20 Replies Latest reply on Feb 13, 2021 11:51 AM by vtrandal

# LM317 in parallel - Answered -> LM2596 circuit question

Hello,

I am building a DC-DC converter to convert the 14.8v 6A power coming from a professional camera battery to 9V 2.5A used by a DSLR, I am trying to build this converter using the LM317 since it's the only easily available power regulator (which I don't have to order 2500pcs to place an order), the problem is the LM317 can only output 1.5A max, can I build the circuit using 2 of them and connecting the outputs of each circuit in paralel to get 9V 3A, is there going to be any kind of problem that may fry the camera?

Thank you

Oscar

Message was edited by: Christos Oscar Kambiselis

• ###### Re: LM317 in parallel

I'm interested to see what answers you get. I googled around and found several sources that said this was ok. Although 1 design used three of them to ensure he got the 3A he wanted. It's been many years since I did electronics professionally, now it's all hobby. I'm all software now. So it would be prudent to wait on an answer from the pros.

Mike

• ###### Re: LM317 in parallel

While I'm not a power supply expert, my understanding is that hooking up the outputs of two linear reguators in parallel is usually a bad idea.  It's very hard to get them to share the load equally, i.e., each put out approximately the same output current.  Usually one dominates, producing most or all of the output current while the other is idle.  I've heard that it's also possible to go into oscillation, with the two regulators fighting each other.

Using linear regulators to convert 14.8V to 9V is a 5.8V drop at 2.5A.  This is 14.5 watts that's going to waste and could make things pretty warm.  You probably want to use a switching regulator, which wastes far less power and runs much cooler.

• ###### Re: LM317 in parallel

In the LM 317 datasheet http://www.ti.com/lit/ds/symlink/lm317.pdf (sorry about the pdf) figure 13 on page 12 shows three of them in parallel (of course with supporting circuitry) to create an adjustable 4A regulator. You would probably have to order 2500 of one of the other parts in the diagram.

The regulator they diagram is 4.5 to 25 volts.

Again, though, I defer to the experts.

Mike

Message was edited by: Michael Conners

• ###### Re: LM317 in parallel

Hi Michael,

The pdf is moved here:http://www.ti.com/lit/gpn/lm317  via http://www.ti.com/product/lm317

I think you talk about page 12 / fig. 13.

It seems ok.

Hi Christos,

I would use the Webench Designer and choose some switching IC. You can request free samples from TI to try the design.

I used many LM317 and LM317K (TO-3 case) but they generate some heat at high current as I remember.

Regards,

Dan

• ###### Re: LM317 in parallel

Danteo I have used Webench to easily find a schematic for my specs but, as I said, many components can't be found or must be order in 500+ quantities, don't know if it is possible to order samples from TI since I don't work in a big company (I only have a sole proprietorship business) and I'm located in Finalnd, but I will check.

• ###### Re: LM317 in parallel

Hello Folks, I have studied regulators and it's my opinion that if you use three identical regulators you could get away with the triple in parallel circuit, but the best bet is figure 14 of the data sheet. You don't have to slavishly follow the designed semiconductors so if one chip or transistor is unavailable try Googling the original part number e g TL080 and you will get access to a data sheet, decide what are the important parameters, in this case its an FET op-amp if I'm not mistaken (look it up) and try to Google for equivalence tables which should be available. The more time you spend in the design stage the better since you will accumulate knowledge and become adept at scanning data sheets. A lot of info on data sheets is superfluous or unnecessarily accurate. Then make up your 1st stab as a breadboard and put it through its paces. You can tell if a device is overheating with your fingers, but I can recommend getting a non-contact thermometer/pyrometer (digital) which are available for under £20. Then you can do some graph plotting. It's a wonderful hobby and also profession, but to be professional you have to surrender the serendipitous side of things and crack your problems. Happy hunting.

• ###### Re: LM317 in parallel

PS the reference M Conners gave for the data sheet was wrong, change 117 to 317 in the URL.

