More current from the LM317T Voltage Regulator

[iLoveJeeps]

I have quite a few LM317T variable voltage regulators and lots of spare electronics parts. I.E. transistors "PNP and NPN", Mostly 2N3906 PNP's and 2N2222 NPN's, and a couple 2N3904's. My question is how can i build a voltage regulator that can put out 2 - 3 Amps at 12 volts with 25.5 volts going in.. Its for an arial photography company i'm starting up with my drone and my power source is a 6 cell 22.2 v 16,000 mah battery. Ive looked at schematics for such a circuit and ive learned that the more 3055's you use the more current the LM317T can put out.. Is there and equivalent transistor i can use such as the ones i've listed that can work instead? If not, then which can i use and can you show me the circuit diagram to build it. It has to be light weight because its for a 6 propeller drone that already weighs 10 pounds without the camera, pan/tilt mount and the 1000 mw 5.8 ghz video transmitter. Which is where I need the 12 volts and 3 amp current rated regulating circuit.. If i have to get different transistors thats fine,, but would like to use my current regulator. Here is the link of the current circuit i would like to build.. It's the third schematic down using 3 NPN transistors in parallel. http://www.reuk.co.uk/LM317-Adjustable-Power-Supply.htm Please keep in mind that i am an amatuer at dealing with electronics, but can read and build schematics.. And this is the first time i've ever posted in a forum. lol any help would be great.. Thanks guys/gals...

 

[Mosaic]

LM317 is a linear reg that needs a heatsink, as will your pass transistors for higher amperage.
3A @ 13.5V drop is = 40.5W of power turned into heat.
Since weight is a constraint, it makes no sense to add heatsink weight as you're going backwards.

Do yourself a favour and just get a few of these:
**broken link removed**

No weight issues..or heatsink, 93% power eff....now your batteries can be smaller as a LOT less power is wasted.

PS: Welcome

 

[iLoveJeeps]

Thanks for the welcome Mosaic! I do have a 12 volt 3 amp regulator circuit i bought off ebay, but it only supports up to 5 cell LiPo batterys which is only a 3.7 volt difference, and the thing weighs about 15 grams... how can it be so light weight yet still give me the voltage and the amperage i need? It did burn up because i was hoping it could handle the extra 3.7 volts

 

[iLoveJeeps]

I looked at that step down link on ebay,, Thats exactly what i need but its gotta come from hong kong, which would takes weeks to get here. This is a time sensitive issue. I should have mentioned that also..

 

[Mosaic]

The regulators of choice these days are BUCK SMPS units when you're going for a lower voltage. Their advantages are very high efficiency therefore small and light and higher current sourcing than the actual supply current.
Thus if you note the details of the item I suggested you'll see that its lower voltage output provides more current than the input requires. Conservation of Power where DC power= Volts x Amps.

Linear regs are fine for cheap, low current applications.

For fast delivery...it costs more:
https://www.ebay.com/itm/DC-to-DC-4...898?pt=LH_DefaultDomain_0&hash=item233a79180a

 

[iLoveJeeps]

I just so happened to be looking at the exact same item on ebay! Shipping cost is no matter.. It takes the fun out of building it myself but this is exactly what i need.. you've been a tremendous help!! Thank you a thousand times!.. If you know of any kits or shematics with a parts list to build a similar circuit, let me know,, but this is all i need! Thank you again!!

 

[Mosaic]

 

[crutschow]

I just so happened to be looking at the exact same item on ebay! Shipping cost is no matter.. It takes the fun out of building it myself but this is exactly what i need.. you've been a tremendous help!! Thank you a thousand times!.. If you know of any kits or shematics with a parts list to build a similar circuit, let me know,, but this is all i need! Thank you again!!

Companies like Linear Technology and TI sell many types of buck converters that should meet your requirements.
They offer test boards for some of them to simplify their evaluation.
It does, however, take a little work to wade through the selection pages to find what you need.

 

[MikeMl]

Don't overlook the potential RFI produced by Buck converters. I hate the damn things. I have never found one that doesn't generate broad-band hash or birdies that interfere with radio reception, especially in the 1MHz to 30MHz range. Same is true for SMPS power supplies used for lighting.

Also, if you are powering analog circuits (sensors, opamps, instrumentation amps, AD conversion, etc), Buck converters have a lot more noise on their DC output than a linear regulator.

 

[MrAl]

If you need current and you can use a linear, then check out the LM338.
That's got a 5 amp limit.
I use one with ultra precision op amp control for my precision power supply back end, with switcher up front.
A linear works pretty efficiently if the input volts is close to the output voltage, but of course you always need that little overhead voltage.

There are still lower frequency buck's around 50kHz that work pretty nice. LM2576 for example.

 

[atferrari]

If you need current and you can use a linear, then check out the LM338.
That's got a 5 amp limit.
I use one with ultra precision op amp control for my precision power supply back end, with switcher up front.
A linear works pretty efficiently if the input volts is close to the output voltage, but of course you always need that little overhead voltage.

There are still lower frequency buck's around 50kHz that work pretty nice. LM2576 for example.

Would you mind showing the circuit of the control part, MrAl? Expecting to start experimenting with that very soon.

What specs has your "ultra precision"? Could you tell range and resolution as well?

Honestly, the full circuit would be even better.

 

[MrAl]

Would you mind showing the circuit of the control part, MrAl? Expecting to start experimenting with that very soon.

