Power Supply ~0.3V , >1A

[pstrag]

Hi,

For my project, I'm looking for inspiration for a simple power supply with a constant and stable (no switched-mode) voltage of about 0.3V (could be 0.4V or 0.5V as a last resort) but with a current of at least 1A - 1.5A preferred.

I know that I can build such a power supply with an LM317, but then I need a negative voltage on the "Adjustment" pin. Could I ask for suggestions on how else I can generate such a low voltage, but at the same time provide a current flow of at least 1A? (1.5A preferred)


Thanks.

[Benta]

What's your input voltage?
Is your load attached all the time?
You might consider a constant-current source and a shunt voltage regulator. But it'll burn power constantly.

[magic]

The simplest possible 0.3~0.4V shunt regulator could be a schottky diode connected to ground. Not very accurate, though; voltage will vary with load current and diode temperature. If this is not a problem then all you need is 1N5819 or a few in parallel, 0.47~1Ω resistor and an AA cell.

For better accuracy, a single-chip but not exactly cheap solution similar to LM317 is LT3080 (and other similar regulators from Linear).

Cheaper and still accurate but more work: NPN emitter follower controlled by an opamp.

[JustMeHere]

https://www.digikey.com/en/products/base-product/analog-devices-inc/505/LT3083/83628

[tooki]

https://www.digikey.com/en/products/base-product/analog-devices-inc/505/LT3083/83628

^^^^ this.

Instead of workarounds to make an ancient part do something it’s not designed to do, just use a modern part that is designed to do what you want.

[shapirus]

https://www.digikey.com/en/products/base-product/analog-devices-inc/505/LT3083/83628

That price is quite steep.

TL431 followed by a voltage divider + LM358 (LM324) set up as a voltage follower with a sufficiently powerful NPN BJT whose emitter gives negative feedback to the op amp will do the job dirt cheap, of course at the expense of using more board space. Load goes between the NPN's emitter and ground. Watch out for heat dissipation, but that's regardless of what parts are used.

The downside is that the op amp might go oscillating depending on what will represent the load.

[mariush]

I wonder if it woudn't be possible to just configure the regulator to output some voltage, put a diode or two in series and then have opamp double or triple the output voltage and send it as feedback to the regulator to account for the variation in the forward voltage of the diode(s) that would happen with current variation and temperature change.

There are LDOs with a voltage reference of 0.8v or even lower, so an opamp would only have to do a 2x..4x on the input to satisfy the feedback.

[JustMeHere]

https://www.digikey.com/en/products/base-product/analog-devices-inc/505/LT3083/83628

That price is quite steep.

There are cheaper versions (just change the last number -1) Time vs price is good enough for me.  Also, this was designed to handle the workload.  It has an adjustable current limiter. Very wide input range.  And would be comfortable delivering 1.5A.   Don't need to worry about a short circuit causing a fire.

[ArdWar]

You pick ADI/LTC parts for things that you need to just work, and work as advertised at that. ^(mostly)

I assume this is an one-off project so price shouldn't be that much of a concern. You pay for the convenience.
If you want to cost optimize for mass production then sure you can afford to thinker around with other topology and design around it. For example "amplifying" the FB sound simple but surely a recipe for oscillating disaster at certain operating point without compensations. Might as well build a discrete regulator at that point.

[bostonman]

I was also looking for a <1v (adjustable) power supply several weeks ago for a work project and came across this schematic online, however, I'm uncertain about how stable it is over higher currents (I haven't built it).

Also, maybe you can tweak the resistors to get a wider or narrower range.

[magic]

I would remove R4 from this circuit to improve accuracy.
To enable really low output voltages, short circuit R3 and replace LM308 with LM358 (or LT1006/1013 for higher precision).
The usual compensation trick could be added: a small capacitor from opamp output to IN-, bypassing R6, T1 and R5.

[shapirus]

There are cheaper versions (just change the last number -1) Time vs price is good enough for me.  Also, this was designed to handle the workload.  It has an adjustable current limiter. Very wide input range.  And would be comfortable delivering 1.5A.   Don't need to worry about a short circuit causing a fire.

Yeah it's generally true that using an IC designed for the job is an easier and more reliable (and predictable) way. Having alternative options is important, however. For example, it (a matter of personal preference) may just be less fun to use a ready made solution instead of building something from a few lower level blocks, of course we're talking DIY/hobby in this scenario. Logistics may be an issue that makes certain parts a no-go: if we take where I live for example, the LT308x chips are only available for backordering with a several weeks lead time (curiously cheaper than digikey: ~$5.5 ordered in ones -- yet this seller seems to be sourcing them from digikey), whereas jellybean parts like TL431 and LM358/324 and transistors are readily available.