overvoltage in my DC-DC

[theChris]

ok, well i decided to make a simple power supply for signal processors in my car, I just hooked up the 4 components, which are 1 1000uF 16v filter cap, 1 5v 3A regulator, 1 470uF 10v filter cap, and a 5v to +-15v DC-DC converter module. I however get +-20v. I also don't have any form of overvoltage protection for the 16v cap, and i know it is possible to get more than 16v in a surge from a car battery. the voltage after the regulator is about 7v if that helps any. I have tl082cp op amps which specify a +-15vDC power supply. What can i do to fix, or diagnose this problem better, or should i do anything?

well i would start by replacing the caps with 25 or 35V ones cause giving yourself a 1 volt leeway (or with most caps being -+ 20% ) a lower rating than your output could cause some cap top poppin. I can't see why it would be acting as you say without looking at it for real but i would reccomend hooking a 15V regulator on the output of your DC-DC if infact it runs at 20 volts with no problems. So it would be:

12V car battery--->25 or 35V filter cap(say 1000Uf)---> 5V Regulator----> dc-dc converter----> 15V regulator---->output smoothing cap (agian 25 or 35V,1000Uf)----> smooth +- 15V out.

[theChris]

yeah, the 16v cap worried me. I had kinda wanted to avoid using + and - 15v regulators because they would just eat more of the output power, which is limited, but should be high enough.

[jackinnj]

especially when it's in the thousands of volts:

http://www.tech-diy.com/RD100R.htm

[nemestra]

Hi theChris,

First thing - a 5V 3A regulator should give 5V near enough - say +/-5% at the outside. 7V is too much and indicates something is wrong in this area. Your 5V input DC-DC may not like getting 7V ...

If correcting the 7V input to 5V doesn't fix the 20V output of your DC-DC then you may need to load the DC-DC to get it to stabilise - some DC-DC expect to see a minimum load and don't react well if it's not present. Add a couple of load resistors - enough to draw a few tens of milliamps may be sufficient - try some experiments to find out the value required.

James

[jackinnj]

about 3ma minimum for an LM317, but I don't think that an LM320 or 78L05 requires loading -- at least it isn't specified in the datasheets (as it is for the adjustable regulators.)

I find the acuracy of most fixed regulators to be between 1% and 2%.

speaking of loading -- smps power supplies switch out the loading when the circuit is up and running -- saves milliwatts. (Do you know that in India, at least in Tamil Nadu, they drive (during the night) with their headlights off to save gas? -- almost got killed one night down there! what they save by keeping the lamps off is offset by what they lose by constantly blowing the horn!).

[theChris]

I will try the loading. i had kinda wondered about that. would it be better to load the DC-DC converter at all times (as current is limited) or load the voltage regualtor at all times. how low of a resistor would i need. i ran a test on another "5v"regulator with a test load pot and the voltage seemed to stay at 7v and the pot just heated up (which i found odd, but it wasn't smelling too good at the end)

[theChris]

ok, i loaded both dc-dc and the regualtor.

i put a 50ohm (49 measured) load across the regualtor's output. i calculated that if the voltage was 5 volts, then the power dissipated should be about a half a watt, and that was the rating of the resistor. next i looked at what the the resistance required for 150mA output would be. i got 200 ohms across the 30v differance. i have some 240ohm resistors, close enough for testing. so i hook up 2 of the 240s, one from +15 to ground and the other from -15 to ground. (which was a mistake). got 44v...

then i looked and said, hmm, 2 240ohm resistors in series, that'd be almost 500 ohms. 2.5 times more than i wanted. i know i'll hook up a 240 ohm resistor across the 30v differance (which actually gave me about 170 ohms...) i finally got the 30v differance and was happy. then i decided to remove the resistors. i hadn't accounted for the extra wattage dissipated by the resistors, and it was hot. i had 3.75 watts on a 1/4 watt resistor. i probably should be more careful in the future...