LM7805 Current boost to 10A

[anishkgt]

Hi All,

I have the below circuit that works well. Would it possible to increase the current from the LM7805 to 10amps using minimal components like a a few resistors and transistors.

[anishkgt]

Came across this circuit but the regulator IC are different. Would this be ideal to go with ? my source is a 12v 111ma with a heat sink of the regulator.

[lapm]

Its possible but you need beefy transistor with good heat sinking... 10A * what ever voltage is over transistor is a lot of heat power to dissipate.

Design pass transistor circuit carefully and mind your power dissipation...

[jeroen79]

https://www.google.nl/search?q=7805+bypass+transistor&espv=2&source=lnms&tbm=isch&sa=X&ved=0ahUKEwiHwpyFy4rTAhWCHxoKHVGcBmUQ_AUIBigB&biw=1920&bih=985

[anishkgt]

Thank you for the reply. Beefy would mean the transistor would be bigger than To-220 package, yea ?

[lapm]

to-220 might do it, depends what your thermal calculations show. Can you keep transistor die cool enough under maximum power dissipation...

Personally i would probably go with to-3 metal can power transistor bolted to big heat sink...

Mind your transistor terminals when attaching heat sink, so you don't create short throw it.

[anishkgt]

The reason was to power this servo http://www.savoxusa.com/Savox_SC1256TG_Digital_Servo_p/savsc1256tg.htm i use uBEC (http://www.ebay.com/itm/HENGE-5v-6v-7-4v-8A-UBEC-7V-25-5V-Input-2-6-Lipo-/111814348386?hash=item1a08a78e62:g:HgkAAOSwT5tWOG-P) to power the servo at the moment that is rated for 8A, 12A Max.

[jeroen79]

You can also use multiple bypass transistors to spread the load.

[langwadt]

first you have to find a 10x larger transformer....

[janoc]

Get a switching regulator - it will be both much smaller and cheaper than trying to manhandle 7805 to such currents.

Otherwise, if you really insist on using the horribly inefficient (= HOT!) linear regulator in this application, the ways to do it are well documented:

http://www.reuk.co.uk/wordpress/electric-circuit/high-current-voltage-regulation/
http://www.electroschematics.com/1961/lm7805-voltage-regulator/

Adapt to needs/taste

[anishkgt]

I would have but that would be an additional board. I am trying to avoid that and have it all in one pcb. Moreover I have very little space on my design at the moment.


Sent from my iPhone using Tapatalk

[janoc]

I would have but that would be an additional board. I am trying to avoid that and have it all in one pcb. Moreover I have very little space on my design at the moment.

Then you are screwed no matter what you do - the pass transistor(s) and some of the resistors will need large heatsinks or maybe even a fan. That you will certainly not be able to keep that on a single board. You will be dissipating (12V-5V) * 10A = 70W of power in that transistor there - that gets pretty toasty.

You also need to replace the rectification diodes and the transformer, 1A trafo will certainly not deliver 10A and the diodes will smoke and blow up if you try to draw such current through them.

A better option would be to ditch the linear supply completely and just use a properly rated switching supply - e.g. a computer ATX supply is a good match for this type of load. You get both 5V and 12V out of it, with loads of current.

[Kjelt]

Moreover I have very little space on my design at the moment. 

The more reason to switch to a smps, the heatsink for your design will be huge.

Seperate supply could also be an option:
https://www.conrad.nl/nl/tdk-lambda-ls-50-5-acdc-inbouwnetvoeding-55-vdc-10-a-50-w-511817.html

[mariush]

You could connect 2 or 3  LM1084 regulators in parallel (with 0.2-0.3 ohm resistors on the output of each regulator for balancing the current)

See http://uk.farnell.com/texas-instruments/lm1084it-adj-nopb/ic-ldo-reg/dp/1685484  (and the datasheet in the page) , for 5v output the regulator has a current limit of minimum 5.5A , typically up to 8A ... so two in parallel should do 10A.

The datasheet for the 1083/1084/1085 regulators made by Linear (LT1083 is a 7.5A regulator, LT1084 is rated for 5A and LT1085 is 3A but LT1083 is practically impossible to get) show a circuit where they use a few feet of AWG18 wire to balance the current (basically a few feet of AWG18 wire would have maybe 0.1 ohm of resistance or something like that)

Here's the datasheet, see page 15 : http://cds.linear.com/docs/en/datasheet/108345fh.pdf

[capt bullshot]

The reason was to power this servo http://www.savoxusa.com/Savox_SC1256TG_Digital_Servo_p/savsc1256tg.htm i use uBEC (http://www.ebay.com/itm/HENGE-5v-6v-7-4v-8A-UBEC-7V-25-5V-Input-2-6-Lipo-/111814348386?hash=item1a08a78e62:g:HgkAAOSwT5tWOG-P) to power the servo at the moment that is rated for 8A, 12A Max.

