Adjustable Current and Voltage Lab Supply. LM317

[TheLaw]

I want to build a +/- power supply capable of putting out 1.5A on each rail with adjustable voltage AND current.

The way I think would be the easiest is running an LM317 in the constant current mode, and then on the output of that put an LM317 in standard voltage regulation mode.

Datasheet: http://www.national.com/ds/LM/LM117.pdf

Thanks.

 

[TheLaw]

Because I can't just put a pot inline with the output as it would fry in an instant, I'm going to use a single throw 5-6 pole switch and just switch like this. I just made this because I was visualizing myself.

All of those resistors will have to be pretty high wattage, but how do I calculate the wattage that they have to be?

If there is user fault and the circuit is pulling infinite power, these resistors could still technically fry. Like if I have it on the "10mA range" and my circuit wants to draw 10A, that resistor will fry I'm sure, right? ...maybe I could use a resettable fuse in addition or something?

Trying to get my thoughts together. If anybody as any other ideas on how to get simple current limiting in a variable supply, be my guest.

Thanks.

 

[rogerk8]

Simple Discrete Regulator

Attaching a simple discrete voltage and current regulator. My recommendation is that you forget the LM317/337-version and build it with simple transistors. My own power supply is actually of the left type. I have however a fixed maximum current of some 0.5A but this may be arbitrary chosen by selecting/switching R2 and only keeping in mind that current will be limited when the voltage over R2 is some 0.7V. KR/Roger

 

[TheLaw]

Attaching a simple discrete voltage and current regulator. My recommendation is that you forget the LM317/337-version and build it with simple transistors. My own power supply is actually of the left type. I have however a fixed maximum current of some 0.5A but this may be arbitrary chosen by selecting/switching R2 and only keeping in mind that current will be limited when the voltage over R2 is some 0.7V. KR/Roger

Thanks Roger

But isn't R2 still dissipating the entire load?

I was actually wondering if anyone had something that could use a potentiometer and not use crazy high wattage resistors...Though I'm not sure if that's possible.

 

[(*steve*)]

If you want a variable current limit that doesn't require a variable resistor to take the full supply current (which is only OK for quite low currents) then you have several options, however the easiest involves a sense resistor, a voltage reference and a comparator.

The sense resistor can be in either the ground or positive lead for a single ended supply (however it can't be in the ground lead for a dual ended supply). This takes the full current, but is of a very low value (typically less than an ohm) so its power dissipation is quite low.

A comparator compares the voltage across this resistor with a reference voltage (typically variable) and uses this to indicate an overcurrent condition.

This indication could act to reduce the output voltage (essentially limiting the current) or to shut down the regulator completely.

Many integrated circuit amplifiers have connections for a sense resistor and incorporate much of the other circuitry internally.

 

[TheLaw]

If you want a variable current limit that doesn't require a variable resistor to take the full supply current (which is only OK for quite low currents) then you have several options, however the easiest involves a sense resistor, a voltage reference and a comparator.

The sense resistor can be in either the ground or positive lead for a single ended supply (however it can't be in the ground lead for a dual ended supply). This takes the full current, but is of a very low value (typically less than an ohm) so its power dissipation is quite low.

A comparator compares the voltage across this resistor with a reference voltage (typically variable) and uses this to indicate an overcurrent condition.

This indication could act to reduce the output voltage (essentially limiting the current) or to shut down the regulator completely.

Many integrated circuit amplifiers have connections for a sense resistor and incorporate much of the other circuitry internally.

Thanks steve. Very well said. I've kind of switched my mindset to an SMPS however. Sorry to make you write all of that, though it was quite insightful.

I'd like to use an LM2576, which is one of the more popular hobby-type switchers available. In the datasheet there is a very clearcut schematic for a 1-30V 3A supply with obviously adjustable voltage. The question that remains is how one would go about incorporating adjustable current as well. Is it a simple matter? Or does it involve things out of my reach?

Thanks for an advice/input.

 

[(*steve*)]

With a linear regulator, providing a current limit can be as simple as adding a "constant current" stage before the regulator. Essentially this limits the current flowing through the circuit.

However this won't work well with a SMPS

If the current sense circuitry is interfaced to the voltage reference in such a way that the output voltage is reduced when the max current is reached, you have an arrangement that will essentially work for any linear or buck SMPS.

Many SMPS chips have a current sense input that is used to limit the peak current. For a buck SMPS this is reasonably equivalent to an output current limit.

 

[TheLaw]

With a linear regulator, providing a current limit can be as simple as adding a "constant current" stage before the regulator. Essentially this limits the current flowing through the circuit.

However this won't work well with a SMPS

If the current sense circuitry is interfaced to the voltage reference in such a way that the output voltage is reduced when the max current is reached, you have an arrangement that will essentially work for any linear or buck SMPS.

Many SMPS chips have a current sense input that is used to limit the peak current. For a buck SMPS this is reasonably equivalent to an output current limit.

Thanks. I'm trying to figure out how to do this.

I was able to find this which is Constant Current and Adjustable Voltage...But I can't find anything else really. Is there anyway to turn this into Adjustable Current Adjustable Voltage?

 

[(*steve*)]

Yeah, sure.

Any voltage/current regulator is essentially something that passes current and either increases or decreases that based on feedback.

That feedback is often the difference between a fixed reference voltage and the voltage across a sense resistor.

In the case of a constant voltage supply, the sense resistor senses all or part of the output voltage. In the case of a constant current regulator that sense resistor measures the output current.

There is no particular reason why you can't employ both forms of feedback.

In general you have a system where the pass transistor is biased on and either /both of the current and voltage comparators can bias the transistor more toward cutoff.

