Constant current power supplies

I'm pretty sure that a constant current switchmode power supply (LED Driver) and a constant voltage switchmode power supply (cell phones) are awfully similar. One would take feedback from the load voltage, and the other would take feedback from current sense resistor om series with the load. Any switchmode chip ought to be able to implement either type of supply with minor changes in wiring.

In "Lab" power supplies, "CCCV" generally means "constant voltage with a current LIMIT", and defines the way the supply behaves when the current limit is exceeded. This is supported by the TND329 datasheet:

(section 5: "Circuit operation")
For output currents less than 1 A the circuit performs as a constant voltage source. ... Although very simple, this current sense circuit will provide a constant current output of approximately 1 A all the way down to an output voltage of 1 V.

In other words, a "CCCV supply" provides a constant voltage up to the current limit, and then decreases the voltage as needed not to exceed the desired current. (as opposed to simpler regulators, that might just shut down in over-current situations.) Some Li-ion battery charging chips assume that the power supply has a current limit built-in to it, but a cell phone supply will normally operate in a constant voltage mode.

In general, that's not quite the behavior you'd want from an LED driver style of CC supply. In that case, you want the voltage to go to any value needed to force the target current through the load. This permits you to use varying numbers of LEDs in series in your light, for instance. I haven't studied the current flood of "LED driver" chips to see how they're different from other switchmode chips; there may be other simplifying factors that permit a cheaper design and justify having a separate chip (oh, like not needing to provide clean DC, for instance.)

westfw:
I'm pretty sure that a constant current switchmode power supply (LED Driver) and a constant voltage switchmode power supply (cell phones) are awfully similar. One would take feedback from the load voltage, and the other would take feedback from current sense resistor om series with the load. Any switchmode chip ought to be able to implement either type of supply with minor changes in wiring.

In "Lab" power supplies, "CCCV" generally means "constant voltage with a current LIMIT", and defines the way the supply behaves when the current limit is exceeded. This is supported by the TND329 datasheet:

(section 5: "Circuit operation")
For output currents less than 1 A the circuit performs as a constant voltage source. ... Although very simple, this current sense circuit will provide a constant current output of approximately 1 A all the way down to an output voltage of 1 V.

In other words, a "CCCV supply" provides a constant voltage up to the current limit, and then decreases the voltage as needed not to exceed the desired current. (as opposed to simpler regulators, that might just shut down in over-current situations.) Some Li-ion battery charging chips assume that the power supply has a current limit built-in to it, but a cell phone supply will normally operate in a constant voltage mode.

In general, that's not quite the behavior you'd want from an LED driver style of CC supply. In that case, you want the voltage to go to any value needed to force the target current through the load. This permits you to use varying numbers of LEDs in series in your light, for instance. I haven't studied the current flood of "LED driver" chips to see how they're different from other switchmode chips; there may be other simplifying factors that permit a cheaper design and justify having a separate chip (oh, like not needing to provide clean DC, for instance.)

Yes, you have explained it all very well and better then me I' sure. There is a difference between a DC power module that has automatic current limiting protection and a true constant current regulator. I'm just trying to make sure that people understand that ohms law is still applicable here no matter what name a DC power module uses. :wink:

Lefty

be80be:
Let's see a show of hands how many of you have used cel phone charger even like the iphone charger......

Let's see a show of hands how many of you built your first computer in the 1970's with an 8008, 256 bytes of static ram and toggle switches for address and data input.

Let's see a show of hands how many of you know the difference between a triode and a pentode (and have actually built circuits using both).

Lots of people here have forgotten more about electronics than you know. Don't be smug.

(rant off).

Me on all counts, although it was 2650 processor and it was 1976.

Cel phone charger are a constant current supply fixed at 5 volts

That sentence is so wrong unless the charger has the ability to change the impedance of its load. But as others have pointed out this is just sloppy talk.

A constant voltage smps can be easily modified to become a constant current power supply.

The same holds true for a linear power supply.

Wow, I finally got a discussion started. Thank you folks.

A bit more inspection of the device shows that it really can power an Arduino, but there are things that concern me. First, it does have a top voltage limit (open circuit) that falls within the range an Arduino can handle as well as a current limit that, when exceeded, shut it down. It doesn't taper off, it just quits. The killer though is that it is very noisy. I don't have a scope so I can't paint you a picture of the noise, but measuring voltage through various capacitors to block the DC shows a substantial noise or ripple coming out of the supply. Comparing that to an apple two board wall wart ( one of the 1 inch cubes ) the apple one has almost no noise at all and the CC supply has a ton. This noise wouldn't make any difference to an LED, but it could to a more complex device, and it appears the Arduino has enough filtering on board to work. This may not be true of something that I threw together in the garage one evening.

