Transformer Alternatives

Hello.

I am building a circuit for controlling 36V high power LEDs. They use about 3 amps. I have decided to use a current sensing controller IC. Basically, it turns on a mosfet until the current reaches 3 amps, and then turns the mosfet off for a set time period. Then it turns the mosfet on again until the current reaches 3 amps, then shuts off again. This cycle repeats, creating a ripple in the output current.

It uses an inductor in series with the LEDs, and a capacitor in parallel. This allows the LEDs to be powered while the mosfet is off, and limits the current and voltage from increasing too quickly while the mosfet is on.

This is all well and dandy… but the thing would work more efficiently if I could use a supply voltage of something like 60 volts instead of 120V. This would be easily accomplished by using a simple transformer with twice as many turns on the primary side as it has on the secondary.

I don’t want to use a transformer… Because I want to design this circuit to be lighter and smaller, and possibly more efficient than what it would be with a transformer.

Anyone have any suggestions how I could cut the supply voltage in half without using a transformer?? Possibly by using mosfets, or triacs, or something else controlled by a microcontroller like the arduino?

Thanks.

attached is a circuit diagram and the datasheet of the controller.

NCL30105-D current mode controller.pdf (613 KB)

This is all well and dandy.. but the thing would work more efficiently if I could use a supply voltage of something like 60 volts instead of 120V.

Not necessarily... Switching supplies, either constant-voltage or constant-current, (such as you have described) can be nearly 100% efficient.

A step-down switching supply will have less voltage coming-out than going-in, and it will have more current coming-out than going-in.

A transformer is also nearly 100% efficient, but if the transistor is followed by an analog regulator, the overall design will be less efficient.

It's also possible that your switching supply won't work from 60VAC.

P.S. Most constant current power supplies that you buy do have a transformer to isolate the power-line from the LED power. But it's a special high frequency transformer, and the high-frequency means the transformer can be smaller & lighter... The AC would be rectified & filtered, then there is an (high frequency) oscillator that's part of the switching circuit. (I assume your power supply is also switching much higher than line-frequency.)

Your circuit is dangerous if the LEDs & LED low-voltage power is not safely insulated. There is a direct connection back to the line voltage through the components.

P.P.S. If your supply is putting-out 36V @ 3A, that's about 100 Watts. Since your circuit is approximately 100% efficient, 100W from 60V is about 1.7 Amps and that's all you'd need from your transformer. And, you'd need about 1A from "the wall".

Hmm… Thanks for that input Doug. I appreciate it. The fact that the LEDs aren’t isolated from the supply was a concern I had and I most definitely will have to find some way to isolate it.

Yes, the LEDs I have are rated at 100w.
Yes, the frequency of the mosfet switching would be much greater than line frequency… upwards of 100Khz, most likely something like 300Khz.

I am kind of confused about the high frequency transformer you are talking about… Where exactly would you place it in the circuit? If it is between the rectifier and the freewheeling diode, then the transformer would be exposed to a relatively smooth DC voltage on the primary side and the secondary side would be pulsed on and off at the frequency of the mosfet switching… but it would always be above 0v, never swinging into the negative…

Take a look at the modified schematic attached… Does this make sense? Would the transformer work in this way? Would it still be able to be reduced in size because of the high frequency on the secondary side, even tho its almost a constant DC on the primary?

diagram ver2.png

Thanks

The supply part of that diagram won't work. Forget about making your own supply. It's a lot better/safer to buy that from e.g. ebay. Search for a 48volt/2.5A (minimum) metal frame supply. ~48volt is the most efficient voltage with 36volt LEDs and this type of switching IC. So the whole top/left part of your schematic can go. Stick to Fig.1 in the datasheet, but add a 100uF/63volt low ESR cap from Vin to ground near the LED driver. Leo..

This could power three of those 100w LED circuits. 24v/7.5A

On the basis that you appear to lack understanding on how transformers work, and for your own (and others’) personal safety may I suggest you opt for the prudent route and use commercially available equipment such as the attached.
http://www.meanwell.co.uk/led-power-supplies/clg-100-36.html
If you feel this is too large for your requirement then you need to ask yourself why it is such a size.

The heatsink for a 100W LED would be twice that size, and might also need a fan. People experimenting with 100w LEDs usually use an old computer CPU heatsink/fan.

