limiting current

Thread Starter

the apprentice

Joined Jan 30, 2005
6
hello all im almost finished building my dual power supply and have hit a snag, i need to add a current limiter so i can a smoke test all my future projects. can anyone please suggest a simple current limiter other than a high power variable resistor in series with the output. if i do decide to use this approach is there a downside to using it? can the tech gurus please help me out? :)
 

Dave

Joined Nov 17, 2003
6,969
I am moving this thread into the General Electronic Chat where the topis is more appropriate, also it will gain better exposure.

Thanks.
 

vineethbs

Joined Nov 14, 2004
56
hello "the apprentice" am trying to make a 5V , 3A power supply , i have an idea which worked just fine in simulation , i used something like a current mirror to produce a proportional current in another transistor which can be used to provide short ckt protection by turning on another transistor to divert the base driving current . But because am ofcourse using a darlington , the current diversion isn't producing that much of an effect :( .
 

vineethbs

Joined Nov 14, 2004
56
the downside to my idea is that the temperature increase can screw up the current ratios , also because the things r discrete and unmatched , it wud require some tinkering before it works fine , i think .

am also thinking of mebbe adding a schmitt to convert the controlling voltage to swing between 5 and 0 so that i can mebbe drive the diverting transistor to hard saturation
 
Generally speaking, current limiting is best built into the voltage regulator. There are a lot of ways to do it, but the choice depends upon your needs. Is this a switching converter or a linear power supply? Is the voltage fixed or adjustable?
How much current and voltage do you want to get? What parts do you have and how much money do you want to spend? How much room do you have? BEWARE OF HEAT which is the most likely killer for linear power supplies. These affect your choice.

Most current limiters tend to be sloppy. You generally have a tradeoff between cost and heat and accuracy.

1) One of the easiest ways is with a "Polyswitch" thermal limiter. These are thermally variable positive-temperature coefficient resistors. When the current rises beyond a point, they begin to heat and as the temperature passes a limit the resistance climbs rapidly. This causes the device to heat more. As long as the maximum voltage is less than the device ratings it will survive OK but beware that after cooling down it takes (probably days) to reset to the initial starting point. The upshot is that they are sloppy and very temperature dependent and not adjustable.

2) An old mainstay of linear power supply design was the 723 voltage regulator. I think that National and others still make it; the applications info may show some current limiting techniques. You almost always have to increase the available input voltage slightly in order to compensate for drop in a current sense resistor.

3) Overall, an easy (but not efficient) approach is to use an integrated linear regulator, as they all offer some sort of thermal limiting and current limiting. These are most practical for current under 3 Amps or so. LM317 will do nicely for up to 1.5 Amps while LM338 or LM350 can handle higher current. Make sure you do an adequate job of cooling these beasts (heatsinking requirements may make us for the easy circuit design.)

Whaddya think?
 

Thread Starter

the apprentice

Joined Jan 30, 2005
6
Originally posted by mozikluv@Feb 5 2005, 02:58 PM
hi

would you like to have an adjustable regulated voltage output and an adjustable current output all in one set up? :)
[post=5051]Quoted post[/post]​
this is essentially what im after, an adjustable output and adjustable current output. At the moment the power supply is as follows: 0V to 22V and the other is 0V to 30V all thats needed is current limiting. :unsure:
 

Thread Starter

the apprentice

Joined Jan 30, 2005
6
while LM338 or LM350 can handle higher current. Make sure you do an adequate job of cooling these beasts (heatsinking requirements may make us for the easy circuit design.)

Whaddya think?
[post=5085]Quoted post[/post]​
[/quote]

i had a look at the lm338 datasheet and found an adjustable current limiter 0 to 5A my question is the second regulator output pin is going to -5 to -10v so i cant really use this unless i have a negative power rail? is this correct? :(
 

mozikluv

Joined Jan 22, 2004
1,435
while LM338 or LM350 can handle higher current. Make sure you do an adequate job of cooling these beasts (heatsinking requirements may make us for the easy circuit design.)

Whaddya think?
[post=5085]Quoted post[/post]​
i had a look at the lm338 datasheet and found an adjustable current limiter 0 to 5A my question is the second regulator output pin is going to -5 to -10v so i cant really use this unless i have a negative power rail? is this correct? :(
[post=5112]Quoted post[/post]​
[/quote]


hi,

base on your 1st post you said you have a dual power supply. anyway you still can make a negative rail out of a single rail. you just have to make a side circuit for that, nothing complicated :)
 
i had a look at the lm338 datasheet and found an adjustable current limiter 0 to 5A my question is the second regulator output pin is going to -5 to -10v so i cant really use this unless i have a negative power rail? is this correct? :(
[post=5112]Quoted post[/post]​
hi,

base on your 1st post you said you have a dual power supply. anyway you still can make a negative rail out of a single rail. you just have to make a side circuit for that, nothing complicated :)
[post=5132]Quoted post[/post]​
[/quote]

Well, at least, I got some ideas cooking! Normally in my design work I would not choose to make a negative supply just to make a positive supply. But as stated, it may be easy, depending upon the rest of your circuit. We still don't know what that is!! It is easier to make a negative voltage from AC input than from DC, but you can make any voltage from any voltage if cost is no object.

It looks to me like National's purpose for the negative supply is so the output can current limit down to 0 volts. If you use ground instead, you will lose the adjustable current limiting when the output goes below a few volts.

