1
\$\begingroup\$

There are several out of the shelf solutions: LR8 from Microchip (formaly known as Supertex product) and NCP785 from ON-Semi. The last one is able to handle up to 700 volts at the input which is impressive.

Both components is very attractive to be used in my application. However, LR8 does not have enough voltage strength (I need to withstand 400 VAC = 625 VDC) and NCP785 is not easy to buy in my country (Russia).

One of the solutions is to put an external transistor like this:

enter image description here

However I think that it should be possible to add an external transistor to linear regulator (LP2981 in my case) without additional components (zener, resistors...) - this would be perfect as I need to reduce PCB size.

From my early engineering career I remember that the external transistor can be used to increase the current of the regulator, something like this:

enter image description here

Is it possible to do same thing to input voltage?

\$\endgroup\$
  • 1
    \$\begingroup\$ If R1 is really 1ohm I expect your regulator to explode quite fast. \$\endgroup\$ – Vladimir Cravero Oct 17 '16 at 10:35
  • \$\begingroup\$ @VladimirCravero this is not my schematic. Just took one from Internet. Actually I don;t really tried to understand how it should work. \$\endgroup\$ – Roman Matveev Oct 17 '16 at 10:37
  • \$\begingroup\$ How much current does your regulator need to supply to the load? \$\endgroup\$ – Andy aka Oct 17 '16 at 12:41
  • \$\begingroup\$ @Andyaka just several hundred micro Amps. MOSFET in my schematic heating for just 10-15 degrees. \$\endgroup\$ – Roman Matveev Oct 17 '16 at 13:23
1
\$\begingroup\$

As far as I can see you need to rectify the incomming phase (diode + capacitor). Feeding from an alternating source is not possible.

The basic model for increasing the current of a regulator should be a little different. Here the corresponding image with the principle.

enter image description here

Realizing regulator with high voltage input is diffent but also feasable. I did not find a regulator but you can build one yourself as shown in the schematic. Ofcourse you can modify the parts etc. according to your needs and parts available.

enter image description here

\$\endgroup\$
  • \$\begingroup\$ So your approach is to skip the regulator at all. This is quite interesting! I will go deep into your schematic! \$\endgroup\$ – Roman Matveev Oct 17 '16 at 10:52
  • \$\begingroup\$ you can also use a normal regulator, if it employs an N device. You power the regulator via zener + resistor, and use an external pass device. Keep in mind that the pass transistor dissipates a lot. You have approximately 250V across it, if you only need 100mA that's 25W and it is a lot. A dcdc is probably the way to go unless you need currents in the order of 1mA. \$\endgroup\$ – Vladimir Cravero Oct 17 '16 at 11:37
1
\$\begingroup\$

I need to withstand 400 VAC = 625 VDC

A common method is to use a voltage dropping capacitor feeding a zener diode. If you have an AC input then you will be using some form of rectifier and the zener can be placed after the rectifier. For a 1 mA load, it's probably advisable to feed 2 mA into the zener and from this you can calculate the value of the capacitor that drops the voltage. This is the general idea: -

enter image description here

So, you get regulation (as good as a zener can provide) and very little heat produced. I estimate that C1 could have a reactance as high as 200 kohms and at 50 Hz, this means about 15 nF. R1 won't need to be a 5 watt type because your current consumption is so low (< 1mA).

PLEASE BE AWARE THAT ANY METHOD OF DERIVING A DC POWER SUPPLY FROM THE AC MAINS WITHOUT USING AN ISOLATION TRANSFORMER IS DANGEROUS.

\$\endgroup\$
  • \$\begingroup\$ Andy, is it still possible to measure AC voltage? One of the reason to use those resistive voltage drop (I'm currently on) is to have a possibility to measure rectified AC voltage by means of a simple resistor devider. I tried several approaches in LTSpice to combine capacitor based power supply and ADC with two resistors: no result :( \$\endgroup\$ – Roman Matveev Oct 17 '16 at 14:51
  • \$\begingroup\$ Using a half wave version you keep neutral as common for power and signal and this makes it possible but you'll have to show your circuits if you want me to comment on them. \$\endgroup\$ – Andy aka Oct 17 '16 at 15:06
  • \$\begingroup\$ img-fotki.yandex.ru/get/54522/855366.8e/… my comments: parts on the bottom right and center left is another story, BSS126 is a current protection (another story as well). The AC measuring going through R17,18 (center top), than filtered at R3,C1 and R8,C2 (bottom left). \$\endgroup\$ – Roman Matveev Oct 17 '16 at 15:15

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.