Shop OBEX P1 Docs P2 Docs Learn Events
Can I use Zener diode to boost up LDO Voltage ? — Parallax Forums

Can I use Zener diode to boost up LDO Voltage ?

john_sjohn_s Posts: 369
edited 2013-12-21 11:28 in General Discussion
Let's say I have a 1.5V LDO and would like to have it working as a 5V regulator - can I simply place a 3.5V Zener diode in its GND leg?

I've never tried it but according to some app notes it's doable with 78XX series etc.
Does the same method apply to modern LDO voltage regulators?

What are the negative aspects of such approach?

Vreg.PNG
Vreg.PNG
417 x 315 - 29K

Comments

  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2013-12-20 10:08
    Hi john;
    john_s wrote: »
    Let's say I have a 1.5V LDO and would like to have it working as a 5V regulator - can I simply place a 3.5V Zener diode in its GND leg?

    I've never tried it but according to some app notes it's doable with 78XX series etc.
    Does the same method apply to modern LDO voltage regulators?

    What are the negative aspects of such approach?

    Vreg.PNG
    Yes you can.
    However, you might want to add a small capacitor, 0.1uF to 1uF, across the zener to improve the transient response.

    Duane J
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2013-12-20 10:12
    I would say just try it and see if it works. A possible source of trouble is that LDO's are pretty fussy about the capacitors used with them. The zener may well affect the output cap's effective ESR between the output lead and the sense/ground lead. As a consequence, you might need to add capacitance between the output and sense/ground lead of the regulator. But I just don't know for sure without trying it.

    -Phil
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2013-12-20 11:51
    Also be aware that it will not be a good 5V regulator. The zener is outside the feedback loop. Zeners have a soft knee and your 3.5V zener may not show its rated voltage until there is 10mA flowing through it. Low-voltage zeners are particularly bad in that respect. If it is 3.5V at 10mA, figure 2.9V at 1mA and 2.1V at 0.1mA. A 7800 series will have a fairly constant quiescent current (~3mA) out of its ground leg. However, most low-dropout regulators have a ground leg current proportional to the output current, so your voltage would end up swinging with the output (with possible ramifications for stability per Phil's comment). You can increase the zener current and stabilize the operating point (at the expense of wasted current), by connecting a resistor from the chip output to its ground leg.

    You might be better off with a blue LED in the ground leg. At those voltages, LEDs have sharper knees than zeners.
  • john_sjohn_s Posts: 369
    edited 2013-12-20 12:08
    Tracy Allen wrote: »
    ...
    However, most low-dropout regulators have a ground leg current proportional to the output current, so your voltage would end up swinging with the output (with possible ramifications for stability per Phil's comment). ...
    You might be better off with a blue LED in the ground leg. At those voltages, LEDs have sharper knees than zeners.

    Thank you all - I appreciate your comments especially in regards to swinging LDO ground leg and stability...
    I sensed that there's more behind it then just adding a Zener :)
  • ercoerco Posts: 20,256
    edited 2013-12-20 13:15
    Great Q & As. What an amazing resource and treasure this forum is. So much to learn! The collective expert knowledge base here covers hardware, software, mechanics, you name it, with informed answers delivered swiftly and politely. It is a joy and pleasure to be associated (albeit on a rudimentary level) with all of you whiz kids here. Merry Christmas and Happy New Year to all.
  • NWCCTVNWCCTV Posts: 3,629
    edited 2013-12-20 18:24
    erco, you been in the egg nog a tad early??
  • ErlendErlend Posts: 612
    edited 2013-12-21 01:50
    I've done the Zener trick many times - with good results. Some LDO's like it better if there is a -say 10k - resistor to the zener to ensure some current is flowing.

    Erlend
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2013-12-21 06:11
    Hi Tracy
    Tracy Allen wrote: »
    Also be aware that it will not be a good 5V regulator. The zener is outside the feedback loop. Zeners have a soft knee and your 3.5V zener may not show its rated voltage until there is 10mA flowing through it. Low-voltage zeners are particularly bad in that respect. If it is 3.5V at 10mA, figure 2.9V at 1mA and 2.1V at 0.1mA. A 7800 series will have a fairly constant quiescent current (~3mA) out of its ground leg. However, most low-dropout regulators have a ground leg current proportional to the output current, so your voltage would end up swinging with the output (with possible ramifications for stability per Phil's comment). You can increase the zener current and stabilize the operating point (at the expense of wasted current), by connecting a resistor from the chip output to its ground leg.