• ###### Re: LM317 in parallel

Interesting, that link was there yesterday. It was the datasheet for the LM117 and the LM317. And yes, now it is the circuit on page 12 , figure 13.

Thanks.

Mike

• ###### Re: LM317 in parallel

I updated the link and the reference.

Mike

• ###### Re: LM317 in parallel

One thing to keep in mind is that you said you will be running this off of a battery. I think John Beetem is correct in asserting that you will be generating quite a bit of heat with this circuit. Heat means power and I would be afraid that you might drain your battery pretty rapidly with a lot of unused potential.

Hopefully we have given you enough ammunition and information to help you decide whether this is still the direction you want to go.

Good Luck,

Mike

• ###### Re: LM317 in parallel

the lm2596 is a good choice for a simple buck converter http://www.ti.com/lit/ds/symlink/lm2596.pdf

it is avalable in through hole, and needs minimal heat sinking. and the math is straight forward right on data sheet

• ###### Re: LM317 in parallel

Sheldon the LM2596 seems to be a good choise, since I can find already assembled circuits using it even on eBay for not more than 4€, seems that I will order a bunch of them since they are adjustable and use them in my other projects too.

• ###### Re: LM317 in parallel

I actually found an assembled LM2596 circuit in a really small package from a Chinese eBay seller for only 1.05 pounds with free postage, with 1 year warranty, ordered 5 of them :-P

Well since I found a better solution for my problem, only one question remains, what would be the ideal placement of a heatsink to this board (since if using it with a greater than 15W load it is recomended) would I have to only cool the chip or the whole unit, by the way this will be placed inside a solid aluminum small box measuring about 8x5x5cm , would it be better to mill the block so that it fits exactly inside, so the case itself is a heatsink, or do you believe it would get too hot.

Module photo:

• ###### Re: LM317 in parallel

it looks like the regulator is heatsunk through the ground plane. the switching regulators are much more efficent than a linear regulator. a moderate air flow around the board should do the trick.

• ###### Re: LM317 in parallel - Answered -> LM2596 circuit question

I invented the Lm317.   It is NOT designed to be paralleled. If you parallel devices, one will supply all the current until it thermally or current limits and then the other one will come on.  This not reliable.

The data sheet show circuitry that allow paralleling by forcing current sharing. A new device, the LT3081 is designed for easy sharing. Also the LT 3080

Bob Dobkin

2 of 2 people found this helpful
• ###### Re: LM317 in parallel - Answered -> LM2596 circuit question

welcome to the forum Bob!

1 of 1 people found this helpful

• ###### Re: LM317 in parallel - Answered -> LM2596 circuit question

Talk about stright form the horses mouth - nice one Bob, you have been responsible for many a power supply project....just goes to show how well your design as stood the test of time...

Gerry

• ###### Re: LM317 in parallel - Answered -> LM2596 circuit question

Bob [Bob Dobkin], your reply in this forum seems largely unnoticed. Thank you for inventing the LM317 and creating such a useful device [and revenue engine] that I am just now learning about. And thank you for suggesting the newer LT3081 and LT3080 as alternatives when needing more current than one LM317 can provide.

Regarding voltage sources in general I think putting battery cells in parallel also has problems. The cells will eventually drain each other unless something is done to prevent that.

• ###### Re: LM317 in parallel - Answered -> LM2596 circuit question

Hi all,

I agree with all comments, better the LM2596 than a linear regulator, probably you would have loses about 15W with linear, so switched mode power supply will be better, because of the battery lifetime, and for the heatsink that would be needed.

The module you show, looks great!! But in order to calculate the heatsink you need, we must know the efficientcy of this module, in order to know the power loses.

Also as Sheldon says it seems to be heatsunk through the ground plane, furthermore if the manufacturer recommends heatsink, probably it will recommend how mount it, or which surface must be in touch with the heatsink.

Do you have a link or the datasheet of the module? Even, you should know how much heat can the external box disipate?

But, step by step, first we have to know the efficientcy and the loses, so a datasheet would be great.

I hope I could help.

Let us know how your project goes on.

Regards,

Miguel