Hi all,

I'd like to install some LED outdoor lights which work on 12v AC or DC. Depending on how many I decide to install I may exceed 50w. I have lots of 12v AC electronic transformers previously used for old MR16 halogen downlights, rated at 50w each. Rather than buy a dedicated transformer I'd rather buy an IP65 box and use these. I'm aware that I can't simply put these in parallel as the phase will not be in sync, please correct me if I'm wrong. My intent was to run these through a rectifier and filter cap of their own to produce 12v DC. I can then put as many of these as I require in parallel to add capacity if required. My thought was as illustrated in the diagram. I've intentionally double fed this from both ends to form a loop to reduce the potential for voltage drop.

I hope this makes sense! If it does please suggest rough values for caps and rectifier assuming 12v DC 60w draw.

 

If the secondary voltage is 12 Vrms, the peak voltage across the capacitors will be above 15 V.

The problem with connecting three secondaries in direct parallel is not the phases; these can be checked with a scope. The potential problem is that one unit with a lower secondary voltage (for whatever reason) will appear as a load on the other two, causing unnecessary extra heating in all three. In a home project, I'm combining the 28 Vac control transformers of three garage door openers as a triple-redundant power source for a control system. Not nearly the same power level, but the same potential problem.

ak

 

Thanks for the response. If I make the following assumptions does this make it OK? I'll need to do some testing to confirm this before I wire it up.

1. The transformers are the same make and model so voltage variance should be less than 0.5v.
2. The 15v - 17v peak voltage is with no load. Once load is added this should be reduced substantially.
3. Since the lights can run off AC or DC they clearly have a driver in each. This driver would be tolerant for voltages up to 15v even though the lights are rated as 12v.
4. The bridge rectifier serves two purposes. 1) AC to DC 2) to prevent any current entering back in and allowing the transformer to act as a load. The diodes will only allow current to flow out of each.

Thanks again. I'm happy to be proven wrong.

 

2. No, or at least not always. The 15+ V comes from the difference between a transformer secondary's voltage rating, which is RMS, and its peak value which is what is stored in the capacitor thanks to the diode(s). Depending on the quality of the transformer design and materials, the unloaded AC voltage can vary form a lot to very little. Typically, the output voltage is at its rated value +/-10% when driving its rated load current.

The DC capacitor voltage will have an AC ripple component, based on the capacitor size and the load current. This can make for a variety of measurements, depending on the meter's voltage input circuit.

ak

 

The Peak-Voltage usually will remain,
but only as a Peak in the Ripple of the not so smooth DC Voltage.

You may not be aware of how much Filtering will be required.
Your LED Regulators may, or may not, play-nice when fed a ton of Ripple on top of 12-Volts.

I would recommend a "Choke-Input" Filter such as the one pictured below.
The Wave-Form is Computer simulated with the component values shown, with a 50-Watt Load.
A Choke-Input-Filter is semi-self-regulating, and will provide a more balanced Load
to the Transformers, increasing efficiency, and reducing waste-heat generation.

You must use a Metal Box for adequate Heat-Dissipation from the Transformers and Bridges.

5-Amp rated, 1mh, Toroidal-Choke .......
3013-L201204-102MHF-ND ~$28.80 at DigiKey Electronics

10- 1000uf, 25V, 10,000hr, Electrolytic Capacitors ........
1189-1245-ND ~$0.82 each at DigiKey Electronics

50V, 6-Amp, Bridge-Rectifier .......
VS-KBPC6005PBFGI-ND ~$2.70 each
.
.
.

 

Thanks LowQCab

You may not be aware of how much Filtering will be required.
Your LED Regulators may, or may not, play-nice when fed a ton of Ripple on top of 12-Volts.

Since the lights can handle both AC or DC I presume they'll be fine with ripple.

I have a bag full of 470uF 35v laying around so was planning on placing 2 in parallel across the output of each bridge.

It sounds like my design isn't fundamentally flawed.