For a switchmode power supply, the feedback tends to stop pulses appearing on the pass transistor's base/gate, r reduce the duration of the pulse (or both)

 

[Resqueline]

But isn't R2 still dissipating the entire load?

Just for the record: No..
At 3A (regardless if the output is at full voltage or shorted) it dissipates 3A * 0.6V = 1.8W
Replace T3 with an op-amp, a pot, and a reference voltage and you have a continously adjustable current limit, R2 dissipating whatever little you choose.

 

[TheLaw]

Yeah, sure.

Any voltage/current regulator is essentially something that passes current and either increases or decreases that based on feedback.

That feedback is often the difference between a fixed reference voltage and the voltage across a sense resistor.

In the case of a constant voltage supply, the sense resistor senses all or part of the output voltage. In the case of a constant current regulator that sense resistor measures the output current.

There is no particular reason why you can't employ both forms of feedback.

In general you have a system where the pass transistor is biased on and either /both of the current and voltage comparators can bias the transistor more toward cutoff.

For a switchmode power supply, the feedback tends to stop pulses appearing on the pass transistor's base/gate, r reduce the duration of the pulse (or both)

I actually understood that. Not bad for being an idiot.Unfortuneatley I can't find the original source of this schematic so I'm not sure which resistor is for feedback for current. The pot circled in the above pic is for voltage. Not sure why they circled it, but that's for voltage...

Now to find out how to control current...

Thanks.

EDIT: http://www.ottomat.hu/Kapcsrajzok/aramgenerator3.pdf

Found it. Now gotta figure it out. The problem is that it's constant current but I want adjustable current.

 

[(*steve*)]

Rsc is the current sense resistor.

Pin 4 (FDB) is normally connected to a voltage divider which causes approx 1.25V to appear on this pin when the output voltage is correct.

To add a voltage limit to this power supply, (for 10V (say)) you would connect an appropriate pair of resistors across the output (approx 1k and 150R) and connect the junction to the FDB pin.

Then place a diode between the output of IC2b and pin 4 of the reg (anode to pin 4 of the reg) so that all it can do is pull the voltage down (also check the current sink capability of the TL082).

This particular arrangement doesn't lend itself to being made variable easily.

 

[duke37]

I bought a power supply from a radio amatuer junk sale giving dual variable voltage and current. Surprisingly it worked! I have no idea how old it is.

It contains a pair of L200C chips which control current and voltage.

 

[TheLaw]

I bought a power supply from a radio amatuer junk sale giving dual variable voltage and current. Surprisingly it worked! I have no idea how old it is.

It contains a pair of L200C chips which control current and voltage.

yeah L200 is a pretty nice chip in that respect. The only problem I have with it is that it is a linear regulator and I really don't feel like paying for a huge transformer and huge heatsinks when it can be done with a switching design for cheaper, smaller and more efficiently.

I'm a newbie still. I get most of the stuff you guys are saying, but not to the point where I could just go design one myself. I'm still a senior in high school trying to figure things out here...

Thanks.

 

[TheLaw]

I bought a power supply from a radio amatuer junk sale giving dual variable voltage and current. Surprisingly it worked! I have no idea how old it is.

It contains a pair of L200C chips which control current and voltage.

You know what?! I don't need a 3A output. Hell, I probably don't even need 500mA. I'm not even sure what I want. I guess that's the first problem.

L200 definitely is more appealing to me now. I believe I will try to use the dual L200 +/- supply detailed in the L200 Designers Guide. Adj voltage and current and +/- in two chips. Sweet. I'll get a cheap 4A transformer. I'll probably set the max current to about 1A on each rail to make sure I don't overload the transformer. Damn inefficient devices!

Well I need some rest but I'm putting the switching design to rest for the time being.

 

[davelectronic]

current limiting power supply

Hi there again.
Why not have a look at this, it a project and a half, nice unit and design from Tony Van Roon's web site, the end result will be a very nice unit indeed, some might say dated, i think its a classy bit of kit. Dave.

http://www.sentex.ca/~mec1995/circ/ps4002/ps4002.html

 

[TheLaw]

Hi there again.
Why not have a look at this, it a project and a half, nice unit and design from Tony Van Roon's web site, the end result will be a very nice unit indeed, some might say dated, i think its a classy bit of kit. Dave.

http://www.sentex.ca/~mec1995/circ/ps4002/ps4002.html

Woowee! That's an old chip! I'm sure someone has it. Seems like the guy who made it has a clue though. Reminds me a bit of the LM741 supplies.

 

[davelectronic]

Ajustable psu

I bet there is a substitute for that IC if its not available and iam curtain it can be made today, and i have a very good " clue " how to build it, but there is often more up to date circuits like the L200 IC's, most bench psu's dont require more than 2 to 3 amps, i like the high current monsters, mostly to drive 0.5 KW linear RF amps, POWER. Dave.

 

[TheLaw]

I bet there is a substitute for that IC if its not available and iam curtain it can be made today, and i have a very good " clue " how to build it, but there is often more up to date circuits like the L200 IC's, most bench psu's dont require more than 2 to 3 amps, i like the high current monsters, mostly to drive 0.5 KW linear RF amps, POWER. Dave.

Well I don't do too much power stuff. Usually low current stuff. So I figured 1-2A should be enough. And with a linear supply, I don't want to spend $999999999 on a transformer to power this inefficient soab.

 

[davelectronic]

Bench psu

Smp stuff is great i use a modded atx to power radio gear, and its as good as gold. Transformers in the Uk are not to steep in price, chassis mount only go so high in VA ratings, teroidals go much higher, i can get a 300 VA unit for about a little over £ 30.00 not a terible price for a new unit, linear psu's are always going to be bigger bulkier in size, early smp's suffered a lot of noise, but current technology is far better today. Dave.