    You might be better off with a blue LED in the ground leg. At those voltages, LEDs have sharper knees than zeners.
    Yes, low voltage zeners are not very "stiff", i.e. the zener voltage can vary quite a bit with varying currents.
    However, the zener voltage can be quite stable when used in this application as the current through the zener is essentially regulated.

    1. The quiescent current coming out of the ground pin of most regulators is fairly stable.
    Quiescent current for the LM340-x1 or LM78xx series is 6mA max and a change of current of about 0.5mA.
    2. The current through R1 is highly regulated by the constant voltage of the regulator itself.

    Common low wattage zeners such as the 1N52xxB operates best at a zener current of about 20mA.

    Calculate R1 for a 10V output:
    LM7805 5 volt regulator
    IQ = 5mA quiescent current @25C (as in the chart)
    1N5231B 5.1V .5W zener
    IZT = 20mA test current
    R1= 5Vreg / (20mAIZT - 5mAIQ) = 333Ω
    330Ω would be the closest standard value.
    5Vreg2 / 330Ω = .075W of power dissipated in the resistor.
    5.1Vreg * 20mAIZT = 0.1W of power dissipated in the zener.

    Duane J
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2013-12-21 08:37
    Duane C. Johnson wrote:
    However, the zener voltage can be quite stable when used in this application as the current through the zener is essentially regulated.

    1. The quiescent current coming out of the ground pin of most regulators is fairly stable.

    Tracy's point, however, is that this is not the case with LDOs, where the quiescent current is roughly proportional to the load current.

    -Phil
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2013-12-21 11:28
    I've used that trick myself in situations not needing or expecting an accurate voltage at the output. The regulator's current limiting and thermal protection are still valuable and operational.

    It can work very well indeed if a synthesized zener such as the LM385 is used in place of a regular zener. The LM385 has a sharp knee and constant voltage for currents from 20µA to 20mA. If you don't happen to have LM385s lying around, though, it's better to buy the voltage you need.

    I was referring to LDO regulators that employ a collector-loaded (PNP) pass transistor. Here is a diagram of ground pin current vs load current, the graph on the left. The graph on the right is ground pin current vs input voltage at several different load currents. This is for an LT1963-1.5, which happens to be both low dropout (LDO) and also low power. The low power refers to the point that the ground pin current drops to about 1mA when the output current is zero.
    Screen shot 2013-12-21 at 11.05.13 AM.png
    Screen shot 2013-12-21 at 11.06.32 AM.png

    In contrast, here is the ground pin current of an LM2940, which is LDO but definitely not low power:
    Screen shot 2013-12-21 at 10.56.28 AM.png

    Even at no load, it sucks over 10mA of quiescent current, and that doubles as it goes toward full load at 1A.

    A regulator such as the MCP1703 has a p-mosfet output, and the drive requirements are considerably less, still proportional, but measured in microamps. This qualifies as micro-power, because the ground pin current at zero load is a mere 3µA.
    Screen shot 2013-12-21 at 11.11.50 AM.png


    A non-LDO, emitter -loaded regulator like the one in the 78xx series does not have much change in ground pin current as the load varies. The current to drive the output stage flows into the load instead of being lost out the ground pin.
    Screen shot 2013-12-21 at 11.27.20 AM.png


    @ Erco & all, happy holidays indeed. It is truly a pleasure to read here and to participate in this community.
    Screen shot 2013-12-21 at 10.41.17 AM.png
    638 x 354 - 48K
    Screen shot 2013-12-21 at 10.56.28 AM.png
    314 x 300 - 17K
    Screen shot 2013-12-21 at 11.05.13 AM.png
    335 x 322 - 27K
    Screen shot 2013-12-21 at 11.06.32 AM.png
    309 x 313 - 31K
    Screen shot 2013-12-21 at 11.11.50 AM.png
    327 x 259 - 22K
    Screen shot 2013-12-21 at 11.27.20 AM.png
    790 x 78 - 16K
Sign In or Register to comment.