I think my plan will be;
Build one times Transformer -> rectifier -> capacitor -> LED
Smoke test
Build another Transformer -> rectifier -> capacitor -> LED. Input and output in parallel
Smoke test
Try a choke if things don't work??
Load up with more LED lights
May build a third if I need the capacity.

 

1. As the lamps are for AC or DC, they will be quite happy with unsmoothed DC.
2. As these are high-frequency transformers, I wouldn't recommend a standard bridge rectifier, use schottky diodes.
3. If your 50W electronic transformers are identical (same type same manufacturer), then there should be no problem putting them in parallel. I would recommend a choke in series with each one, the rectifier will prevent any power being fed back.

 

A further thought - will an "electronic" transformer really take kindly to the pulse currents associated with smoothing capacitor, especially if it's the crude self-oscillating circuit?

 

Sorry .......
I missed the "Electronic" part of "Electronic-Transformers"........
That throws a potential wrench into the gears.
Some "Electronic-Transformers" are VERY FINICKY and may do weird stuff when paralleled.
You have NO IDEA of what comes out of those Transformers without an Oscilloscope.
In light of this info,
I would now say that EACH Transformer must have its own "Choke-Input" Filter,
BEFORE connecting them all together.
This will also insure that any high-frequency trash / noise will be filtered-out.
This type of "Electronic-Transformer" is designed to be just good enough to
provide power to a Resistive - Incandescent-Load, which doesn't care about
anything except "Average-Power".
ALL OTHER BETS ARE OFF.

It may be better to use 3 separate Lighting Circuits.

Be careful, those "Transformers" may not be "Isolated" from the Mains Power.
.
.
.

 

As these are high-frequency transformers, I wouldn't recommend a standard bridge rectifier, use schottky diodes.

If I go down this path then it's not worth it. Shottky diodes seem expensive. Are you suggesting that the basic silicon bridge/diodes won't be able to handle the high speed switching?
Any estimate how fast the switching would be on an electronic downlight transformer? I don't own a scope.
What's the main disadvantages of using basic silicon bridge/diodes in this circuit?

If I don't use capacitors and instead use chokes in series would this keep the voltage closer to RMS rather than a capacitor holding it at peak?

Also, for practicality I can't use individual lighting circuits. I really want to wire all lights in a loop to prevent voltage drop and since all lights follow their own natural loop layout it would be best to keep it as a single loop. I can't fit that much wire through my conduits.

Thanks all.

 

One "real" Transformer would make things much easier .........
237-2038-ND ~$59.oo

 

Maybe I'll get my hands on a few old iron core transformers for halogen lights. At least I know that the output of these is 50Hz. These are throw away items these days.

I checked eBay and it seems I can get 10 x 10A 45v schottky diodes for a few dollars. If I go back to using a 50Hz secondary then I may not need the schottky and can stick with any old bridge rectifier.

 

A standard Bridge will be fine in any case.
The difference between a "Standard" Bridge, and a "Schottky" Bridge is the
amount of heat dissipated by the Bridge.

The "Schottky" Bridge is simply a bit more efficient.

A "Choke-Input" Power-Supply will help to reduce Heat-Dissipation in the Bridge, and the Transformer.

In either case, the Bridge will need some kind of Heat-Sink.
A Standard Bridge will probably dissipate around ~5-Watts in this scenario.

.
.
.

 

Hi guys,

The one thing that I learned here is that the output frequency from these basic electronic downlight transformers is extremely high, way higher than I anticipated. I presumed it was 50Hz but it turns out that they range between 40kHz - 80kHz without load and increase frequency when load is applied. This makes them completely unsuitable for my needs. No wonder why my good quality multimeter didn't pick up any current on AC or DC. It must have been switching too fast for it to recognise.

I ended up picking up a couple of old iron core, wire wound ATCO transformers which I can easily turn into DC with rectifiers, chokes and caps.

Hopefully this lesson will teach someone else trying to go down this same path.

Ref: http://www.ledbenchmark.com/faq/Transformers-Output-and-Compatibility.html

Thanks for all your input and assistance.