These things usually draw power only for short moments (when they are moving fast). If they would draw that power continously, they get cooked. The average current should be less than 1A, otherwise - you name it: toast servo motors. So what you want is some kind of buffer for the power supply (maybe a LiPo battery or a SuperCap  of appropriate size). If your primary power source is beefy enough, you may get away using a TO220 transistor with reduced heat sink, that depends on the load cycles. A switching mode supply is kind of critical here, since you must start the design using components beeing able to run 10 Amps continously (and then do some derating on the heat sinking).

[firehopper]

you really need a bigger transformer and diodes as well. it all depends on what kind of load is on the servo, if its a constant load or not.

[macboy]

I would have but that would be an additional board. I am trying to avoid that and have it all in one pcb. Moreover I have very little space on my design at the moment.


Sent from my iPhone using Tapatalk

That circuit you posted technically could provide the power you need. A MC78xx is another manufacturer's equivalent of the LM78xx regulators. They work identically for all practical purposes. You would probably need two transistors with separate base resistors (R) and current sharing resistors as well. This makes it even more large and complicated. But you are asking for a 10 A linear regulator so that comes with that territory. The biggest problem is that you would need a heatsink at least the size of a paper back book, probably larger. No joke. Do you have space for that? I doubt it. Are you OK with wasting that much battery power as heat? I doubt that too.

You need a switching regulator like the uBEC you linked to. There is no other way. If this is a one-off, why not incorporate the uBEC into your design? Literally just add a footprint to your PCB to install that thing directly onto it. Or buy another suitable buck converter module and use that? If you really want to integrate the thing into your PCB, pick an appropriate swtiching converter IC and use a reference design from a datasheet and/or app notes. That would give you the best chance of success but I think you'd still be in over your head.

[Nusa]

I agree with the rest, a switching solution will almost certainly be smaller and more efficient.

But if you do want to continue with the linear approach, I'd suggest designing for 5V output to start with. Choose a transformer with a 6V secondary instead of 12V. Most of the heat and inefficiency comes from the size of the voltage drop, so if your unregulated DC starts out close to the target, you will generate way less heat. Of course, with your current requirements, that transformer needs physical space. I'm thinking about a 3-inch cube.

[Theboel]

if You can handle the heat than a multiple (10 pcs in parallel) 7805 can do the job off course You need balanced resistor for each 7805 You use.

but if I agreed with most answer switching is the best option 

[tszaboo]

I would have but that would be an additional board. I am trying to avoid that and have it all in one pcb. Moreover I have very little space on my design at the moment.


Sent from my iPhone using Tapatalk

If you have very little space, you should have avoided linear regulators in the first place. Switching regulators are smaller.
For linear: You need a much bigger transformer, you need a heatsink, even if the average current is only 1A. You need bulky capacitors.
I know, if feels like: "Oh no, I dont know switching, it is hard, and EMI and I'm scared", but you only need to pick up your man pants once.

[anishkgt]

Thank you all. It was valuable.

My design so far has an independent pin for supplying power to the servo via the ubec which also gets its power from the PCB. I had gone with 12v power because i could not get a transformer with 115 x 2 primary and 9v 200ma secondary to add to my PCB. well i could draw one on eagle but was lazy to go that way.

The pcb is for a spot welder.

[Monkeh]

And just how do you expect to get 10A out of any part of that design?

[anishkgt]

obviously not from the existing. I was just thinking of it and then redesign it once i get some help from here. I guess i will just stick to this.

[rstofer]

From the PCB, the transformer is rated at 1VA.  So, V*A=1 and given 12V output, you can expect a secondary current of 1VA / 12V = 0.08A.

If you want 10A at 12V, you need to look at a transformer with a rating of 120VA or just about 120 times as big!  Fortunately, transformers don't scale up that way but it's still going to be huge!

A manufactured switching power supply would look like this:
https://www.amazon.com/Switching-Transformer-Regulated-Computer-Project/dp/B0146IAXYO

As noted above, you really want to get the secondary voltage down.  You are dissipating as much heat in the regulator/pass transistor(s) as you are in the load.  A linear supply doesn't make much sense without multiple 2N3055 pass transistors and a huge heat sink.  Even then, it doesn't make sense.