Single, or 2-6 string LEDs are way easier to start experimenting with. Whole driver boards are a only a few dollars. The 12 or 24volt supply they need are 1/3 of the price for the same power.

Why a 100W LED. It's a LOT of concentrated light in one place, only suitable as a work light, or outdoors. Leo..

wawa… Jackrae…

It’s all well and good to say that the schematic I posted won’t work… But if you can see that there is an error in the operation of it, please do us all a favor and describe what the issue is that you have with it.

Would you put the transformer in a different location? Would you add a ground connection on the primary side of the transformer aswell? What exactly is the issue that you are seeing?

Wawa… You say “The supply part of that diagram won’t work.”
Can you explain why it wouldn’t? I’ll attach a similar circuit that was professionally designed and is part of the IC datasheet LM3447. The supply part of this circuit is basically the same as the one I posted…

Why 100 watts? Why not? That’s just the tip of the iceburg… This will be the building block of a larger 1500w lighting system… I’ll be adding in a power factor correction and harmonic distortion minimizing block so that it meets all the national energy efficiency standards, and hopefully it will end up being above the 85% efficiency mark. That seems to be the standard goal for power supply designers.

There’s no way I will be buying a power supply… If I wanted to do that I would already have bought one and wouldn’t be posting this question in the first place! I am definitely building it.

lm3447.png

here
3v 300mA x 10s x 1p.png

This is a pretty standard smps current controlled buck converter with an emi filter.

The efficiency of this is 88.4% at 85VAC rms and about 83.6% at 265VAC rms. I’d be supplying it with 170vac rms… which would be around 85% efficient… but if I could put a transformer in there somewhere to get my 170VAC rms supply down closer to the 85Vrms level, it will be much more efficient… That’s the reason I wanna lower the voltage in the first place… So ya lets here the suggested locations / applications you guys think would work.

The schematic you posted now is completely different from the first one. The inductors in the first post are drawn as a transformer with a rectifier in front of it. In the second drawing, they are radio interference chokes.

I think you are trying to play with us.

Who would post a schematic with a 1ohm/486mA fuse and 62.5nF/1mohm polarised capacitors.

Did you forget to take your pills?

And apart from that anyone who thinks it's OK to ground one side of the non-polarised mains supply (even via a bridge rectifier) clearly isn't aware of the dangers. The transformer in your second last post is a high frequency unit which acts as a mains to LED isolation device, there is no transformer in your last post which means the LEDs are directly connected to the mains - a potentially lethal set up.

We really are trying to guide you, not knock you down.

I had a look at this guy's history. He seems to know more about SMPSs than most of us. Leo..

Wawa: The schematic you posted now is completely different from the first one. The inductors in the first post are drawn as a transformer with a rectifier in front of it.

Ya.. I was trying to show other LED driver circuits that have the same power supply setup... AC mains followed by rectifier followed by transformer.. which is what the second schematic shows.. it is a flyback topology of smps..

In the second drawing, they are radio interference chokes.

I think you are trying to play with us.

Not playing... it is a flyback topology. very common.

Who would post a schematic with a 1ohm/486mA fuse and 62.5nF/1mohm polarised capacitors.

Did you forget to take your pills?

I guess the guys over at fairchild semiconductors forgot their pills.. because that's exactly where I got that schematic from. If you don't like their values, take it up with their engineers... the 62.5nF capacitor is part of the EMI prefilter section... not even part of the converter..

Other than the prefilter section, the circuit is very similar to the one I plan to build, just with smaller values.. it's a basic current controlled buck converter.

jackrae: And apart from that anyone who thinks it's OK to ground one side of the non-polarised mains supply (even via a bridge rectifier) clearly isn't aware of the dangers. The transformer in your second last post is a high frequency unit which acts as a mains to LED isolation device, there is no transformer in your last post which means the LEDs are directly connected to the mains - a potentially lethal set up.

We really are trying to guide you, not knock you down.

Dude.. if you are talking about the second schematic with the rectifier negative to ground... That schematic is directly from the datasheet for the IC LM3447... Thats the layout that the manufacturer of that IC recomends! It's texas instruments by the way.. I kinda doubt they would mess that up... The transformer in the second last post is a flyback converter circuit..

