Power Supply Question

Guido

I have a simple power supply question, using 1 transformer, lets 18-24 vac and one full bridge rectifier....Can you run three seperate regulated voltage regulators.....example, 5vdc, 9vdc and 12 vdc? Or is only 1 regulated supply per transformer?
Thanks
Guido

Post Edited By Moderator (Chris Savage (Parallax)) : 4/6/2005 2:14:13 AM GMT

Chris Savage

Guido,

·· This question has two answers...The short answer is yes, you can run several different power supplies from one transfomer.· The secondary winding of the transformer should have a bridge rectifier, and filtering capacitors, after which you can hang several regulators of different voltages off from it.

·· Remember though, your input voltage will affect how much each regulator has to dissapate as heat.· IOW, your 5V regulator will be running quite toasty to drop the 34 Volts you will likely see on the front side of the rectifier.·

·· If I had to choose an appropriate transformer for a 12V, 9V and 5V regulator system, I would use a 12.6 Volt transformer.· The reason is, after rectification and filtering, you'll be seeing almost 18 volts on the front side anyway.· Each regulator (Especially the 5V) will require adequate heat sinking to handle rated current.

·· You should be able to achieve around 1A using a 7805, 7809 and 7812.· Heck, while you're at it, add a 7815 and get 15 volts DC too!· LOL



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com

Guido

Thanks for the information Chris,,,
The problem is I have tried this and burnt out a couple of Regulators. I was using just 5 volt and 9 volt regulators, 12.6 VAC transformer and no load they both got real hot and burned out.I am positive I am wiring the right, just thought maybe the same ground circuit was causng a problem
Guido

Chris Savage

Guido,

·· Not sure which regulators you were using, but two things to make sure...First of all, that they have heatsinks.· You usually don't need anything huge,· Just a normal TO-220 heatsink (Even a clip-on) if you're running 750mA or less.

·· Also, if memory serves, the TAB on the regulator is·not always·ground.· I have seen a few which were different.· And so if they're tied to ground you can have problems.· Check the datasheet.·

·· I have used at least 2 on the same output.· In this case I have a 3-way supply, all running off one transformer and rectifier, and the outputs are +12V, -12V and +5V.· Well the 12V and 5V regulators come from the same source.· I have no problems, but then I am not drawing much current either.· It's for an audio controller.



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com

steve_b

You might try cascading your regulators too!
If you have 36Vac at the output of your transformer...but need 24Vdc for motors...then put in a 28Volt regulator. Then from this 28V regulator output, put in your 12V regulator (might want to use a zener to drop the voltage down to 16Volts or so. Then another 6 or 7Volt zener and then your 5V regulator.
You need to have a few volts over your regulators rated voltage (ie. need 7volts input for a 5volts regulator to regulate).

Another thing to note is that each regulator in the chain must supply adequate current to each regulator down the chain. This might become a problem as you might need to heat sink some or put higher rated components in.

Just a thought...not sure I'm dead on though!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://ca.geocities.com/steve.brady@rogers.com/index.html
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

Beau Schwabe

Along the lines of "cascading your regulators"..... Could you use a LM317 (Adjustable regulator) as a constant current source to
limit the amount of current to the other regulators?

There is an example here for charging batteries that could apply...
ourworld.compuserve.com/homepages/Bill_Bowden/page7.htm

A Quote from the link above said...

The circuit (lower right) illustrates using a LM317 variable voltage regulator as a constant current source. The voltage between the adjustment terminal and the output terminal is always 1.25 volts, so by connecting the adjustment terminal to the load and placing a resistor (R) between the load and the output terminal, a constant current of 1.25/R is established.

If you look at the datasheets for a LM317 and a LM78XX regulator, there are similar schematics to regulate the current.

I suppose that you wouldn't even need to use an adjustable regulator... If you used a 7805 and treated the GND terminal as the
ADJ, then the voltage between GND and the output would always be 5V thus making the constant current equation 5/R instead.

The question now becomes... How much current will your circuit demand.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

---

Beau Schwabe - Mask Designer III

National Semiconductor Corporation
Latest Company News
(Communication Interface Division)
500 Pinnacle Court, Suite 525
Mail Stop GA1
Norcross,GA 30071

---

Post Edited (Beau Schwabe) : 4/6/2005 8:10:51 PM GMT

David B

Another trick you can do, if your regulated circuit is going to draw a known current, is to stick a heavy-duty power resistor between the DC supply and the regulator. That way, the resistor disapates a good portion of the heat. Just be sure you leave enough voltage for the regulator to work properly.

Out of curiosity - whenever I read regulator spec sheets, they almost always say they are internally thermally protected, that they will shut down before they are destroyed by excess heat, and will restart when they cool down. Did your regulators really get destroyed by overheating?

Beau Schwabe

In some cases it's better to place your resistor across a regulator to limit current rather than at your source.

Take for example if your power supply is 30V and you are regulating down to 5V. A resistor across the 25V difference
to regulate your power down to a workable level is going to generate a lot of heat, however a resistor across the output
of a regulator and the ADJ or GND will generate far less heat with similar results.


