DC/DC step-down converters: what trade-offs prevent lower minimal input voltages?

Question

I'm currently looking into isolated step-down DC/DC converters, to convert from high voltages (a few hundreds volts) to 24 V.

I'm surprised that the minimum input voltage is often quite high (like 200 V for a 300 V nominal input voltage (example)).

When the input voltage is close to the output voltage, I understand that keeping a few volts of difference makes things far easier.

But I'm wondering : what tradeoffs can explain why most DC/DC converters with 300 V nominal input have a lower input limit >150 V?

I'm referring to switching power supplies, in the 500-2000 W range. For the topology, I haven't looked at those in detail yet for this kind of power (and manufacturers often don't specify it), so whatever classical topology in this power range is interesting.

Answer 1

I'm currently looking into isolated step-down DC/DC converters, to convert from high voltages (a few hundreds volts) to 24 V.

Apart from some exotic exceptions, all these DC-DC converters use transformers that provide both isolation and a good efficiency at high step-down ratio. There are many different topologies: flyback, push-pull, resonant... In the case of flyback, the transformer is used as coupled inductor.

But there's always a primary winding, a secondary winding, and therefore there is a turns ratio between primary and secondary. This turns ratio will usually be optimized for best efficiency at the nominal input and output voltage ratios.

Therefore, the further away you get from this nominal ratio, the worse it gets. In fact, it's already pretty good to get a 1:2 input voltage range with a transformer that has a fixed turns ratio.

There are other factors too: at constant output power, input power is inverse proportional to input voltage. As input voltage gets lower, more current is needed to provide the same power. Extending the input voltage range down will require switching devices capable of higher current, larger dies, with more capacitance, and more switching losses. Also it needs thicker wire in the primary, which costs more and makes the transformer bulkier. If the application does not require it, there's no reason to incur the extra costs and loss of efficiency.

The usual "universal wall wart" supplies with 85-240V AC input generally use flyback topologies, which offer a wider voltage range at a low cost, but at the expense of higher losses. It's okay for a low power converter, but at 500W there are much better options.

For high power AC-DC converters with "universal input" you'll usually get a boost power factor correction stage or a capacitive voltage doubler, which compresses the allowed input voltage range into a much narrower DC bus voltage range, followed by a resonant converter which allows very high efficiency. So the actual isolated DC-DC converter in these supplies does not have to handle the full input range.

When the input voltage is close to the output voltage, I understand that keeping a few volts of difference makes things far easier.

This would apply to a buck converter, which is a non-isolated topology using an inductor. It's a completely different topology from a transformer-isolated converter.

Answer 2

When the input voltage is close to the output voltage, I understand that keeping a few volts of difference makes things far easier.

If you are talking about switching DC/DC step down converters then that is incorrect.

The reason why it is incorrect is because as the input voltage falls below a certain point, sufficient energy becomes harder to put into the inductor for a given switching frequency and load.

You can use more complex switching circuits that are called buck-boost converters but, these generally aren't used internally in those small encapsulated cheap converters you get from Recom, Traco, XP and Murata (to name-drop a few).

To use buck-boost conversion techniques in isolated converters can be done. Typically a fly-back converter can do this but, as power levels reach over around 200 watts, they are becoming quite inefficient. The converter you liked in your question is rated at 1200 watts so, flyback is out of the question.

If you are thinking of linear converters then it is an easier job when there are only a few volts of difference.

what trade-offs can explain why most DC/DC converters with 300V nominal input have a lower input limit >150V

Basically, as the input voltage level drops to a certain point, the device won't be able to guarantee full output power for the price people expect.

Answer 3

The duty cycle should be held in 10-90% range, otherwise converter is not efficient because of increased switching losses.

But the combination of wide input voltage range together with load regulation cannot fullfill the 10-90% rule.

Note: With frequency controll it can be overcommed a little.