Hi all - I'm still at it (putting together my lab PSUs) & the doctor says the bandages should be off my hands in a week or two...

Question: Can a selectable current limiting switch be introduced to a variable voltage line to provide the user a set range for the current to load?

I've done multiple searches on this site, read lots of very complicated (to me, lol) stuff about BJTs w/ 317s, etc., as well as looked around the web. I just can't seem to find an explanation that's translates to something in my comprehension range (read that - "newbie dumbie").

Background: I have a very clean 12V supply to start with. I'm not sure of the current out. The PS was part of a very expensive RF rack panel from the 80s & it's still in great shape. As an example, the LCD meter in it is from Texmat & currently lists by itself for $79 - I got the whole unit for $15. The number of bourns variable resistors on the RF board alone make up the total cost, lol, but I digress.

Eventually I want to build a 30VDC supply to run the adjustable voltage sub-circuit, but for now, the 12 will do. The sub-circuit is the simple 317/337 design, widely referred to on this site & the web. I just would like to add a selectable range to the current (which won't be metered, just the volts are metered) so I can limit the current from <1.5A to down around 50mA (i.e., 1, 500, 250, 100, 50). That's to say, turn a knob, light an led at the selected current range.

I've seen a lot of items for limiting current w/ transistors, but just having a tough time finding one that uses resistors.

My brain is telling me this is way easy. I'm just not getting my mind around it for some reason. Anyway, thanks for the time.

 

If you want to limit the current, you have to measure it somehow. how else do you want to detect if your load tries to draw more than the allowed amount.
What exactly is your goal:
- limit the current such that the maximum amount is regulated at your desired level?
OR
- just some kind of simple, resettable fuse that switches off your circuit if the current exceeds the maximum allowed amount?

 

Do you want to vary it with a voltage input, a variable resistor, or something like a rotery switch?

 

Yes, your goal needs clear. There are two current control modes that can be used in a bench type general purpose power supply. First is an adjustable current limiter and the second is an adjustable constant current source.

An adjustable constant current source will try and maintain the desired 'setpoint' current into the load by raising or lowering the voltage even into a short circuit. Of course it can only raise the voltage to the limit of the voltage regulator's capacity. With this kind of function one could wire an LED without a series resistor, set the constant current to say 20ma and the circuit would automatically adjust the regulators output voltage to maintain 20ma.

An adjustable current limiter is a circuit that is passive unless and until the current to the load increases to the adjusted 'setpoint' current value and will then start lowering the voltage to maintain current to equal or less then the adjust value. That is it acts like a constant current source only once the current reaches the setpoint adjusted value, below that value it is passive and the power supply stays in a constant voltage mode.

Constant current is a mode of power supply operation where current limiting is more of a protection feature in case something goes wrong in the circuit being powered. Both are useful for different reasons. The popular 3 terminal linear regulators already have a over current limiting protection built in, but it's current setpoint value not adjustable.

Lefty

 

I've attached a schematic that lays out my problem - how to simply connect the two ideas of using a 317 for voltage & current regulation. I hope it helps clarify which type of current limiting I need to achieve.

Some other thoughts on this line:

Adding a course/fine set-up trimmer, but since it won't be metered @ this time, seemed pointless.

Replacing some of the lower end resistors with a resistor & trimmer combo to be able to fine tune the current <100mA. Means I'd have to measure it manually (again, sans meter).

Adding overcurrent protection from shorts

Replacing the rotary switch w/ a decade counter/push switch set-up.
​

For those who would ask, the supply will be within the limits of the 317 specs (>7VDC, <36VDC; 3A).

Once I understand & have breadboarded this set-up, I intend to duplicate it for a negative rail; then connect the two together with an op-amp(s?) & combine the adjustments so both rails are adjusted simultaneously from single sources.

Thanks for the help.

 

Your design is pretty much the standard way to achieve your goal. Please note the following comments:

1. R16(5K VR) of the second LM317 is wrongly connected and should be connected to 0V instead.

2. R17 value of 240Ω is suitable for LM117 only and for LM317, one should use a smaller resistor value like 120Ω to meet the min. output current load specification.

