Need advice for 74VDC power supply

[mister_e]

yes, and make sure you have some heat-sink around. Some switching regulator could help. have a look to LM5010

[Pic_User]

yes, and make sure you have some heat-sink around. Some switching regulator could help. have a look to LM5010

I agree.
The voltages involved can quickly rack-up Watts.
He should keep in mind the heat generated is from the voltage drop within the device, and the amperage through it. With 90 Volts input and 12 Volts output you have considerable voltage differential. The data sheet claims assume you have dealt with the heat, and removed it.

Second pitfall:
Use an input (bypass) capacitor.

Third pitfall:
Mind the "Minimum output current to maintain regulation".

As you said the switching regulator could help. More efficient and a lot less heat to dump. Little more engineering though.

-Adam-

[Christopher4187]

Thanks for the replies everyone. I have looked at the datasheet and have some further questions for electronic wizards.

1. The minimum current is not a problem. I will just put an LED to indicate power and that should take care of that problem.

2. The data sheet says, "An input bypass capacitor is necessary
only when the regulator is situated far from the input filter." I will put one anyway just incase but would a .1uF cap be sufficient?

3. The data sheet also says, "An output capacitor, although not required, improves transient response and protection from instantaneous output short circuits. Excellent ripple rejection can be achieved without a bypass capacitor at the adjustment terminal." Again, would a .1uF cap be sufficient?

4. I don't understand the capacitance issue. The data sheet says, "The TL783 regulator is stable without bypass capacitors; however, any regulator becomes unstable with certain values of output capacitance if an input capacitor is not used." In another section the data sheet says, "when an instantaneous short circuit is applied to the output, both ringing and excessive fall times can result. A tantalum or aluminum electrolytic bypass capacitor is recommended to eliminate this problem." Is the electrolytic acting like a small battery? How can I measure the output capacitance? Is there a cap that will work for every value?

Thanks for the help, this chip seems to be exactly what I need.

Chris

[mister_e]

usually 10uF tantalum + 0.1uF ceramic at the input and at the output cover almost everything.

[Christopher4187]

For my information, what does the 10uF tant do? What does the .1uF ceramic do? I am assuming the material has a lot to do with the type of operation it is used for.

[mister_e]

to make a short story.. both do the almost the same job (filtering, noise supression, ... ,.. ) but on different frequency range. Of course the 10uF react slower.

that was the short story. Now for a better explanation you could Google for decoupling Capacitor.
http://en.wikipedia.org/wiki/Decoupling_capacitor
http://hsi.web.cern.ch/hsi/s-link/de...c/decouple.pdf

http://www.satcure-focus.com/tutor/page6.htm

[Pic_User]

Hi Chris,

The capacitors Mister E mentioned are important for input filtering and noise suppression.

These active voltage regulators look simple, like a power transistor. But if you look at the schematic, of the IC, inside the package, they are complicated. There are about 20 transistors inside. The circuit is a high speed error amplifier, that responds quickly to anything that throws it off equilibrium. When the load changes, the error amplifier corrects to balance the output back to normal. Any noise or variations are also reacted to.

Because of the high gain of the regulator, sometimes the correction overshoots.
If the AC filter capacitor is far enough (few inches) from the regulation device, the circuit has enough inductance in the leads to cause a small time delay. The inductive delay and the high gain of the regulator amplifier, create oscillations.

I have seen regulators rated at 1 ampere output get smoking hot and shut down with a minimum load. Checking with a fast oscilloscope, shows a strong AC waveform at 20 MHz. Adding a 10 microfarad Tantalum capacitor near the input pin cured the problem. That’s how I learned input caps are needed on these guys.

Also, do the math on the Watts, you will need a huge heat sink as Mister E said. Depends on your final circuit but looks like 40 Watts or so!

-Adam-