What specs has your "ultra precision"? Could you tell range and resolution as well?

Honestly, the full circuit would be even better.

Hi there,

Actually i am happy to hear that other people are interested in these kind of circuits. Since i left the industry i still like to work with these circuits as a hobby and it's nice to see other people do too.

As far as showing the *entire* circuit however, let me think about that. I suppose i could.
It's not entirely difficult however, as the basic idea is to use an op amp to control the LM338. That way we get very precise control over the output voltage but need only one TO220 package as the pass device.

The real difference though came about through an investigation into 'fine' control techniques. It seems that there were not many solutions out there that i could find that would allow for very precise voltage settings, and the resolution of these adjustments always seemed to vary with output voltage too, which i did not like at all. For example, at 3v output we might adjust by 10mv per tenth turn of the pot (0.1 of the full rotation) but at 30v output we might only get 100mv resolution in the adjustment with 0.1 of the full rotation. What i wanted was at least 5mv resolution with maybe 1/2 turn of the pot, and i didnt want to have to use a 100 ohm pot either as the good ones are much more expensive then ordinary 1k or 5k pots.
It turns out that there is a fairly cheap and effective way to do this too, and it worked out pretty well. I needed to be able to set voltages for battery charging for different types of batteries, at least to do some testing anyway, but it's nice to have for everything else too.
So the real difference is the low cost and highly effective output voltage adjustment.

The temperature spec comes in mainly as the choice of voltage reference diode and low drift op amp. I did not expect a really huge difference in temperature indoors, but i found that the LM3xx devices had unacceptable temperature drift when used for precision bench power supplies, at least when used alone. That was because the voltage reference for these devices is located on the same chip as the voltage regulation circuit itself, which means it is subject to a large temperature rise for high input to output voltage differentials. It's amazing how bad it can get too, and i found that unacceptable also. Taking the reference off chip (away from the main heat sink) means it is always subject to room temperature, which makes a heck of a difference in the temperature spec when used indoors in a roughly controlled temperature environment like 65 to 85 degrees C.

So the main design part is just an op amp that controls the LM338, and that's not too difficult to come up with, and the adjustment comes from the pot connected to the op amp rather than the LM338.
I'll try to post a schematic by tomorrow afternoon or sooner.

 

[atferrari]

I'll try to post a schematic by tomorrow afternoon or sooner.

Really interested. Gracias

 

[MikeMl]

Why I hate switchers (SMPS).

**broken link removed**

 

[MrAl]

Hi Mike,

Switchers are a necessary evil of the future because we are always trying to reduce the power lost that goes for nothing like a whole bunch of wall warts plugged in and not doing anything at the time. But i do share some of your remorse. The older 'regular' unregulated wall warts would last for years plugged in, but i had a switcher type for a scanner a while back that gave out after only a few years. Something went wrong inside and smoked some components. I dont think i even had a 'regular' wall wart fail though because they are made solid and with few parts.

Yeah the noise is an issue that has to be dealt with and if the manufacturer doesnt do it right then we have a problem.

But some of my power supplies would not do so well without their switcher front end, so i am stuck in the middle...sometimes i have to use them. For the PC computer, we'd be lost without them because we'd need heavy duty power transformers to power one computer. So it's a trade off...a little more noise for reduced weight, reduced cost, and reduced power consumption.

 

[Tony Stewart]

As others stated linear reg makes no sense to supply 36W while drawing 75W from a drone battery source. Without knowing your Radio susceptibility to conducted and radiated noise, a linear regulator only makes sense if used in series with a SMPS designed to regulated the minimum voltage drop. (2.5V poor old LDO types or <0.5V with modern LDO types)

Both need to be EMI isolated for conducted and radiated noise to give best tradeoff between noise quality and efficiency target of 90%.

Rushing into a linear solution may be a compromise in battery life.

 

[MrAl]

Hello again atferrari,

Here is the basic circuit. Note the resistor for the voltage reference diode depends on what kind of diode you use and what the input voltage range is.
Also, i used this circuit to charge batteries, especially 12v lead acid batteries, so you might have to check for oscillation for different loads.

Also remember to think about Tony's notes before you use it with any significant output current.

I almost forgot to mention, i used the LM358 for testing and it works pretty well, but the better choice of op amp is one of the higher slew rate types with like 10v/us slew rate, and maybe 10MHz bandwidth. That would make the response as fast as the original LM317 but with better accuracy and temperature characteristics.
BTW since the temperature characteristics are now solely a function of the voltage reference diode (given a quality op amp) the better the reference voltage temperature profile is the better the temperature characteristics of the entire power supply is.
For fine control the usual method is to insert a 100 ohm pot in series with the top terminal of the COARSE adjustment pot.

 

[Mosaic]

Dealing with SMPS RF noise ..use a conductive housing as a faraday cage, such as a mint can. Or rig one up a using PCB Box tied to the ground plane, solder the seams with desolder braid or seal with conductive foam made for the job.

 

[shokjok]

Mr. Al, why are you counteracting the op amp? Sharing the output with ground should almost certainly burn it out. Using a capacitor is better than the trimpot.

 

[Tony Stewart]

Mr. Al, why are you counteracting the op amp? Sharing the output with ground should almost certainly burn it out. Using a capacitor is better than the trimpot.

If only a cap could limit the regulator control voltage. I don't see any ground or OA output violation but details like noise filtering were omitted for simplicity. The OA is just a feedback voltage with diode OR control for reducing the output.