If you are looking for 22 or 30 volts at 5 Amps, that is mucho heato for a linear regulator, amigo. Hope you have liquid nitrogen tanks ready. Actually your linear regulators will go into thermal shutdown......

With 35 volts input and 30 volts 3A output from a linear regulator, you will produce 5V * 3A = 15W of heat. That will be alot for LM338 or any of those. Big heatsink + fan will be required. As Vout goes down, the heat goes up if current is the same.

How are you normally adjusting your output voltage?

Best regards,

topedopudding
 

Thread Starter

the apprentice

Joined Jan 30, 2005
6
1.) We still don't know what that is!!
2.) If you use ground instead, you will lose the adjustable current limiting when the output goes below a few volts.
3.) Actually your linear regulators will go into thermal shutdown.....5V * 3A = 15W of heat.
4.) How are you normally adjusting your output voltage?

1. the power supply is based on lm317 with pass transistors on large heatsinks and the other channel is a replica of the first one.
2. so does this mean i can still use current limiting over say 4v? 5A is way too much im only after say maybe 0 to 2A maybe 3A max limiting.
3. but if you had a pass tranny it would be handling most of the current (heat)? yes?
4. a 5k and 10K pot
i have actually tested the PSU loading it with a window motor in series with a rheostat the heatsinks do get a little warm but i will be replacing them with extra large ones (free at work). the rheostat gets very warm when i limit it to say just under an amp, the motor consumes upt to 4A on start up, 1.80A running. this is the reason i dont want to use rheosats as a limiter i think there has to be a better way. thanks for all your reply please keep em coming as im still learning :)
 

mozikluv

Joined Jan 22, 2004
1,435
Originally posted by the apprentice@Feb 4 2005, 03:36 AM
hello all im almost finished building my dual power supply and have hit a snag, i need to add a current limiter so i can a smoke test all my future projects. can anyone please suggest a simple current limiter other than a high power variable resistor in series with the output. if i do decide to use this approach is there a downside to using it? can the tech gurus please help me out? :)
[post=5032]Quoted post[/post]​

hi,

now we get a clear picture of what you intend to do. you didn't mention anything about a motor as your load. with that in mind that would be a different ball game. there's a power ic that can handle that but i just can't pin down the ic # for the moment. will look into my files, but i believe other members can come up with a circuit or suggestion suited to your need. :)
 
Originally posted by the apprentice@Feb 9 2005, 02:12 AM
1.) We still don't know what that is!!
2.) If you use ground instead, you will lose the adjustable current limiting when the output goes below a few volts.
3.) Actually your linear regulators will go into thermal shutdown.....5V * 3A = 15W of heat.
4.) How are you normally adjusting your output voltage?

1. the power supply is based on lm317 with pass transistors on large heatsinks and the other channel is a replica of the first one.
2. so does this mean i can still use current limiting over say 4v? 5A is way too much im only after say maybe 0 to 2A maybe 3A max limiting.
3. but if you had a pass tranny it would be handling most of the current (heat)? yes?
4. a 5k and 10K pot
i have actually tested the PSU loading it with a window motor in series with a rheostat the heatsinks do get a little warm but i will be replacing them with extra large ones (free at work). the rheostat gets very warm when i limit it to say just under an amp, the motor consumes upt to 4A on start up, 1.80A running. this is the reason i dont want to use rheosats as a limiter i think there has to be a better way. thanks for all your reply please keep em coming as im still learning :)
[post=5160]Quoted post[/post]​
Ahhhh Hahh!

Now, If you are using an LM317 with a pass transistor, the easy way is with a PNP pass transistor. I have attached a quick sketch of an easy-sloppy way to do it.

1) If you mount at least one of the pass transistors very close to the LM317, then the thermal protection of the LM317 may also be able to protect the pass transistors.

2) The 3 diodes I drew (can be 1N5400, 1N5401, etc.) are an approximate voltage clamp to limit the voltage between the TIP32 transistor base and the + supply input. With the 3 diodes operating at significant current, you might get 2.5 V when the regulator is at current limit.

3) The transistor Vbe (maybe 1V at high current) get subtracted from this 2.5V or so and the remainder is across the 0.75 ohm resistors. So with these numbers, each transistor will limit at around 1.5V/0.75 ohms = 2 Amps. You get 4A total.

4) I re-iterate that this is very rough, so you may need to adjust to get the current limit you want. You can go to 2 or 4 diodes instead of 3 or adjust the emitter resistor value.

5) An LM317 when used normally should have an output capacitor. In this boosted configuration, you should have more of it. Best to use a "low esr" type, and I will guess something like a Panasonic FM series (avail from Digi-Key) 120 uF 50V assuming that 50V is more than your max output. It is fine to use several smaller capacitors in parallel or to use a lot more if you have something older.....

6) Really, really, really, for this much power, you want a switching regulator instead of a linear one. Switchers conserve power instead of burning it. But we would be at this all year for me to coax someone how to build a switching regulator to make it work correctly. It turns out that the right schematic diagram is only one part of the task. Getting the layout correct is very important. It is sort of like giving an interior decorator the parts and drawings to build a modern automobile engine and expecting them to build a good one.

Please let me know how you do.

Best regards,

Torpedopudding
 
Top