This doesn't happen to me as much as it used to, but from experience, when a regulator shuts down due to thermal reasons it seems as though
the regulator is never quite the same afterwards.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

---

Beau Schwabe - Mask Designer III

National Semiconductor Corporation
Latest Company News
(Communication Interface Division)
500 Pinnacle Court, Suite 525
Mail Stop GA1
Norcross,GA 30071

---

Post Edited (Beau Schwabe) : 4/6/2005 9:26:59 PM GMT

Paul Baker

Believe me, just because a chip has thermal protection does not mean it cannot fail due to thermal breakdown. A famous case of this is with AMD processors, they have a thermal diode that is on chip to scale back power consumption in the event of thermal run away. The only problem is that it wasn't designed properly and the diode is too far away from the critical portions which tend to overheat, this coupled with the diode's lackluster·response time meant that the chip would still fail if something as trivial as heat sink removal or the fan died. I snagged a supply line to my·processor fan without realizing it one day, I immeadiately notices a severe degredation of performance and intermittent rebooting. I quickly deduced what was up and fixed it. Had I been using an AMD processor it would have fried itself.

Guido

I am glad I started this topic! Dam good information guys...For a little more information I was using the basic LM7805, LM7809, with a secondary of 12.6 VAC. I seem to have no trouble hooking up either one and have it work normally, but when I try and insert the other, Everything Gets hot....Maybe thats normal I don't know. I know it is wired right and would like to run two regulators (fixed) off of one transformer for a project I am doing. I just can't afford to keep smoking them...Also I am aware that on some regulators the Base may not be grounded....so I had no heat sink on them.
Thanks Again

Forrest

"Also I am aware that on some regulators the Base may not be grounded....so I had no heat sink on them."

That doesn't sound like a good reason to not have a heat sink on them. I'd install an insulator between the regulator and heat sink, along with a non-conductive shoulder washer on the screw.

Guido

Forrest,

I can not think why they shoulkd get so hot, just because I add another regulator with no load. To my way of thinking if I use the 5 volt regulator with no load and it runs semi cool and then add the 9 volt regulator! Why would both get extrtemely hot? Neither one has a load, I do realize that the % volt rtegulator should get a little warmer because of the voltage difference, but it will run some what warm if it is alone, with no 9 volt regulator??????

thanks

Guido

Chris Savage

Paul Baker said...
Believe me, just because a chip has thermal protection does not mean it cannot fail due to thermal breakdown. A famous case of this is with AMD processors, they have a thermal diode that is on chip to scale back power consumption in the event of thermal run away. The only problem is that it wasn't designed properly and the diode is too far away from the critical portions which tend to overheat, this coupled with the diode's lackluster·response time meant that the chip would still fail if something as trivial as heat sink removal or the fan died. I snagged a supply line to my·processor fan without realizing it one day, I immeadiately notices a severe degredation of performance and intermittent rebooting. I quickly deduced what was up and fixed it. Had I been using an AMD processor it would have fried itself.

Paul,

·· Not only can I confirm that the AMD processors are vulnerable to heat, but I have a video clip from Tom's Hardware Guide showing what happens to them when the heatsink is removed.· Very interesting video, since they also did it to a Pentium III and I think a P4, and they survived.· Both Athlon chips went up in smoke!



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com

David B

It does seem like something is wrong for them to get hot with no load.

Maybe they're oscillating. Most design notes I've seen say you've gotta include external capacitors close to the regulator leads to supress oscillations. Does your design have these?

Paul Baker

Oh as a general rule I don't run regulators without some load, for instance my lastest PCB has a socket for a SD/MMC card which requires a 3.3V supply. Its the only part on the board running at this voltage and it isn't always going to be there, so I added a 3.3K resistor to provide a load to the regulator when a card isn't present. The spec sheet for the regulator warned about driving highly capacitive loads and how this could lead to oscillations within the regulator. A near 0 Ohm load will appear as a capacitave load to the regulator through the output bypass cap and parasitics. Also the type of capacitor you use can effect the stability of the regulated output, this is a factor of the ESR of the cap.

Harry Stoner

Beau said...
This doesn't happen to me as much as it used to, but from experience, when a regulator shuts down due to thermal reasons it seems as though
the regulator is never quite the same afterwards.

I would agree with this based on a 7812 regulator I overheated last week. The input transformer was about 25VAC, so I was certainly at the high end of allowable input voltage for the regulator. I had a heat sink on it but it got too hot to touch and did shut down when I put a big load on it. After that it kind of worked, but not very well.

I need to drop the input voltage.

Related to heatsinks, here's my stupid question of the day: for a TO-220 package, which way should the fins of the heatsink point? Away from the chip or "towards" it (surrounding the chip). I googled quite a bit and didn't find anything definitive. If someone can point me to a good web page or provide an explanation I'd appreciate it. It's not obvious to me which way would dissipate heat better.

Thanks.

Harry

Paul Baker

Most TO-220 heak sinks have the TO-220 mounted horizontal to the board with tab down and the heat sink sandwiched between the PCB and the TO-220 with fins pointing upwards. If you don't care about lost efficiency, placing two power resistors in a voltage divider can drop the input voltage for you. Or go with a switched power supply.

Harry Stoner

Thanks Paul - that makes sense.

Harry