Resistor divider or another VR in front of a VR?

[eigenvictor]

Hello, I know it is not the best way to solve this problem, but just for the sake of learning. Let's say I need 3.3V but my transformers output is very high, say 17V, plus the transformers's regulation is pretty high (drops to 12V at 300ma).  17V s is to high for my favorite low drop 3.3V 662K.  So I tried to use two regulators in series, first regulate down to 5V using  78L05 and then feed 5V into 662K.  But I am getting about 2.3V instead of the expected 3.3V volts.  Any idea why? Headroom shouldn't be an issue.  I measure 5V at the output of the first VG and the same 17V DC at the transformer output. (This is a small cheap electronic transformer  that outputs 12V DC at 300mA).

Also,  I am thinking shouldn't I be able to put a simple resistor divider in front of a VR?    I understand this is inefficient, but shouldn't this work since a VR's input impedance is very high?

[rob77]

what current are you drawing from the 78L05 ? that's rated for 100mA only... furthermore there is a thermal ptotection.. dropping 12V (17-5) at even 100mA will trigger the thermal protection..

[eigenvictor]

It is open circuit.    78L05 seems fine, I am measuring 5V at the output.  This is intended to be a micro controller circuit, so no currents above a few mA.

[mij59]

The output of the low drop regulator could be oscillating, did you use the proper in and output capacitors ?

[eigenvictor]

The output of the low drop regulator could be oscillating, did you use the proper in and output capacitors ?

Haha, I just got back to post about my findings which exactly what you said!  I first connected a 7V lipo battery instead of transformer and it worked fine 7 to 5 to 3.3V.  Then I looked at the output of the transformer on the scope and the waveform was just horrible.  So I connected two additional 10uF caps (transformer to 5V, and 5V to 3V) in addition to a 10uF cap I already had on the micro input and everything is working fine now!  Thank you.

But I still have a question.  Could I just use a simple resistor voltage divider instead of the 5V VR in front of the 3V VR?   
Since its input impedance is very high, that shouldn't cause any issues, correct?

[mij59]

But I still have a question.  Could I just use a simple resistor voltage divider instead of the 5V VR in front of the 3V VR?   
Since its input impedance is very high, that shouldn't cause any issues, correct?

If the load current is constant you could, but the voltage regulator may not like the high input resistance created by the resistor.
Also at no load current the maximum input voltage will be exceeded .
A 3.3V voltage regulator with  the proper input voltage rating would be the best solution. 

[edavid]

1. A voltage regulator's input impedance is not "very high".  If it's a linear regulator, it will be a little lower than the load impedance.

2. You can use a series resistor dropper in place of a preregulator in some situations.  You usually wouldn't use the lower resistor of a voltage divider, since that just wastes power.

[rob77]

So I connected two additional 10uF caps ....

2 x 100nF ceramic is kind of mandatory for a 78xx (one across input the second across output). for 3V3 LDOs it's a pretty common requirement to have ~ 10-22uF tantalum on output to make them stable (e.g. the widely used 1117)

[electr_peter]

How did you set up transformer? Did you use configuration [transformer] -> [full diode bridge] -> [capacitor bank] -> [linear regulators]?
Omitting [full diode bridge] or [capacitor bank] components will lead to high voltage ripple which will result in misbehaviour of linear regulators.