3. only one resistor is needed for the all LEDs and the resistor be connected from the junctions of all LEDs to ground. The common of the selector switch should be Vin instead of the first LM317's output or else the LED will dim if the unit goes into current limit. Use a suitable value to match the current of the LED. Keep the existing connection if this dimming effect is what you actually wanted.

4. use a make-before-break selector switch for current selection. If you fail to get one, you can make do with the use of a large value of input capacitor for the second LM317 input(C9) so that the output current is not cut off during switching. A 2200μF capacitor is suitable.

 

Thanks for the review. I made the corrections (I think, lol) & attached the updated schematic. Couple questions (as always )...

Why couldn't the LED resistors be placed on the lead to the switch common? Is the resistance through the switch significant so that those who know what they're doing compensate for that in the design? I'm just use to putting my LED resistors on the anode side (force of habit I guess).

When I compute the POWER for the current resistors, it ain't pretty. The lower mA settings aren't too bad, i.e. 50mA resistor has to get rid of 0.065ω so can use a 1/4 watt on it (using the doubling technique). The 100mA works out to 0.13ω, so really should go to 1/2 watt since doubled places it above the .25 mark. Going further upwards ain't so pretty, especially as I get closer to that 1A mark! Unless, again, I'm fat fingering the calculator (P=(ExE)/R, correct?) 1.25V squared is 1.5625; divided by the various resistors, yields the following: 250mA, 0.325ω (3/4 or 1 watt); 500mA, 0.625ω (2 watt); 750mA, 0.919ω (2 watt); & finally, 1A, 1.9ω (4 watt). Am I on track with these numbers?

Finally, 0V is not ground, correct? I'm treating them as seperate entities for wiring & design. Is this correct?

 

Thanks for the review. I made the corrections (I think, lol) & attached the updated schematic. Couple questions (as always )...

Why couldn't the LED resistors be placed on the lead to the switch common? Is the resistance through the switch significant so that those who know what they're doing compensate for that in the design? I'm just use to putting my LED resistors on the anode side (force of habit I guess).

You can put the resistor on either side.

When I compute the POWER for the current resistors, it ain't pretty. The lower mA settings aren't too bad, i.e. 50mA resistor has to get rid of 0.065ω so can use a 1/4 watt on it (using the doubling technique). The 100mA works out to 0.13ω, so really should go to 1/2 watt since doubled places it above the .25 mark. Going further upwards ain't so pretty, especially as I get closer to that 1A mark! Unless, again, I'm fat fingering the calculator (P=(ExE)/R, correct?) 1.25V squared is 1.5625; divided by the various resistors, yields the following: 250mA, 0.325ω (3/4 or 1 watt); 500mA, 0.625ω (2 watt); 750mA, 0.919ω (2 watt); & finally, 1A, 1.9ω (4 watt). Am I on track with these numbers?

Finally, 0V is not ground, correct? I'm treating them as seperate entities for wiring & design. Is this correct?

The node you labeled 0V should be connected to ground. Ground is not necessarily earth ground, it's just circuit common. It's the negative terminal of your power supply.
I would change that 2200uF cap to 100nF. This cap has to discharge before your programmable current limiting takes effect, so the smaller the better. 100nF is generally required for stability.
Keep in mind that power dissipation in the two LM317s will limit your combinations of voltage and current out.

 

Thanks, Ron, for the clarifications.

Glad you confirmed I could move those LED resistors (works out better for me on the board). As mentioned, I was just concerned I was missing something in regard to loss through the resistance in the switch. I suspected is was negligible, but don't know these things.

I will change the schematic to reflect the "0V" is ground per your recommendation. Eventually this line will be the centertap common off the power rectifier for the +35/0/-35. When I get to that point, I'll be asking if I should also add an earth ground into the network, but that's for later.

Ron wrote:

...Keep in mind that power dissipation in the two LM317s will limit your combinations of voltage and current out. Yesterday 09:49 PM

Planning on it. Counting on it actually when computing the LED resistors.

