6
\$\begingroup\$

I'm using a AMS1117 fixed 3.3V linear regulator and I want to increase its output voltage to about 4V. The input voltage to the 1117 is in the range of 9-10V.

Wikipedia suggests that it's possible to put a zener or resistor on the ground terminal of the regulator to change its output voltage but it provides no citation or reference. Anyone know how this works?

\$\endgroup\$
  • 1
    \$\begingroup\$ If you use a Zener as a reference the stability of your supply degenerates to that of the Zener. Use a variable regulator. If you must do this, bypass the Zener with 47uF or more. \$\endgroup\$ – user207421 Jan 26 '13 at 23:24
  • \$\begingroup\$ The emperor has no clothes on!!! The "fixed" versions of the AMS1117 are every bit as adjustable as the ajustable version. They have the same basic internals but an added internal divider. While this COULD influence the Adj (or ground) pin current, the spec sheet indicates that it hasn't. Vout = Vreg x (R2+R1)/R1 where R1 is from Vout to adj pin and R2 is from Adj_pin to ground. Read data sheet for current,s minimum load etc. Add Iadj current effect in R2 for more precision but usually irrelevant. To get 4V from 3.3V regulator.Say R1 = 270R, R2 = 0.7/3.3 x 270r = 55 Ohms. Try 270R, 56R. \$\endgroup\$ – Russell McMahon Jan 27 '13 at 0:35
5
\$\begingroup\$

Fairchild datasheets for their LM78XX series of linear regulators contain a circuit for increasing the output using a resistor to hold the reference above ground. I guess how well it works in practice will depend on how stable the quiescent current is over the operating range for the AMS1117.

enter image description here

The above image was from the Fairchild 7805 datasheet that may be found here:

http://www.fairchildsemi.com/ds/LM/LM7805.pdf

\$\endgroup\$
  • 3
    \$\begingroup\$ +1 for not just saying "that's not how you should do that!" \$\endgroup\$ – justing Jan 26 '13 at 23:46
  • \$\begingroup\$ Thanks! I thought I remembered something like this but I couldn't remember where I saw it. I used 470Ω for R1 and 100Ω for R2 which gives me around 3.9 to 4.0 volts output depending on load and supply voltage. The AMS1117 datasheet shows a quiescent current of 5-11mA but based on my results according to the formula above it must be lower, less than 1mA under my conditions. Thanks again! \$\endgroup\$ – CR. Jan 27 '13 at 13:25
2
\$\begingroup\$

The main difference between a fixed linear regulator and an adjustable one is that in the adjustable version there is a feedback pin (usually called FB) that is used to set the output voltage using a resistor network. In the fixed linear regulator, this is all internal to the regulator. Unless that pin is exposed, it is not possible for you to do what you want in any reasonable way.

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

If you want a well regulated voltage then you would do way better to simply replace your 3.3V fixed version of the AMS1117 with the adjustable version. Same package but then you add two external resistors and a capacitor. The following circuit shows how it is hooked up.

enter image description here

This was taken right from the AMS1117 data sheet where you can read information on how to apply this circuit to your specific design.

\$\endgroup\$
  • 2
    \$\begingroup\$ Thanks, yes I know about the adjustable regulators. I'm working with the fixed version though. I believe they operate in a similar manner and can be adjusted with external components in a similar but not exactly the same fashion but this is where is has been difficult to find information. \$\endgroup\$ – CR. Jan 26 '13 at 22:34
  • \$\begingroup\$ @CR The adjustable regulators of this sort in the same type of package that you have are easy to acquire for reasonable cost. Trying to use a regulated one in an adjustable manner leads to having to add external diodes or zeners which adds a significant additional error term to the output accuracy. The adjustable part by itself has its internal reference adjusted to about 1.5% and external 1% resistors would keep accuracy of the output to within a very reasonable margin. With external diodes or zeners you will be lucky to achieve better than 10% accuracy or repeatability from unit to unit. \$\endgroup\$ – Michael Karas Jan 26 '13 at 23:52

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.