Tracking adjustable power supply with two positive rail regulators

[mcbmcb0]

Hello
i wonder if you could tell me if this is a good idea or not.
I want to make a dual tracking power supply with the usual positive and negative outputs.
There are the usual circuits with positive and negative rail regulators such as this (copied from somewhere) with the negative side following the positive side via an op amp.


BUT would it work with two positive rail regulators? The advantages would be (1) there are better LDO positive rail regulators than their negative counterparts and (2) when i add mirrored current limiting i won't have to bother about all the op amps being rail-to-rail (the final version will need 75v op amps for the tracking etc).
here's a basic diagram:


I breadboarded a test circuit an it worked OK. The inputs are from different transformer windings/rectifiers/caps etc. The op amp inputs and outputs follow down to the negative rail. the inverting and non-inverting inputs are reversed from the first schematic above.

One potential problem is the op amp's supply voltage changes according to the output voltage...
Does this - or anything else - look like a bad idea?

thanks for any suggestions
Mike

 

[dougy83]

Hi,

Can you make it clear in your schematic what the input power supply is. You have two input rails marked '-ve' - one is supposedly the -ve output while the other is the 0V/GND.

You can use two positive regulators if the inputs to each are isolated from each other in some manner.

If you're worried about 75V rail opamps, why don't you just use a transistor to that can take the voltage swing and power the opamp from a lower voltage?

Cheers,
Doug

 

[mcbmcb0]

Thanks for your comments Doug.
The dc inputs (on the left of the page) are two independent supplies - from different winding with different regulators and soothing caps. yes - they are isolated.
(sorry i'm just learning how to draw these things out on the computer, so some details vague)

re your transistor suggestion - I've got myself worried that the op amp must take the full rail to rail voltage so the inputs are always within the vcc - vss. As if if took a big spike on one rail then would not the rail splitter spike well away from ground, until it regains tracking? Or maybe i'm not understanding your suggestion

and... any concern about the op amp's supply voltage varying?

as you can tell, my knowledge is a bit limited here - all guidance appreciated

 

[MikeMl]

Look at your chosen opamp's Power Supply Rejection Ratio (PSRR) spec.

 

[dougy83]

re your transistor suggestion - I've got myself worried that the op amp must take the full rail to rail voltage so the inputs are always within the vcc - vss. As if if took a big spike on one rail then would not the rail splitter spike well away from ground, until it regains tracking? Or maybe i'm not understanding your suggestion

the voltages on both the inputs of the opamp will be very close to 0V always; if there's a spike on one it should be safe, but you can connect some diodes on the opamp input if you wanted.

and... any concern about the op amp's supply voltage varying?

As per MikeMI, the PSRR tells you how much of the noise on the power supply will make it to the input of the opamp. (as per 3rd paragraph https://en.wikipedia.org/wiki/Power_supply_rejection_ratio)

 

[mcbmcb0]

OK thanks. so that's what PSRR means!
for LT1013 its 120dB. I'm *guessing* the gain is 1, but either way figure i'm talking a few millivolts (even across a 30V change in supply voltage), which i can live with.

I'll check out diode protection across the op amp inputs, and experiment with the op amp supply voltage connecting to the tracking regulator's input supply.

sounds like i'm good to go... thanks for the help

 

[crutschow]

I believe you could power the op amp from the unregulated minus supply to avoid the change in op amp voltage with output voltage.

One possible problem is instability of the op amp due to the high open-loop gain of the op amp. You may need to add compensation, such as a capacitor from the op amp output to the minus (-) input to prevent oscillations.