I've been using these Chinese power supplies for years. They are great! This is the older model that has been pretty much the standard for powering 100w LEDs (3A @ 33v).

And here is a new model that's I've never seen before. And it is even cheaper than the old model, as it should be just by looking at it. The board is smaller, the heatsink is simpler and there is only 1 transformer now!

What does it look like is going on with the new design? How did they get away with just 1 transformer now?

 

A switcher only needs one transformer to reduce (or boost) the line voltage, but that is typically for a non-isolated output. You can manage to isolate the output with a single transformer but the regulation tends to wander as you are left with things such as opto isolators to give the voltage feedback.

 

The other trend worth mentioning is the move to higher frequencies. The aircraft guys long ago realized that power transformers that work at 400 Hz. are smaller and lighter than those required at 60 Hz. Similarly SMPS switching frequencies have increased by nearly 2 orders of magnitude in the last 20 years or so.

 

You can manage to isolate the output with a single transformer but the regulation tends to wander as you are left with things such as opto isolators to give the voltage feedback.

Perhaps old design was isolated? And new cheaper one is not?

Looking at the new one I think I see an opto-isolator right in the middle of the single transformer. So it might be just as you suggest, isolated by the transformer with opto feedback. Is there something inherently bad about this design?

 

Both supplies shown are fully isolated. And yes, that is an opto coupler in the 2nd photo. Without seeing the bottom side of the pc board it is hard to say, but the supply in the first photo might be power factor corrected. If so, the small magnetic device to the right is an inductor, not a transformer. It is part of a non-isolated boost converter that stores energy in the large cap. That drives the 2nd magnetic device, the main transformer.

ak

 

I notice there is a toroid on the second one that I don't see a matching part for on the first. I don't suppose that's taking the place of one of the transformers?

 

I notice there is a toroid on the second one that I don't see a matching part for on the first. I don't suppose that's taking the place of one of the transformers?

On the high side next to the bridge rectifier? Seems like that just a more budget version of an EMI choke.

 

the first "transfo0rmer is probably a choke to keep noise off the ac line. the second one uses a toroid for the same thing.
the biggest change in switchers I have seen recently is the use of a double switcher, a switcher powered off the ac line that supplies power for another switcher that provides the output. on a couple of different brands, it is so they can use different output modules for different output voltages, and others are single output. I havnt found a reason for that yet.

 

You might be describing a modular power supply, which has a common front end followed by a frame that holds interchangeable modules for various output voltages. Depending on the manufacturer, what is distributed to the output modules is either high voltage DC around 360V - 400V, or switched DC at the same level. Either way, the modularity enables a variety of load conditions to be met with a standard set of components. Compared to a dedicated switcher the modular ones cost more per watt and take up more room, but efficiency is about the same. That two stage is now standard in all switchers above a couple of hundred watts.

The reason is power factor correction (PFC), or harmonic current correction. Although the first thing in the door in a "normal" switcher (after the bridge rectifier) is a large capacitor, the current spike that input causes looks inductive in that the input current comes off of zero very near the point when the input voltage is at its peak. So while the average and rms currents both are whatever it takes for a particular output power, the peak input current can be 10x, 20x, 50x that amount. This causes great grief in the infrastructure that is delivering power to the device. For a connector or terminal block or breaker panel *or generator*,delivering 10A rms and 14A peak is BIG difference compared to the rms current staying at 10 A but the peak current goes to 100 A.

The most common PFC technique has a boost converter after the input bridge, followed by the holdup cap, followed by the "real" switcher circuit and transformer. The boost converter has two control loops, one to maintain a peak voltage level at the cap, and one to shape the input current so it follows the input voltage. In modular supplies the dividing line between the front end and the modules usually is after the big capacitor. The rules and regulations regarding when PFC is required change about every two years, with each change lowering the power level that requires PFC. Most of this is being driven by European regulatory agencies because their average power quality is so much worse than in the US.

In the first photo, the dual-winding device behind the bridge rectifier is a common mode choke that is part of the EMI filter that keeps switching current noise off of the power line. In the second photo the torroid to the left is the same thing. For the same performance, a torroid usually is smaller and more expensive than the bobbin style choke. The second photo supply is not power factor corrected.

ak

 

common mode choke that is part of the EMI filter that keeps switching current noise off of the power line

Doesn't really help it. Those switchers knock off local FM stations for me. A few more ferrite chokes on the power cable feeding it help though.

The second photo supply is not power factor corrected.

I ordered one of the new power supplies, and I have a Kill-A-Watt that shows PF. Will post results.

 

As both supplies are open frame types, they radiate all over, not just from the power line.

ak

 

So I tested the "new" design power supply, and sure enough power factor was 0.61. The "old" design bounced between 0.99 and 1.00 power factor. Both about 86-88% efficient.

Any ideas why the Chinese manufacturer would include power factor correction at an expense of roughly 20% higher price? Maybe so it's exportable to the EU market? Do they require power factor correction? This is a small consumer based "appliance". No regulations in US as far as I know.

 

No one clear answer. The minimum power level that requires PFC varies by country, industry segment, and application.

ak