Recommendation Request - Neg to Pos regulator

[rfdes]

Hi -
I am looking for a solution for generating 5Vdc 25mA from a -24Vdc source.  This seems like a good candidate for a switched capacitor DC-DC converter.
Obviously trying to keep parts count and price low.  Any recommendations?
Thanks.

[Doctorandus_P]

I don't know what part you consider "GND" in your project, but overall "GND" is just an arbitrarily named node that is used as a reference.\

For the rest, switched capacitor power supplies only work for doubling or inverting a voltage. They work under the assumption that a capacitor is charged to the input voltage, and then deliver that same voltage at their output. They are not fit for generating other voltages.

For modifying voltages, look into all the variants of SMPS circuits with an inductor. All those are based around putting current into the conductor, and then delivering that current to some other part of the circuit. And because the current though the inductor can be controlled by the circuit, and the output voltage is "arbitrary", a lot of different power supplies can be created around switching an inductor.

[rfdes]

I don't know what part you consider "GND" in your project, but overall "GND" is just an arbitrarily named node that is used as a reference.
Yes, I realize that and initially considered this.  I am designing a daughter board that will be installed into a piece of test equipment.  I was hoping to keep the same GND as the system.  It still may be the way I go, I just need to isolate the different references.


For modifying voltages, look into all the variants of SMPS circuits with an inductor. All those are based around putting current into the conductor, and then delivering that current to some other part of the circuit. And because the current though the inductor can be controlled by the circuit, and the output voltage is "arbitrary", a lot of different power supplies can be created around switching an inductor.
This is my least desired option as I'm trying to avoid magnetics. Something simpler would be desirable.

Thanks for your input.
Jim

[Manul]

You can use of the shelf isolated converter module. Lots of manufacturers make them, just search. Very simple.

Or you could use a standard boost converter referenced to -24V as ground. Then reference it's output to "real" ground. The only caveat is that you need to translate feedback reference (and that it's not affected by -24V supply variations). Can be done with current mirror or some other tricks maybe.

[ajb]

An off-the-shelf isolated DC-DC converter is a good idea.  If you really want to avoid any kind of inductor for whatever reason and are willing to accept horrible efficiency in return, you could use a linear regulator to provide -5V from the -24V supply, then use a switched capacitor boost arrangement to provide +5V from that.  A boost converter that provides +5V directly from -24V is going to be the most efficient and probably lowest cost option, but the isolated DC-DC should also be pretty cheap and would be the simplest.

[Doctorandus_P]

Just before reading about the isolated DC-DC suggestion I also thought of this and it may indeed be a good option.
5V@ 25mA is also only 0.125W.
A linear regulator for 5V would also only waste 0.5W as heat, which may also make it a viable option.

But the isolated DC-DC converters cost around EUR3. But do remember they are usually not very well regulated and noisy, and they are just a building block and need extra buffer and decoupling capacitors. For a "cleaner" output you can add a "capacitor multiplier" and/or an extra voltage regulator. Combined with proper PCB layout that should get rid of  the noise.

[PCB.Wiz]

I am looking for a solution for generating 5Vdc 25mA from a -24Vdc source.  This seems like a good candidate for a switched capacitor DC-DC converter.

Not really, switch cap work best for simply integer ratios.
For your spec, you would need 4 or 5 caps in series during charge phase and then those flip to parallel, but outside the top voltage, during discharge phase.
That's a rare combination, so nothing commercial comes like that.

Obviously trying to keep parts count and price low.  Any recommendations?

-24 to +5 is a sitter for a simple flyback converter. You just need to level shift the feedback.
A PNP matched pair transistor gives tempco matching, and 4 or 5 resistors will shift the +5V feedback to a typical 1.25V reference above the -24V

[Zero999]

Obviously trying to keep parts count and price low.  Any recommendations?

-24 to +5 is a sitter for a simple flyback converter. You just need to level shift the feedback.
A PNP matched pair transistor gives tempco matching, and 4 or 5 resistors will shift the +5V feedback to a typical 1.25V reference above the -24V

Something like this? I haven't simulated the switched mode, just the concept.

There are errors due to Q1's base current and the Early effect. Adding base resistor equal to R1 to Q2 will compensate for Q1's base current and cascoding Q1 with a small P-MOSFET will reduce the Early effect.

 voltage mirror.asc (1.72 kB - downloaded 47 times.)

I'd be tempted to use an op-amp and P-FET. The only downside is P channel JFETs aren't so common nowadays. It will also work with a MOSFET, but then there are the issues with the much higher gate capacitance and protection against exceeding the maximum V_GS rating. The J-FET solution also has the advantage of the op-amp been able to be run off 5V.

 voltage mirror op-amp.asc (1.22 kB - downloaded 61 times.)

EDIT:
Another option is to change J1 in the above schematic to a Darlington pair. That will make the error due to the base current negligible. If space is an issue, then go for a Darlington in a single package such as the MPSA64, BCV46TA, BCV26 etc.

How to apply it in the real world.