Assuming the source voltage Vs is regulated in below diagrams, is there any advantage of using this rail-splitter topology:

enter image description here

instead of this simple one?:

enter image description here

Is there any good reason or scenario which makes the first topology superior to the second one if the power supply Vs is a regulated supply?

  • \$\begingroup\$ For both cases think what happens if you place a resistor of 1 kohm between V+ and VGND. Placing the resistor between V+ and VGND means that the supply is loaded asymmetrically. \$\endgroup\$ – Bimpelrekkie Oct 11 '17 at 13:04
  • \$\begingroup\$ You emphasize that Vs is regulated, but does that actually matter for splitting the supply? The function of the splitter is to create a virtual ground point. That can still be done when vs varies. \$\endgroup\$ – Bimpelrekkie Oct 11 '17 at 13:07
  • \$\begingroup\$ @Bimpelrekkie I plan to use one of these above to power a dual supply opamp's rails. Or to power a Wheatstone bridge. I dont think in my case there will be balance issue (?) I was wondering if the second one is sufficient or the first one is better. \$\endgroup\$ – user1245 Oct 11 '17 at 13:22
  • \$\begingroup\$ I dont think in my case there will be balance issue If an opamp's output has a voltage other than zero and this causes a current to flow to ground you will have an imbalance. \$\endgroup\$ – Bimpelrekkie Oct 11 '17 at 13:28
  • \$\begingroup\$ But then this splitter cannot be used for powering opamps or Wheatstone excitation? I mean in what application for example it would be okay to use the first topology? Is it totally useless? \$\endgroup\$ – user1245 Oct 11 '17 at 13:32

The first one can sink more current through the "ground."

Its impedance is set by the 10 ohm resistors and the transistors. Fairly low impedance.

In the second diagram, the current is set by the resistors and the voltage rails. 4.7k is going to limit the current drastically in comparison to the first circuit.

  • \$\begingroup\$ I dont get how this transistor circuit work. But if the load is between V+ and V-, I see no imbalance in simulation. But if there is a load is between lets say V+ and VGND, then the V+ and V- changes with that load... I want to use that for this sensor:crlsensors.com/prodDocs/sa-107ln.pdf which requires +-15V excitation and it has built-in amplifier. I want to power four of these sensors by using the transistor splitter(+-12V) from a 24V regulated supply. Since I dont know the inners of the sensors Im a bit confused if it would be alright to use this topology for this application. \$\endgroup\$ – user1245 Oct 11 '17 at 18:53
  • \$\begingroup\$ Those sensors draw 10mA each. Four in parallel would be 40mA. At 24V that's a load of 600 ohms. Or 1200 if you only consider the halves. Your second circuit has no chance at all. Put a 1200 ohm resistor in parallel with one of the 4.7k resistors and consider what will happen to the voltage divider. \$\endgroup\$ – JRE Oct 11 '17 at 19:19
  • \$\begingroup\$ The first would probably work. You confuse me, though, with your comment about V+ and V- changing when you put the load between V+ and Vgnd. The voltage from V+ to V- should stay the same (unless your powersupply is utter crap) when put a load from V+ to Vgnd. From V+ to Vgnd might change under load, if the load is too much (resistance too low, draws too much current.) What are you using for a load in your simulation? Probably ought to be like 600 ohms. \$\endgroup\$ – JRE Oct 11 '17 at 19:25
  • \$\begingroup\$ Sorry I meant the "voltage difference" between V+ and VGND changing changing when I put the load between V+ and Vgnd. The voltage from V+ to V- is well regulated. What I mean and confused is the "voltage difference" between V+ and VGND is not regulated. Does that matter for this application? I mean I dont know if there can be very different loads between between "V+ and VGND" and between "V- and VGND". Is there a way to estimate the imbalance for this application if I use the first transistor circuit? \$\endgroup\$ – user1245 Oct 11 '17 at 19:31
  • \$\begingroup\$ The voltage shouldn't change much. What are you using for a load in your simulation? \$\endgroup\$ – JRE Oct 11 '17 at 19:37

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.