Current limiter for my power supply

[Windows9Pro]

Recently, I bought a lab power supply, wich has no short circuit protection. Is there any component or circuit wich limits the current without any voltage losses? I know, I could use a Resistor, but they cannot output for example 2.9 amps at 1 Volt.

My power supply supports 3 amps and 15 Volts (15.8)

I wish I could just connect the circuit to the annode of my power supply, so I can make an Adapter with banana plugs out of it.

The limiter shoult have a max output of 2.9A

 

[dknguyen]

Not without losses. The closest thing is a PTC which is just a lowish resistor whose resistance exponentially as it gets hotter. Maybe just put in a convenient accessible fast-blow (or slow blow) fuse.

 

[Windows9Pro]

I think I will just buy an 5Ohm Resistor... But I'm still looking for a simple circuit.

 

[MikeMl]

The only way is to add a current sensor which monitors and reacts to the voltage across a shunt resistor which is inside the control loop of the existing power supply. You would have to get intimate with the innards of the power supply. Is the supply linear or switch-mode? Do you have a schematic of it?

 

[unclejed613]

most active current limiters use foldback limiting, which means you will get full voltage up to the current limit, and above the limit, the supply goes into constant current mode, and the voltage begins to drop linearly with the resistance of the load. with a resistor, the load can continue to draw more current as the resistance decreases, and the voltage will also drop.

 

[crutschow]

Deleted

 

[rjenkinsgb]

Remember that a resistor will drop more and more voltage as the load current increases; with the 5 ohm example you mention, you will lose 5V at one amp load or 10V at 2A load etc.

It defeats the object of having a regulated power supply that can deliver a precise voltage.

Not without losses. The closest thing is a PTC which is just a lowish resistor whose resistance exponentially as it gets hotter. Maybe just put in a convenient accessible fast-blow (or slow blow) fuse.

That's the best, simple, option - a "polyswitch" type solid state fuse.

Examples: http://uk.rs-online.com/web/c/fuses...ses/?applied-dimensions=4294584304,4294584339

The next best option for near "zero loss" would be electromechanical. A big glass reed switch typically operates at around 100 ampere-turns so eg. start with around 35 turns of suitable wire, with plenty spare for adjustments.
eg. http://docs-emea.rs-online.com/webdocs/1484/0900766b814841ac.pdf

This shows one with an added winding for use as a current sensor, mounted on some perfboard; you need more turns, possibly as two layers, for it to operate at lower current:
http://www.aeroelectric.com/Pictures/Misc/CurrentSense.jpg

You should be able to tweak the turns so the reed switch operates at exactly the current you need; add turns if the operate current is too high, remove turns if too low.

Use the contacts to operate a small relay that latches itself in and also disconnects the PSU output to the load via a second normally-closed contact.

That will give an instant disconnect on overload, reset by turning off or disconnecting the power supply.
If you build it as a separate unit to connect inline with the PSU, just add an on-off switch at the input connection. That will also be your "reset".

(We've used the same reed-switch-with-a-few-turns-of-wire current sensing system as an extra safety sensor in some machine tools as an additional interlock for a magnetic chuck, so the machine cannot start unless the chuck is actually drawing current. They are simple and reliable with near zero loss)

 

[Windows9Pro]

@MikeMl That's my power supply: https://cdn.sos.sk/productdata/36/dc/848c3a75/bt-153.pdf
It has a current limit function, but when i short it, the electronics get stupidly hot (the manual says it). And I don't want to kill the supply.

And yes, it's linear!

 

[Windows9Pro]

@rjenkinsgb what exactly does mean trip current ?

 

[MikeMl]

@MikeMl That's my power supply: https://cdn.sos.sk/productdata/36/dc/848c3a75/bt-153.pdf
It has a current limit function, but when i short it, the electronics get stupidly hot (the manual says it). And I don't want to kill the supply.

And yes, it's linear!

Are you sure it is getting too hot? A temperature rise of 40degC is normal for electronics, but it seems "stupidly hot" to the human hand... (actually, the human mind).

 

[unclejed613]

a lab supply like that should be able to handle being in constant current mode for any foreseeable amount of time. yes it may get hot, and it might help to have an external cooling fan for it, but that's what the current limiter is for.

if you operate the unit at full voltage and current into a short, yes it can overheat the supply, but for most applications that would take the supply into current limiting, it's not something to worry about. for instance, if i set the unit to 5V output, and set the current limit at 1A, and the circuit i'm running from the supply makes the supply go into constant current mode, the current is limited to 1A, and it's not going to generate as much heat inside the supply as it would if the current limit was set wide open. when i use a lab supply to do amplifier troubleshooting, i set the current limit at 100mA. if something is shorted in the amp, the supply goes into constant current mode at 100mA, and the short in the circuit under test doesn't cascade to other components. with the current set at 100mA, the lab supply won't be as likely to overheat, because the current is limited.

if the voltage and current were set to wide open, and a short is placed across the output, you are generating 45 watts of heat in the supply (actually probably higher, around 60 watts assuming the preregulated voltage around 20V). if the current limit is set to something a bit more sane, like 1A, a short across the terminals only generates 15-20W inside the supply.

