Modify LM2576 dc-dc converter

[LinkZ]

Hi guys,
I've bought lots of those dc-dc converter based on Lm2576 IC (http://www.ebay.it/itm/LM2576HV-5-60V-to-1-25V-30V-DC-DC-Step-Down-Adjustable-Power-Supply-Module-/281117875719?pt=LH_DefaultDomain_0&hash=item4173ee5e07) I know that this IC can output up to 3A current, but as you see the description says that the continous working current is only 1.2A  Perhaps because the PCB cannot dissipate enough heat. Now, what if I repleace the LM2576 with a proper TO-220 version and add a heatsink? It's sufficient to work at 3A without changing the rest of the components? Is that inductor "big" enough to work with 3A CONTINOUSLY? Thank you guys.

[ElektroQuark]

This LM2576 is $3.43@500 at Digikey.
With the other parts (inductors and electrolitics are not very cheap) you can guess de total price.
Maybe they are "second market" units and can not handle that current like the proper ones?

[madires]

You would have to change both electrolytic caps and the inductor for 3A. The datasheet includes a nice step-by-step guide how to select the components.

[Zero999]

Even if you manage to find an inductor which will fit on the PCB and is capable of handling the current, the PCB traces will still only be thick enough for 1.2A.

[David_AVD]

So, replace almost all of the parts and the PCB and you'll be good.   

I've used those cheap boards before for hobby projects but would never dream of trying to pull more than an Amp or so from them.

[Kjelt]

Put three parallel at 1Amp each, resistor to balance finished.

[LinkZ]

So, replace almost all of the parts and the PCB and you'll be good.   

I've used those cheap boards before for hobby projects but would never dream of trying to pull more than an Amp or so from them.

Even if you manage to find an inductor which will fit on the PCB and is capable of handling the current, the PCB traces will still only be thick enough for 1.2A.

To cheap to be true, what a garbage! Anyway I build a proper LAB PSU  starting from zero, and yes, I make a SWITCHING variable lab psu, because I like switching power supply. And by the way alone the 100uH inductor cost more than 5€ on RS.

[madires]

And by the way alone the 100uH inductor cost more than 5€ on RS.

I've salvaged several from e-junk. Look for iron power toroids with yellow/white colouring (body is yellow and one side is white, Amidon type 26). Those are perfect for 50kHz. And broken transformers are a good source for enamel wire.

[Paul Price]

Go figure, there is a catch,  the diode must also be bigger.

[LinkZ]

Thank you guys, do you think that an EMI filter on the AC input is needed in this case, since the transformer will isolate the DC to DC converter from the outlet? Also do you believe that a toroidal transformer will perform better than a standard laminate one? Or it will be the same thing? I don't have any doubt about the EMI filter when you use a "true" switchmode psu (hooked directly on mains), but this is basically just a DC to DC converter.

[Paul Price]

No EMI filter is needed, for the reasons you mention. A standard transformer is cheaper and performance is the same as toroidal.

You can save the cost of the inductor, you can salvage high-current toroid coils from discarded Desktop PC power supplies.

The switcher IC itself if probably 100% in spec, if you just swapped the inductor and catch diode and added a heatsink to the switcher IC, you would realize the full output current the IC can deliver, probably without problems from the original circuit trace widths.
If you chose to hack this board, you would be adding the more robust parts and this would require wiring to them, and you should use higher current wire, eliminating the trace size limitation problem.

In other words you could just salvage and reuse most of the parts, take all the parts, switcher IC and input/output caps and move them to a perfboard, but upgrade the diode and inductor.

 If you are careful to use large enough wire in the common ground and high current connections in/out of the switcher to catch diode/filter caps and to the output, it would work ok at the chip's spec'd power output,  even if you are cloning the new circuit on a simple perfboard. I've had done this successfully.

[kizzap]

Also you can look into the LM2596. It is basically the same chip, however it uses a higher switching frequency, which in turn allows you to have smaller components.

[LinkZ]

Also you can look into the LM2596. It is basically the same chip, however it uses a higher switching frequency, which in turn allows you to have smaller components.

No EMI filter is needed, for the reasons you mention. A standard transformer is cheaper and performance is the same as toroidal.

You can save the cost of the inductor, you can salvage high-current toroid coils from discarded Desktop PC power supplies.

The switcher IC itself if probably 100% in spec, if you just swapped the inductor and catch diode and added a heatsink to the switcher IC, you would realize the full output current the IC can deliver, probably without problems from the original circuit trace widths.
If you chose to hack this board, you would be adding the more robust parts and this would require wiring to them, and you should use higher current wire, eliminating the trace size limitation problem.

In other words you could just salvage and reuse most of the parts, take all the parts, switcher IC and input/output caps and move them to a perfboard, but upgrade the diode and inductor.

 If you are careful to use large enough wire in the common ground and high current connections in/out of the switcher to catch diode/filter caps and to the output, it would work ok at the chip's spec'd power output,  even if you are cloning the new circuit on a simple perfboard. I've had done this successfully.

  Thank You!