Variable Pwr Supply project 0-15v, 20amp

Thread Starter

hermhart

Joined Feb 16, 2011
20
I would like to build a variable DC voltage output power supply. It does not need to be very accurate, but it needs to start at 0 volts and go to about 15 volts and be capable of about 20 amps. I have seen some projects listed out there. With that high of amps, they tend to be fixed voltage. Some cannot go any lower than 1.2 volts. I haven't given up searching, but I decided to put a query out in the forum and see what help is available.
 

KMoffett

Joined Dec 19, 2007
2,918
Do you really need 0.5V at 20A? Too often people, new to the field, want one power supply that will run an LED and jump-start MACK truck. It's much more sensible to make several power supplies that do what you will really need. If you need to go down to nearer 0VDC, what current do you really need at that voltage. Maybe more details about what you will be powering. Then we can make suitable recommendations. :)

Ken
 

wayneh

Joined Sep 9, 2010
17,496
Well, this would relatively easy except for the voltage below ~1.5v. That's because the voltage regulator itself needs to see a 1.25v drop. There are such things as LDO (low drop out) regulators, but that only narrows the gap. Are you sure you need 20A at, say 0.5v?

Note that a conventional linear circuit will dissipate a LOT of wasted heat and you'll need big heat sinks and a strategy for removing the heat. Or you need a more advanced, efficient design such as a SMPS (switch-mode power supply). These are great but more complex, and I really don't know about chances of getting the 20A over a wide voltage range.
 

Thread Starter

hermhart

Joined Feb 16, 2011
20
Well, my client is now saying that 10 amps is plenty. I was being over generous. Also the current available can drop as the voltage drops. At low voltage, low current is OK. I don't know if that's how typical simple power supply circuits work, though. Also, we probably can manage not going all the way down to zero if it's difficult to accomplish that. One entry I read talked about a bipolar supply. We'd like to do something reasonably quick and dirty.
 

Jaguarjoe

Joined Apr 7, 2010
767
Look at the datasheet for National's LM317 3 pin regulator. In the app notes you find a way to add more pass transistors to boost current as well as a way to run the 317 down to zero volts. Both of these ideas can be combined to give you the power supply your client requires. Depending on input and output voltage the 317's power dissipation can be quite high. This will limit current at low voltages.

http://www.national.com/ds/LM/LM117.pdf

Switching power supplies are much more efficient but I don't know if they have variable voltage outputs.
 
Last edited:

KMoffett

Joined Dec 19, 2007
2,918
Well, my client is now saying that 10 amps is plenty. I was being over generous. Also the current available can drop as the voltage drops. At low voltage, low current is OK. I don't know if that's how typical simple power supply circuits work, though. Also, we probably can manage not going all the way down to zero if it's difficult to accomplish that. One entry I read talked about a bipolar supply. We'd like to do something reasonably quick and dirty.
It sounds like we are three steps away form the needed information, and that this is not your field of expertise. What has your client given you as to the purpose of the power supply, and are there actual spec's?

Ken
 

wayneh

Joined Sep 9, 2010
17,496
Look at the datasheet for National's LM317 3 pin regulator. In the app notes you find a way to add more pass transistors to boost current as well as a way to run the 317 down to zero volts. Both of these ideas can be combined to give you the power supply your client requires. Depending on input and output voltage the 317's power dissipation can be quite high. This will limit current at low voltages.
+1

Exactly what I was thinking of earlier. Not elegant or efficient, but quick and dirty, and cheap except for the transformer and heat removal. Even 10A calls for a non-trivial transformer. If the OP doesn't have one on hand, this could be the key part.
 

Thread Starter

hermhart

Joined Feb 16, 2011
20
Well, you're all giving me some excellent information. The off-the-shelf is a kindof overkill, simple, but not cheap. National Semiconductor looks like a good path to be on. Their apps are helpful. At the moment we are looking at the LM338 which gives us 5 amps out the chute. (http://www.national.com/mpf/LM/LM338.html#Overview).

Then this circuit ...

http://electronics-diy.com/electronic_schematic.php?id=650

...would seem to indicate that we could put several LM338s in parallel to give multiples of the current capability.

Here's what we're doing. We have a small gasoline engine. We're using an automotive alternator as a load for the engine. This power supply supplies power to excite the field magnet of the alternator...higher voltage on it gives greater load experienced by the engine. A 12 volt battery is too great a load. So we know we need less than that.

So that's where we're at at the moment.
 

Jaguarjoe

Joined Apr 7, 2010
767
The last time I fiddled with alternators, the rotating field only needed 2 amps or so. Measure it with an ohmeter to be sure. A 5 amp LM338K by itself will handle this. There is no real need to go to zero volts, 1.5 volts on the field is inconsequential.
The car battery and a rheostat will do this too.
Your client will need a load for the alternator output. Something that can eat around 90a * 12v or ~ 1kw.
1hp being 746 watts, is 1 alternator enough?
 

Thread Starter

hermhart

Joined Feb 16, 2011
20
Well, we're doing engineering work on this tiny little engine. This alternator/load is a sort of make-do dynamometer. We have commercially made eddy current dynos in labs designed for full sized engines. But we got this tiny little engine and for it we are quickly putting together this little "dyno" load for it.
 

Thread Starter

hermhart

Joined Feb 16, 2011
20
JaguarJoe, We're getting about 50 amps max out of the alternator. It's a 5hp engine. We're feeding the alternator's power into a car battery. But we'd be glad for suggestions on what to use for a load bank. Any ideas?
 

wayneh

Joined Sep 9, 2010
17,496
But we'd be glad for suggestions on what to use for a load bank. Any ideas?
Lightbulbs are cheap, easy and provide good feedback. Resistive heaters, like the kind for starting charcoal or warming up a water bath, also take a lot of juice.

One thing about your power supply: Do you know if you really want constant current, or constant voltage? I don't think they're the same once you start spinning the magnets around. My guess is you need constant current so that the magnetic field created by the field windings is constant.
 
Top