Page 1 of 1

Split Power Supply From Unipolar Using a Rail Splitter

Posted: Fri Aug 08, 2008 8:38 am
by mediatechnology
I just ran across this power supply in the dbx163 schematic sent to me by Ian. It provides +/-12 using a 24V regulator and op amp-based rail splitter. The 24V regulator appears to be an LM723H in TO-99. The pin outs do not appear to be correct for the DIP package.

A rail splitter is used to derive ground. This technique is useful when wall-wart transformers are required. I remember NY Dave looking for one. While there are a lot of circuits similar to this one posted on just about every DIY forum, this one serves as a typical example. Note that the rail splitter appears to be running open loop until you realize that its closed by the ground symbol - a ground that it creates.

Image

Re: Split Power Supply From Unipolar Using a Rail Splitter

Posted: Sat Aug 09, 2008 10:43 am
by JR.
I recall using the 723 in my first kit back in the '70s before three terminal regulators became ubiquitous.

I'm not sure I see the rationale for using the rail splitter in this design, since a real ground exists at the common point between the two capacitors. If they are trying to deal with one rail being loaded heavier than the other, their circuit is just loading down the lightly loaded rail, driving the synthesized ground to be similar to their actual PS common. About the only thing they are accomplishing is preventing a lightly loaded rail from climbing too high.

I used a rail splitter on a non center-tapped wall wart application, because I needed to full wave rectify the transformer output to get higher current at lower voltage. I likewise balanced the load on two equal rails by shunting the lighter one, but I didn't have a ground I could grab, like the dbx circuit. I needed a high current 5V rail for digital stuff, and +/- 8V or so for audio. I was switching the audio path with CMOS transfer gates so +/- 8v was a CMOS friendly working voltage.

Note: In my case where I applied a rail split to FW rectified winding you can no longer parallel multiple (different) products from one heavier transformer winding. The dbx circuit is still effectively relative to ground although indirectly.

JR

Re: Split Power Supply From Unipolar Using a Rail Splitter

Posted: Tue Aug 12, 2008 3:10 pm
by mediatechnology
I think a lot of this had to do with them stocking LM723s. I was recently told that the engineers couldn't use 5534s either because they were, at the time, sole-sourced. In the later 163X they use three terminal regulators. There may have been some early corporate "bias" against the 7815/7915.

It's interesting that virtually every open-frame linear supply uses the 723.

I posted this because the subject of rail splitters comes up often when there isn't a center-tap. If you Google "rail splitter" a lot of them lead to HeadWize and the battery-powered headphone amp dilemma.

BTW John there's that trick you did in the TS-1 power supply...

Re: Split Power Supply From Unipolar Using a Rail Splitter

Posted: Tue Aug 12, 2008 7:24 pm
by JR.
mediatechnology wrote:
BTW John there's that trick you did in the TS-1 power supply...
?? JR

Re: Split Power Supply From Unipolar Using a Rail Splitter

Posted: Tue Aug 12, 2008 8:38 pm
by mediatechnology
Didn't you bootstrap the 5V regulator?

Re: Split Power Supply From Unipolar Using a Rail Splitter

Posted: Tue Aug 12, 2008 9:46 pm
by JR.
The trick in the TS-1 was a cross between synchronous rectification and a light dimmer. I used a switch (bipolar transistor with hysterisis) in series with the transformer winding to effectively disconnect it when I had enough voltage. This allowed me to pull relatively high current for 5V digital circuitry from the same winding as my +/- 18v analog supply without having to dissipate 15v across the regulator, not that the transformer would have been happy putting out all that power.

JR