Here’s a couple more examples of the same thing…

I didn’t draw these… These come direct from the datasheets of these IC’s! I think these ones are both resonant converters… similar to flybacks.

tps92010.png

So without inspecting all these designs too critically, I gather your purpose is to use a switchmode circuit to deliver a constant-current supply to your LED.

In one stage.

Because that is exactly how many if not most of the supplies commercially used in such light fittings work, and the only sensible way to do it.

Note - one stage, not a voltage stage and a current stage.

Note - no transformer, either mains or switchmode frequency. The LED is operating at mains voltage and must be adequately insulated and isolated from the user, contained in a grounded or double-insulated housing.

Hi, I would suggest that the first transformerless circuits are designed as LED drivers for stand alone LED lights, the lights and driver all in one sealed package. Hence the NCL IC only in surface mount form. Such as a flood light with an array of LEDs.

They are not intended as a bench LED supply or a stand alone supply with external wiring from the driver to the LEDs.

Tom..... :)

Actually.. Now I remember the original reason I wanted to put a transformer in there.. Kinda forgot the real issue I was having..

I'll be implementing a power factor correction boost stage infront of the buck converter stage.... The problem is that the output voltage of the PFC boost stage is about 400VDC. I don't need anything near that voltage input at the buck converter stage... The higher the input voltage to the buck converter, the worse the efficiency.. so if anyone can suggest a method to knock that back down to like 100VDC without too many losses, I'd love to hear it.

So what's the deal.. transformers don't work with DC? I can't believe I didn't know this by now! I shoulda been able to figure it out.. only a change of current over time in the primary can induce a changing magnetic field in the core, and only a changing magnetic field over time can induce a current in the secondary. Wish you guys woulda just said that instead of just saying "u don't understand transformers. buy a supply cuz ur gonna kill somebody."

Thing is.. none of us are experts in every single facet of electronics / electrical physics. I've got a huge wealth of knowledge and experience with all kinds of circuits, schematics, sensors, controllers, etc. etc. etc. And I'm sure you guys do too. There are gonna be things here and there that I might not have picked up along the way, and that's why there are forums like this.. So we can discuss things and learn from each other to fill in those gaps of knowledge that we all have.

I find it really disappointing that so many people on electronics boards tend to highlight all the problems without giving any solutions.

I mean is it really helpful to tell a person "hmm.. ur problem is you don't quite understand how transformers work. Just give up, because they are so complicated I can't even think of explaining it to you."

Hi,
Can you post a spec sheet on the 36V 3Amp LED array that you have, as well as a picture please.

Tom… :slight_smile:

Hi,

Actually.. Now I remember the original reason I wanted to put a transformer in there.. Kinda forgot the real issue I was having..

I'll be implementing a power factor correction boost stage infront of the buck converter stage.... The problem is that the output voltage of the PFC boost stage is about 400VDC. I don't need anything near that voltage input at the buck converter stage... The higher the input voltage to the buck converter, the worse the efficiency.. so if anyone can suggest a method to knock that back down to like 100VDC without too many losses, I'd love to hear it.

PFC Whyy???????

Can you draw a block diagram of what you are trying to do to get from mains power AC to a DC supply for your LEDs.

Tom...... :)

PFC… power factor correction and THD reduction (total harmonic distortion)… It’s a requirement for commercial electronics of this size to have a power factor above a certain point and THD below a certain % if they are to be connected to a grid power source (which it will be… 120VAC). This is because of the larger number of digital circuits being used by everyone that implement things like switch mode power supplies… If no PFC is used then there can be bad harmonics, interference, noise, etc. in the rest of the line…

In the case of a bulk capacitor used after a rectifier, without PFC, the input current will be mostly flat, except for a brief period of each cycle where the capacitor is charging. This can even cause a ‘sag’ in the line voltage waveform at this point.

Here’s an example of what I’m talking about:
pfc.png

Here’s a block diagram that’s pretty similar to what I intend on doing… It’s not exactly what I’m doing but it’s close. It’ll give you an idea of what’s going on.

pfcbuck.png

So, if you want a different AC supply voltage because of an alleged difference between 83% and 88% efficiency, has it occured to you that having an extra transformer is likely to waste at least 5% efficiency anyway ?