# Constant current: Does the current change when the source voltage changes?

#### Tonyr1084

Joined Sep 24, 2015
7,998
I built a CC source for some LED's. Set it to run at 250mA (not your typical 5mm LED). When I powered it from a 12VDC source I had to adjust the current to get what I wanted. Almost couldn't get to the 250mA. Tried powering it from 24VDC and got a whole lot more current. I thought the CC was supposed to hold current through all the changes. I'm thinking I must be wrong.

What I am more confident in is that if the load changes the CC will remain the same. But I thought as long as the source was sufficient, going from merely sufficient to plenty of headroom in source voltage the CC would remain the same. Is that a wrong assumption?

#### LowQCab

Joined Nov 6, 2012
4,300
It depends on the capabilities of ALL of the components in the "chain" and
on the capabilities of the Power-Supply.

It only takes one weak-link in the chain to degrade the expected results.

Do You have a Schematic with part-numbers ?
.
.
.

#### Tonyr1084

Joined Sep 24, 2015
7,998
I've got some stuff. Will have to find an older post of mine. Will need some time this afternoon. Trying to do two things at once, both suffer. Let me clear my pallet and I'll get back.

#### MrChips

Joined Oct 2, 2009
31,087
If the current is changing then it is not constant current, is it?

For the constant current controller to work you need some headroom on both current and voltage.
If your power supply is a voltage source then you need a series resistor. This does not give you a constant current source but it is better than nothing.

The higher the resistor value the better it is at providing close to a constant current. An ideal constant current source has infinite series resistance and almost infinite supply voltage.

My rule of thumb for simple voltage source and series resistor, make the supply voltage greater than twice the load voltage.
Thus, if the LEDs run on 12VDC, make the supply voltage 24VDC. Calculate the required series resistance from there.

If the LED current is 250mA then the series resistance is
(Vs - VLED) / 0.25A = 48Ω
Wattage required = 5W

#### dl324

Joined Mar 30, 2015
17,137
I thought the CC was supposed to hold current through all the changes. I'm thinking I must be wrong.
It depends. Was there enough supply headroom for it to operate as intended? A schematic would be helpful.

#### ElectricSpidey

Joined Dec 2, 2017
2,885
As well as the needed overhead you also need "line regulation" as well as "load regulation" to achieve what you desire.

#### Ramussons

Joined May 3, 2013
1,418
I built a CC source for some LED's. Set it to run at 250mA (not your typical 5mm LED). When I powered it from a 12VDC source I had to adjust the current to get what I wanted. Almost couldn't get to the 250mA. Tried powering it from 24VDC and got a whole lot more current. I thought the CC was supposed to hold current through all the changes. I'm thinking I must be wrong.

What I am more confident in is that if the load changes the CC will remain the same. But I thought as long as the source was sufficient, going from merely sufficient to plenty of headroom in source voltage the CC would remain the same. Is that a wrong assumption?
If your source is not able to supply the required current, the CC unit will not perform as it should.

#### Tonyr1084

Joined Sep 24, 2015
7,998
Still looking for the schematic. If I have to I'll reverse engineer the one I built. As simple as that is, the unit is in a housing sealed from weather. It's easy enough to open up but I'm trying to not go there.

Will continue to search.

As for the question as to whether there's sufficient headroom, yes, there is. But without a schematic we're discussing this blind. Sorry. I'm still trying to get all the info together.

#### MrSalts

Joined Apr 2, 2020
2,767
Changing the change in voltage can change the base current to the biasing transistor which changes the gain to the current controlling transistor. Show your circuit.

#### Tonyr1084

Joined Sep 24, 2015
7,998
This is all I can find. Too generic I suppose, but the power source is a 24VAC transformer full wave rectified and filtered with caps. The LM317 is supposed to be set to provide 240mA for a set of three UV-C LED's. They're inside a fish pond filter where they can cause no harm to human or aquatic eyesight.

When I was experimenting and adjusting it I tried a 12VAC source rectified and filtered but had difficulty reaching that current. But when I supplied it with 24VAC the current jumped way up. According to the calculations, 250mA = 1.25V/5Ω. Had to parallel two 10Ω resistors to get 5Ω. 250mA would do quite well as the LED's are rated up to 300mA.

Sorry, I just can't find any more info without disassembling the whole thing. And right now the pond is frozen over.

Tell me, did I get it wrong?

#### MrChips

Joined Oct 2, 2009
31,087
Check the LM317 pinout.

#### MrSalts

Joined Apr 2, 2020
2,767
I built a CC source for some LED's. Set it to run at 250mA (not your typical 5mm LED). When I powered it from a 12VDC source I had to adjust the current to get what I wanted. Almost couldn't get to the 250mA. Tried powering it from 24VDC and got a whole lot more current. I thought the CC was supposed to hold current through all the changes. I'm thinking I must be wrong.
That all seemed ok until you posted this...

