led drivers typically aren't specified in voltages. they may list a maximum output voltage, and an input voltage. the metric you should be concerned about with a commercial LED driver is the output current. that is because leds are current-driven devices, not voltage driven.
the first request sounds like any ordinary 5v power supply. I have a bunch of the, they claim to be universal 90-260vac @ 47-70 hz, providing 5vdc @ 1a (most of mine are 2-3a actually)
any good surplus vendor should have those.
the other request is a bit harder, since you're looking for a buck-boost or sepic converter :/
Yes, I know LED driver should be constant current but I need two difference output output1= 14 Vdc and output2 = 4 Vdc all are 1A. If I use the power resistor for 4 Vdc LEDs, the Resistor getting very hot. Do you have any idea
for using 1 driver for two difference voltage?
How many LEDs are you powering and what colour are they?
Have you contacted the LED manufacturer?
They might be able to suggest power supplies for those voltage ratings.
If you want to go for the DIY route then the best option is to look at a SEPIC converter for the DC-DC option and use a small mains transformer with a switching regulator on the output for the mains version. It is possible to build switching regulators that run directly from the mains but it isn't recommended unless you have lots of experiance.
The reason I'm asking for difference voltage is PWM. I build PWM by LM317T.
It's very hot if I use LED driver V-output = 18 V. Do you have any circuit to build PWM to control The LEDs (current output for LEDs should be 1000mA).
There are R-G-B-G LEDs. Totally is 15W/14V/1A. All-in-one Package (one big LED have four dices inside). If I change current or Voltage (by PWM) for each individual LED I'll have difference color. If I turn on all four dice at the same current and same time I'll have White color.
That looks expensive but it's probably worth it as it will do exactly what you want and more.
95% is its peak efficiency which occurs when the input-output differential is very low; it's more efficient to convert 5V to 3.5V than it is to convert 32V to 3.5V.
I don't like the idea of connected LEDs in parallel without balancing resistors (see datasheet) but there again I suppose those LEDs must be so well made that they have exactly the same series resistance and voltage drop.
that diagram is right from Philips - they condone wiring Luxeons in that manner, provided you're using a constant current power supply (they advocate their Xitanium driver of course).