 

[Windows9Pro]

That's what I thought, too @unclejed613 and MikeMl, but at this Situation, I think the manual wins.

Manual: Caution! The lab power supply has a protective circuit that limits the current in the even of a short circuit. To prevent the lab power supply from overheating, always turn it off immediately and disconnect all connected units in the event of a short circuit. Allow the lab power supply to cool down and make sure that it is well ventilated. The air vents (12) must be clean and free of dust, do not cover the housing openings. The lab power supply‘s maximum continuous operating time is 8 hours. After this period, turn off the device and let it cool down to room temperature.

 

[unclejed613]

if you set an appropriate current limit for the circuit you are testing, it will generate a lot less heat inside the lab supply should there be a short. for instance, a circuit using a few op amps, that operates normally with less than 100mA of current. set the current limit to 100mA, and if a short occurs, the current is limited to 100mA, and the heat generated in the lab supply is 2 watts, hardly enough to overheat the supply. one thing about the warning you posted is rather odd

The lab power supply‘s maximum continuous operating time is 8 hours. After this period, turn off the device and let it cool down to room temperature.

what bothers me about that, is if that's what it takes for the unit to last through the warranty period, then maybe the electrolytic caps used in the supply will dry out prematurely if the supply is left on 24/7. i've worked in a calibration lab, and most of the equipment in a cal lab is left on 24/7 to minimize thermal drift. when i was working at a service center, i left my oscope and supply/function generator/frequency counter running 24/7 for 9 years without them failing. thermal cycling can cause failures in solder connections a lot sooner than caps will develop high ESR (depending on the quality of the caps).

 

[Windows9Pro]

I think, before a week or sth., I accidently touched the alligator clips, when the supply was on. But after a few seconds I thought ''It just can't jump from room temperature to around 100 DegreesC'' so I continued using it.

The problem is, that I sometimes want to make a coil and test it. So I need a current limiter, because I want the power supply to run around 2 mins. with the coil. A short for like 1 second, it would just survive easy.

 

[unclejed613]

depends on the wire you are making the coil with. if you are making an electromagnet, the determining factor on how strong it is is "ampere-turns". the following examples are using the power supply in constant current mode, so presetting the supply to the given current limit must be done before connecting the coils.

if you have 100 turns of #20 wire, you can safely run 1 amp through it, and it will pick up "x" number of paper clips. a coil of 1000 turns of #36 wire running at 100mA will pick up the same number of paper clips, because the number of ampere-turns is the same (100 ampere-turns for both examples), the voltage across the coils will be different, because the wire resistance is different.

the constant current mode is useful here, because you can set a safe current for the wire being used before connecting the coil, instead of just dialing to 15V and hoping the coils won't overheat.

 

[Windows9Pro]

The coil is just an example. Sometimes I play around with graphite. You know, it's conductive like hell, and it gets quite hot, when you run 400mA through it (that's the limit of my resistor. above this level, it starts to smoke). I don't wanna know, what happenes when you run like 2.8Amps through it.
This experiments are the reason, why I need something that limits the current.

 

[unclejed613]

all more reason to use the supply in constant current mode. you can also build a constant current source fairly easy. the following circuit provides a constant current up to the limits of your power supply. Rset determines the current through the load, and the formula is 0.7/I=R (approximately, the voltage term of 0.7 could actually be anything from 0.5 to 0.8 depending on the Vf of Q2).

Q1 should be mounted on a heat sink.
Rset should be at least a 2 watt resistor for currents of 1 or more amps, and 5 watts for 3A. actual wattage in Rset is 0.7*I, and best practice is to use a resistor twice the calculated wattage.

for a current source on the positive side (so the load is ground referenced), turn the circuit upside down and use PNP transistors.

 

[rjenkinsgb]

@rjenkinsgb what exactly does mean trip current ?

The value of current (or voltage, or whatever a circuit is monitoring) that causes a protective or alarm device to act.
eg. Turn off the output, in the case of a current protection.

A "trip" (or tripswitch) is also a common name for a circuit breaker, overload detector or protector in electrical equipment.

eg. https://www2.schneider-electric.com...-catalogue/abted205160en/EN/pdfs/page_036.pdf

https://www.ti-soft.com/en/support/help/electrical/libraries/curves/devices/mcb/tripping-curves