Yet another Opamp based regulator

[banana]

This is not to challenge the well regarded Jung's Regulator. I just want to try something easier.

Schematic attached.

 

[banana]

The prototype PCB.

20Vdc input was supplied by an LM317 pre-regulator.

 

[banana]

A little noise measurement. The MC33078 regulator was compared with an ordinary LM317 regulator, under the same loading condition.

Output voltage 12.8Vdc, 20mA loading.

Red line is MC33078, C load=10uF
Noise level: 7uVrms (BW=500KHz), 5uVrms (BW=20KHz)

Blue line is LM317, C Adj=100uF, C load=10uF
Noise level: 41uVrms (BW=500KHz), 23uVrms (BW=20KHz)

Note that some low noise regulator IC such as LT1762, was specified at 20uVrms.

 

[martin clark]

Now try a larger value C39 to reduce noise gain, to 10uF or more.. I'd also try increasing the values of R21 and R5, so you can make at least EC11 a film cap (= no leakage current noise).

 

[banana]

Never though of leakage current noise... Thanks for the tip.

In fact I've aready tried larger C39, but the change in noise level is not significant. Perhaps it's because the DC gain is only 1.28

C5 compensation cap is a bit larger than expected. At the beginning I used 50pF, but there was static oscillation around 300KHz. C5 might need to be re-adjusted when different kind of Opamp is used.

Also I had worried about stability in case of low ESR cap at the output. Surprisingly, it appears that lower ESR cap actually improve load stability.