballpark limit. In practice you might want a smaller time constant to give time for the reservoir capacitor being charged in a cycle or two, allowing for some impedance in the transformer and rectifiers.

The voltage lost across the resistor is also a consideration, and the AC input might need to be raised a little to compensate.

Last edited: Oct 1, 2011
4. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
If you are going to use a regulator like an LM317, you will need small caps; 0.1uF, from both the input and the output terminals to ground. A 10uF cap from the output to ground will improve response to transients (sudden heavy loads).

Did your instructor specify what regulator to use, or are you free to choose whatever regulator you wish?

Are you supposed to use a certain transformer, or can you select any transformer you desire?

A PI filter will help a great deal in ripple rejection, but will not provide regulation by itself.

Dec 26, 2010
2,147
300
The OPs original statement of the problem refers to ripple, but voltage accuracy is not stated, and the load is a fixed resistance. Regulation may not therefore be necessary. This problem may be solved very easily if a regulator is allowed to be used, achieving a tiny ripple voltage and starting up in a cycle or so.

With passive components only, on the other hand, it more challenging to meet the requirements. In particular, if we assume that the reference to the 80ms start up limit presents a significant difficulty it seems likely that a simple filter solution was intended.

6. ### HunterDX77M Thread Starter Active Member

Sep 28, 2011
104
2
Yikes! Thanks for the help, but I guess I should have made it clear that I am completely new to the field of Electronics. If you wouldn't mind, could you please "dumb it down" (for lack of a better word) for me? I'm sorry, it's just that I only recently finished an introductory course on electric circuits and the class I am taking now is an introductory Electronics class. I barely know what a diode is at this point. That is why I need help with this assignment.

7. ### HunterDX77M Thread Starter Active Member

Sep 28, 2011
104
2
Yes, I think a voltage regulator is allowed (I vaguely remember my professor talking about them). And yes, this is probably solved by a simple filter as this is my first assignment (and we are only a few lectures into class). What do you recommend?

Nov 30, 2010
16,665
7,313