I can't tell you what the chips are that it uses, the markings are blurred or missing on almost every part; the inductor has rings, but I can't tell exactly what colors they are. You've all seen this before.

Regarding the size, the two CC supplies I have are about 3/4" x 1/2" x 1/2", and I got two of them under 5 bucks, so you see why they are appealing as a possibility. Remember, the link I posted was just the first one I hit on ebay, that may have been a mistake since several folk grabbed on that as a negative. The point was that these CC devices are incredibly easy to find and don't cost much, which only a couple of people noticed. Tiny little constant voltage supplies that run off wall power are not anywhere as numerous or cheap. I venture to say that you'll find 10 or so CC led supplies or more for each little CV supply and they'll cost substantially more (not counting getting a usb charger and gutting it).

In answer to the various comments about the apple chargers ( 1" cubes ) the one I have torn apart provides a constant 5.1 volts regardless of the state of the battery in the device it's hooked to. If I hook it to a resistor and force it past the rating of 1A by lowering the resistance, it drops below the 5V level until it reaches some point where it shuts off. That's usually just after the resistor burns a blister on my hand. This little device lists a ton of certifications on the bottom in print so small I have to get the big magnifier out to read it. So, it appears that something in the device with the battery helps the charge cycle along somehow. I don't own a single apple device to look at to see; I probably wouldn't pull an expensive device apart to check either.

See, I don't just discount a possibility simply because it doesn't have the right name; I want to actually see if it will work or not. Bouncing it off other folks helps fill in the gaps that I don't think of or understand (yet). That's why I asked for your thoughts in the very first post.

Your various comments led me to check for noise in the output, so I don't think this particular kind of device is the answer to my need for a tiny power supply, even if it was a compelling possibility.

Get a scope and put it on that sucker, you will be surprised.

Mine (a large collection of apple stuff) shows 100mv or so Vpp, ranging well into Mhz. Linear regulators have practically zero rejection to those ripples.

300mA+/-5%. Now, why on earth would I want to power my Arduino board using something
like this? Can anyone explain why my board would want to have a device trying to drive a
fixed 300mA current into it? Makes no sense, unless maybe the p/s were jiggered after the
fact.

You would never want to power an arduino board with a true constant current power supply. An arduino board with nothing wired to it's output pins or 5V or 3.3 pins draws only around 80ma, and a true CC power source would raise it's output voltage as high as it could trying to force 300ma into the arduino board, which would most likely burn up many components on the board due to the high voltage.

There are some Asian DC constant current LED driver modules designed to power 1 and 3 watt power leds that also have a TTL level input that allows for PWM control for the module effectively allowing a dimming function if one wants to use such a feature, or as a simple on/off control using an arduino digital output signal. However you have to look carefully at the maximum PWM frequency that is allowed for the specific CC drivers that have this extra feature.

Lefty

draythomp: can you post a link to the actual modules that you have purchased? And/or photos or reverse-engineered schematics (without chip numbers, since they're unreadable.)
eBay will let you link to completed auctions for quite a while after they're over...

Also, what voltage to these end up putting out when you connect them to an Arduino?

Here's the closest thing to what I got that is on their website. I actually ordered it through sourcing map and they don't have it up any more. After I got it and played with it a little there was a series of mails back and forth with them in extremely broken English and Chinese (we were both using google translate) to determine the lowest voltage (6V) and the highest (12V) instead of what's on their site. Seems there was a mistake somewhere. I paid practically nothing for them; I'm unclear on the exact price though because there was a few cents adjustment after shipping for currency differences that happened during the day or so it took to get it actually ordered.

http://www.microtek-led.com/ArticleShow.asp?ArticleID=116

So, you see that the voltage, 6-12, was right in the sweet spot that the arduino likes and its regulator would take care of the voltage, and of course, experimental results were different from either the specs or the conversation. I was pretty bummed when I saw how much noise it puts out, but nothing ventured, nothing gained.

Letfy, you're right, no one (well maybe me) would want to put a perfect constant current supply on any device that has some limits on acceptable voltage; it would blow up in a heartbeat. However, there ain't no perfect CC supply, and this ones parameters showed possibilities for a really cheap and easy power supply.

oric, you don't have to do this. I never suggested you do this. Don't do this. For crying out loud, I brought up a subject for discussion and got attacked like a teenager that told her dad she got a tattoo on her face.

