[time-nuts] Two V(in)s walk into a bar...

[Chris Caudle]

On Wed, June 24, 2015 4:54 pm, Neil Schroeder wrote:

Better late (reply) than never?

> Main question: if I want to isolate the quiet and noisy power
> supplies, or just supply them differently, can they share a
> common return?

All conductors have resistance and inductance.  Any current flowing on a
shared return will be seen by both supplies, so be aware of when that
return conductor is also used as a reference (meaning that if the voltage
on that conductor increases or decreases, the output of the supply you are
worried about increases or decreases by the same amount at the same time).
That will often be the case, but there could be configurations where there
is a reference separate from the return line.

> The basic answer has to be yes.  Eventually everything has to
> get  back to ground, and the Earth itself is a fair equalizer of
> all things on it.

I see two misconceptions in a single sentence.

> Eventually everything has to get back to ground

Current has to flow in loops.  That is the only requirement based on the
physics of electricity.  When you say "ground" that term is ambiguous.
Obviously not everything has to get back to "the ground," i.e. earth
reference, because car electronics run while suspended on insulating
tires, and aircraft electronics run at several km above the "ground." So
maybe you mean the reference conductor, but if that is the case then  you
really should say eventually all current has to get back to the reference
conductor of the source supply.  It is a cumbersome mouth full, but you
really have to get that concept in your head, especially when thinking
about transformer isolated supplies where current just has to complete a
loop with the transformer secondary, but you may need to think about
leakage current through the interwinding capacitance and understand the
path the the leakage current takes back to the primary side.

> Earth itself is a fair equalizer of all things on it

Typical soil resistance varies between 10 and 1000 Ohm-meters (according
to Wiki the All Knowing:  https://en.wikipedia.org/wiki/Soil_resistivity
).
Copper resistivity is 17 nanoOhm-meters.  0.000000017 Ohm-meters. Nine to
eleven orders of magnitude worse than a "good" equalizer does not come
anywhere close to fair in my terminology (assuming by equalizer you mean
the ability to hold two points at equal potential when there is current
flowing).  I would class earth as the poor equalizer of last resort.

> Take the PRS10 - it has two cables for Vin - power and signal
> supplies  but only one return cable.

I think that must just be so that you don't have to use a regulator which
can hold regulation over the entire current range from electronics plus
warm heater (low current) to electronics plus cold heater (high current). 
You just need a regulator which can hold regulation at the current draw of
the electronics, and another supply which doesn't have to be as tightly
regulated for the lamp and heater.

But yes, the heater current flowing down the shared return can modulate
the output of the "clean" power supply, because to the electronics in the
oscillator module there is no difference between the 24V dipping low, or
the 24V staying constant referred to some external reference, and the
voltage of the return conductor increasing referred to the same external
reference.  The only thing the electronics can see is the difference
between in and out.

> Can i isolate, via transformers couplers what have you the two
> supplies yet return the common ground to one or the other?

Sure, just keep in mind that with a shared return any current flowing from
either supply causes a voltage drop on that return conductor which is seen
by both supplies.

So lowering the resistance of the return conductor can help, as well as
lowering the inductance of the return conductor.  In cases where you have
separate supplies, if there is any high frequency demand on the supply it
may help to have separate twisted pairs of supply and return for each
supply, with the returns joined at the common pin on the oscillator.  The
two wires will halve the total resistance, and the twisted pairs will
reduce any noise coupling in or out of the supply lines.

>  I think the single return option on the PRS10 is what's hanging me up
> here.

Don't get so twisted up, it is just so the heater current doesn't cause a
lot of ripple on the supply to the rest of the electronics.  It might be
better if it had a separate return line, just keep the return line as low
resistance as practical and don't worry about it, the guys at SRS did the
worrying for you.

The manual seems to indicate that in most applications you can just run
them from the same supply:
"In most applications, both +24 Vdc supplies (heater and electronic
supplies) will be connected together and operated from a +24 Vdc supply"

The internal electronics will be running at some lower voltage, so there
has to be additional regulation in the module, so it makes sense that it
wouldn't be too critical.

-- 
Chris Caudle