The LM317 is supposed to be set to provide 240mA for a set of three UV-C LED's
how many UV-C LEDs are in series? These will have forward voltages of 4 to 8v each. The 4-volt versions out there are not likely UVC but closer to visible.

Have you measured them or do you have the datasheet? I'm guessing that, at 12v supply, your headroom isn't what you think it is.

#### crutschow

Joined Mar 14, 2008
34,842
The LM317 needs at least 3V headroom in that circuit between the input voltage and the load from proper current regulation.
What voltage do the LEDs require?

#### Tonyr1084

Joined Sep 24, 2015
7,998
how many UV-C LEDs are in series?
Three.

OK, I NEED TO MAKE SOME CORRECTIONS TO MY EARLIER STATEMENTS. First, now that I've found some more information I recall more clearly that I was running the LED's on a current of 135mA and NOT 250mA. It's been a while and I'm pulling this out of a faulty memory (not the computer memory). Here is the LED I purchased (the three chip version with a 5.6Vf per LED)
https://www.ledsupply.com/leds/uv-c-275nm-seoul-viosys-led

ANOTHER CORRECTION: I did attempt to run it on 12VDC using THIS "Bucktood DC LED Driver". I had issues with it and ended up sending it back for a refund. I couldn't get it to maintain the necessary current. It's recommended voltage range is 5V to 28V DC. Obviously 5V wouldn't work since the LED's run each with a 5.6Vf. When I spoke with an agent on the phone he said the BuckToot would easily power the three chip LED. As I said, I purchased the 3X LED version with sales receipt that I did in fact purchase the 3X. They sent me a 1X LED. That was shipped back when they replaced it with the one I paid for, the 3X LED. Either the sales agent didn't listen to what I was explaining OR I failed to convey clearly and correctly what I was trying to accomplish. As I said, he sent me a single chip type when I wanted the 3X chip. Something went wrong.

I had issues with these people, which is why I decided to build my own CC source. I mis-spoke before, saying "I powered it from 24VAC". I used a transformer with 6V, 12V and 18VAC outputs. The transformer was full wave rectified and filtered. So the input voltage to the 317 was approximately 25.5 volts DC.

Hopefully this clarifies my situation. My initial question, which seems to have been answered is that I wanted to know if changing the input voltage to a CC regulator circuit will change the amperage output.
Changing the change in voltage can change the base current to the biasing transistor which changes the gain to the current controlling transistor. Show your circuit.
As for showing my circuit, sorry, I just can't find the original schematic. And since the unit is now live in the pond and iced over, I can't get to it just now. Nevertheless, I am not seeking in solving a problem, I'm seeking greater knowledge. Will a CC circuit as shown in post #10 with a set current change when the source voltage changes? I think I got my answer. I have been assuming that as long as there is sufficient headroom the voltage can change but will maintain a constant current regardless. I'm now believing I was wrong in that assumption.

#### BobTPH

Joined Jun 5, 2013
9,276
think I got my answer. I have been assuming that as long as there is sufficient headroom the voltage can change but will maintain a constant current regardless. I'm now believing I was wrong in that assumption.
No, your initial understanding is correct. With the LM317 and 3 x 5.6V LED chips in series you need about 16.8+2= 18.8 volts input. so, of course it did not work at 12V.

#### crutschow

Joined Mar 14, 2008
34,842
With the LM317 and 3 x 5.6V LED chips in series you need about 16.8+2= 18.8 volts input.
The LM317 requires closer to 3V when operated in the constant-current configuration (≈1.7V across the LM317 plus 1.25V across the current-limit resistor), so the minimum voltage required is more like 19.8V.

#### BobTPH

Joined Jun 5, 2013
9,276
The LM317 requires closer to 3V when operated in the constant-current configuration (≈1.7V across the LM317 plus 1.25V across the current-limit resistor), so the minimum voltage required is more like 19.8V.
Ah, I didn’t know that.

#### MisterBill2

Joined Jan 23, 2018
19,453
I have used a circuit like that with a 7805 voltage regulator IC and it worked very well in a very successful product. .BUT the supply voltage must exceed both the required maximum load voltage PLUS at least5 volts for the regulator to develop across the resistor, PLUS the required overhead voltage for the regulator. So there is a very definite minimum voltage required. above that minimum voltage the current remains constant withing just 1 or 2 mA. That was delivering a current of 100 mA.
It was not very efficient because of the higher source voltage required, but it used much less power than the system it replaced. And it never required recalibration.