I asked for thoughts, not a diatribe.

For crying out loud, I brought up a subject for discussion and got attacked like a teenager that told her dad she got a tattoo on her face.

Good, so you learned a lesson. :smiley:

Lefty

Yep, I ain't never getting no darn tattoo on my face.

Straight out, it is unlikely something like this can be used for a device needing a constant voltage.

The topology, however, is likely a fly-back smps - assuming that it is isolated. That means the led string is floating above a current sampling resistor that controls the output voltage to maintain that constant current (thus constant voltage over the current sampling resistor - likely via a zener + optocoupler).

So all you need to do is to identify that resistor and put in place a divider. That resistor is very easy to identify: it must be between the cathode (led-) and the ground.

After that, you have turned this constant current regulator to a constant voltage regulator.

Sorry folks, I just have to do this:

<flame_on>
Oh, for goodness sake, I asked a question. Yes a question. I started off light-hearted and wanting to have fun with the possibility, then I was told that I was basically an idiot for even bringing up the subject, it wasn't constant current, had a step-down transformer, didn't have certifications, it didn't make sense to attempt such a thing, and that others would never ever do such an unsafe thing. Fine, don't do it.

As for the number of people that thought it was a bad idea, a couple of them actually said something reasonable about why it might be a bad idea. Things like isolation, noise, etc. That's the kind of thing the inexperienced folk should be learning, not that someone THINKS it's a bad idea because they hadn't looked into the possibility.

The tone of this board is sad. I tolerate it because there are a lot of people that have really good ideas and experience that have helped me a lot in the past. However, there are some that just want to belittle the poster as quickly as they can. I'm way too old to be put off by people that use sarcasm as their initial response because I know there are others that will contribute really good ideas.

So, thank you very much to the folks that came up with good suggestions and novel ideas; the rest of you,...

Have a nice day.
</flame_on>

I just have to do this:

It is a process thing: you will learn to talk to those whom you can talk to and don't talk to those whom you cannot talk to.

http://www.microtek-led.com/ArticleShow.asp?ArticleID=116

It looks very much like the "fake" Apple power supplies (one or two transistor flyback), though not misleadingly labeled and bare for anyone to look at the details (and nicely priced, and not needing "decasing.")

It doesn't look like they'd come much closer to meeting safety specs. I love the "fuse" implemented as a squiggly little trace on the PCB (clever and probably effective, but probably not "legal.")

Here is one out of a broke led light bulb I bet it's made right off the data sheet
the chip on this one is LNK562-564. It was rated at 300mA

The circuit shown in Figure 5 is a typical implementation of
a 6 V, 330 mA, constant voltage, constant current (CV/CC)
output power supply.

http://www.powerint.com/sites/default/files/product-docs/linklp_family_datasheet.pdf

Led Suppy.jpg

If you prowl around that site a bit there's example after example of flyback power supplies and how they can be made. I gather that, given their chips, it all boils down to the flyback transformer design and choice of resistors that are used. They even offer a transformer prototyping service that can build a transformer for you based on their application notes. I don't want to build one, I want to use one.

Now, all I have to do is find a tiny, cheap, constant voltage power supply that isn't already stuck inside a wall wart to experiment with.

See, I have this dream of a device that houses an XBee, a minimal arduino and a couple of screw terminals all built inside a wall wart like this: GS-2415 | Wall-plug Electronics Plastic Enclosure . That way I can deploy a sensor and read its data remotely. Things like the inside temperature of the chicken coup, voltage of the generator battery, open-closed state of the North gate, become easily done. Add a few more terminals and a couple of tiny relays and you have a sprinkler system, a remote gate opener for the UPS guy, an auto drain for the horse watering trough (mosquito prevention). With a simple sealed float switch I can remotely tell when the septic tank starts to rise so I can go out and clean the output filter and hose out the input. Yes, the chicken coup has power, this is Arizona, they even have a swamp cooler in the summer.

I have practical uses for this kind of thing.

find a tiny, cheap, constant voltage power supply that isn't already stuck inside a wall wart to experiment with.

The simplest would be your cell phone / usb chargers. Buy one and break it apart.

LNK562-564.

That is one of those gated oscillator smps types: it turns itself on/off between 1.69v - 0.8v (on the feedback pin), making it suitable for low-power applications but largely not usable for a device needing a constant voltage.

The grandfather of those designs are the Linear LT1107.