Thanks again to all for the help. One of these days I gotta post a picture of this contraption in the Projects Forum, lol.

 

Doug, can you explain why you have two resistors in your LED circuit?

 

I had the same question as Ron_H.

If you're planning for having a widely variable input voltage (looks like you put in 9.9v to 32.9v, but I can't read it that well) then you could instead of R10 and R11, use an LM317L as a current regulator with an appropriate resistance value from OUT to ADJ. 100 Ohms would give you about 12.5mA, and power dissipation in the regulator would be about 388mW when Vin was 33v (assuming the LED Vf is around 2v).

 

Ron & Sarge: LOL! You guys....just when I think I'm on track...oh well.

I needed the 2 in series to get the exact 1680Ω I calculated for 20mA when the voltage was 32.9V & still be able to have an acceptable amount available for voltage adjusted down to 9.9V (although it's dim, but still clearly visible).

I like the 317 idea better. Thanks a lot. Will also make the board look cleaner.

I also plan on doing the heavy duty resistors in series to get the ohms I calculated were necessary for the current limits. Hope that's okay.

(evil laugh) My "dealer" also had some 5V to 12V, DC to DC converters at a good price. Then I had a brain storm & bought one. Thinking about adding a DPDT heavy duty switch to swap between the dedicated supply set for the circuit & another one coming of the ATX unit (running off both the 5V rails). That'll end up giving me two variable supply sources capable of 0 volt, one capped @ 12V (well, actually around 10V) & the other up to around 30V.

Anyway, thanks for the input & let me know if I'm way off target here.

Doug (I know, it's either I'm gonna blow myself up or put out an eye)

 

Ron & Sarge:

I made your recommended corrections/additions & have attached the updated schematic.

Thanks to everyone for reviews/comments.

Doug

 

A couple of comments:
Why did you put the ground for the pot so far from the pot? It makes the schematic look really weird.
With a 5k pot, much of the range is unusable. You could change R17 to 82 ohms, change the pot to 2k, and have most of the pot range usable. A second pot somewhere in the circuit could be used to trim the full scale to 30V, if desired.

 

Ron_H wrote:

Why did you put the ground for the pot so far from the pot? It makes the schematic look really weird.

lol, so I would recognize it as a pad location. There is another schematic I'm working with (it doesn't have the datasheet circuits on it) to develop a board layout in Eagle. I've attached wirepads at various locations so when I lay out the board I can adjust their location. I'll clean-up the schematic & update the previous post to make it more correct.

Ron_H wrote:

With a 5k pot, much of the range is unusable. You could change R17 to 82 ohms, change the pot to 2k, and have most of the pot range usable. A second pot somewhere in the circuit could be used to trim the full scale to 30V, if desired.

Excellent! I had wondered about the pot. Was just basing the size on various readings around the web since I wasn't sure what math to use to determine an approriate size. Can't remember what I was working on but I've actually ran into a situation in the past where the pot "maxed out" with very little adjustment. Very frustrating.

But now I'm confused for a different reason. Why a smaller pot? If I want decreased sensitivity, that is to say a larger dial movement for less affect on the circuit, wouldn't I want to go larger in capacity (i.e., 10k)?

If a 2nd pot is added, would that create a "coarse/fine" set-up? In this type arrangement would I place Pot 2 in series with Pot 1? What can I expect to happen in the meter display since it only goes to 1 decimal place?

I had thought about doing something similar on the current side (see earlier posts). From what I read in Chap 6 of the tutorials, I thought for current that I could set a range of say 50mA to 250mA & then use a pot as a rheostat to adjust between these two. I shelved the idea because I'd have to modify my panel for a second meter (which may happen anyway, lol).

One other question...

Ron_H wrote:

I would change that 2200uF cap to 100nF.

Why the nF? Why not the μF units & designator? As an aside, I had put in the 2200μF cap at the suggestion of eblc1388. In his response to the original schematic I posted, he recommended it if I didn't have a MBB switch. I'm not sure what type switches I have, so I went with it just in case.

Ya know, if I keep plugging away at this, I may actually learn something.