Symmetric floating bench power supply tracking problem

[Onemangang]

Hi all.

I am in the process of designing my own fully symmetric bench power supply. It will go from 0 to around ± 70V @ 3 amps.
The voltage controls will be floating, while the current limiters probably will be in respect to true gnd of the main supply.

The whole thing will consist of 4 internal supplies to make it all work. 1x 75-0-75 and 3x 15-0-15 DC. One of the 15-0-15 will share gnd with the
75-0-75 main supply. The two others will be floating, which is where I am a bit lost.

I cant wrap my head around to to make the tracking work between the two floating supplies. The only thing I can think of is to
create some variable current sources in the common 15-0-15 supply, and pull that current through a resistor in the floating
supplies. That should work, but it somehow doesnt feel right.

I dont need mcu control or programmability. I just want to turn one reference voltage from true gnd into the same voltage on
each floating supply. I have thought about using a stereo pot for this, but the channels usually dont track too well on pots.

I also havent found any info on this, nor could I find any schematics on this issue.
Any insight on this is welcome

Thanks in advance
Jørgen

 

[jan.didden]

How about putting two equal resistors in series between the +70 and -70. If they are the same voltage, the midpoint will be at 0V. You can then use that midpoint to correct one supply to keep the midpoint at 0V and thus ensure tracking of the two supplies.

Jan

 

[MarcelvdG]

Just to get the question clear: you want to make three adjustable symmetric supplies that can be adjusted up to +15 V/0/-15 V, two of which float, and they all need to be controlled by one and only one knob? So essentially six supplies, three positive and three negative, tracking each other and controlled by that one knob?

On top of that, there will be an adjustable symmetric supply up to +70 V/0/-70 V, but that doesn't need to track its colleagues.

You can use current sources to take a trimmable reference voltage from one domain to another, but only as long as you don't get into voltage compliance issues. Are the floating supplies to float completely or is there a limit to how much voltage difference there can be between the 0 of one supply and the next?

 

[Elvee]

Even for the maximalist case, there are solutions: isolation amplifiers, matched dual optocouplers, V/F+F/V converters, digital transmission through opto's, inductors or even optical fibers and radio links.
The fact that it is possible doesn't mean it is desirable: the requirements should be kept to the bare minimum possible (AKA the KISS principle)

 

[Onemangang]

Hi again.

I implemented my first idea using currentsources in ltspice, and it seems to track pretty much perfectly.
Schematics are now attached to post #1. Would there be a better way to to it ?

@jan.didden
I would like to not reference the output from one to input of the other.

@MarcelvdG
I think i didnt express what i meant very well. I refrased it.

@Elvee
Thank you for the suggestions. I also ended up doing something "Kissy"

What do you think about the reference system? Most of the relevant stuff is inside the rectangle on the schematic.

The schematic is not complete. Current limiting is not yet completed, but I though about using highside sensing
and tuning it for 1Volt = 1Amp. This way I can use a voltmeter to readout my current limit setting.

Off topic:
Please dont mind the transistors used in the output. It will use darlingtons that come from a big surround receiver,
which is also why the supply is made like this