linear supercapacitor charge circuit protection?

[coppercone2]

So if I have a linear regulator, I was thinking about this one, https://www.analog.com/media/en/technical-documentation/data-sheets/158125fa.pdf

to charge the super cap. On the bottom pages the document goes into detail about 5000uF being their considered maximum. A supercap is way more then that.

If I could tolerate presence of a current limiting resistor and beefed up those diodes, it would drain current through the protection diodes at the maximum set by the current limiting resistor, right? On Page 9 they show diodes to vcontrol and vin.

I.e. for power loss with charged capacitors.

[jonpaul]

you need a current source not voltage regulator to charge the supercap.

I = C vd/dt

many simple current regulators available both linear and switching

jon

[coppercone2]

I suppose a current source would be better then a resistor limited voltage source, but I am mainly worried about what happens when you unplug it with super capacitors attached. Usually the capacitance of circuits are small on input and output and usually have bigger input then output so power off behavior is never really considered.

if you have a linear 10 amp part that is constant current with low heat sinking requirements and compatible with voltage on output when turned off that would be interesting. I have 5V supplies (go up to 6) to charge these with. Last time I did LT3080 constant current source for some LED, it turned out the input voltage requirement was pretty high.

[magic]

The problem is not with the input unplugged, but with the input shorted. Then the capacitor discharges through the regulator.

ESR limits the discharge current in such case so protection diodes need to be dimensioned for that, or further resistance added.

Alternatively, a diode (Schottky?) could be put in series with the input if there is enough voltage headroom.

Some LDOs with PNP outputs can withstand reverse voltage. MIC29752 is one example, but not quite 10A.

[coppercone2]

Hmm. Well these use switching meanwell power supplies to power the linear regulators. If I switch those off, its not going to do anything remotely similar to shorting the input will it? I wanted a mains switch only, no interlock between the meanwells smpsu and the linears besides a fuse, then to have a fuse and small resistor between the linear regulators and the super capacitor*.

Yeah and my idea was to test this with a ammeter with a current limited supply first to see what exactly is going on

The supercapacitor I have has a ESR of 0.5 Miliohms (0.0005 Ohms). I was going to put a series resistor between the power supply and capacitor to limit it.

*Not exactly sure about this one. It will increase charge time. I don't really want it but it seems like a good idea rather then dead shorting the power supplies.


Also I was wondering what protection if any I should put to protect the linear regulators from the super capacitors that are powering a spot welder. Like diode or choke. I would prefer not to have a series diode.

On the LT1581, internal resistors limit internal current
paths on the ADJUST pin. Therefore even with bypass
capacitors on the ADJUST pin, no protection diode is
needed to ensure device safety under short-circuit con-
ditions. The ADJUST pin can be driven on a transient
basis ±7V with respect to the output without any device
degradation.

I assume that means it might be OK to just run this thing short circuit, so long heat is managed?

[DavidAlfa]

Linear regulator? It'll waste a lot of power in heat, better get a buck current regulator with adjustable max Vout.

[coppercone2]

power efficiency is not my concern for something used so sparsely.

[coppercone2]

I think it might be wise to install relays that open when the power is turned off so that the capacitors are isolated from the electronics

[DavidAlfa]

power efficiency is not my concern for something used so sparsely.

It's not just the efficiency.
If you want to charge the supercap quickly enough, instead hours, you'll need several amps.
With a buck regulator it's easy, compact and keeps everything relatively cool.
Try that with a linear, you'll have a bulky thing requiring heatsink, perhabs a fan...easily getting more expensive than the buck converter.
Plus it'll heat everything up, may affect other components, not sure about the heat effects on supercaps lifetime.
You say the power is not a problem, but it's always nice to be able to reduce the power supply requirements and save some bucks.