I have a 17V DC power supply which can provide around 50A. I want to build a battery charger using it which outputs 14.6V constant voltage. Therefore, I'm planning to manipulate low cost, widely available LM2596 buck converter module like this. LM2596 typical application diagram

High current modification

Everything inside red dashed line after A-B points in typical LM2596 application (which is basically the filtering part) will be replaced by this - a power MOSFET (gate coupled to LM2596 Output), a schottky diode (that can handle 50A) and an appropriate inductor. Feedback will be connected to the output of this part.

Would this work as expected or have I got it wrong? Has anybody tried this before? Your ideas are highly appreciated. Thanks.

  • 2
    \$\begingroup\$ Switchers with internal mosfets are not meant to drive external mosfets. The part you're looking for is a buck controller - the chip includes the oscillator, feedback, protection and gate driver, but it doesn't include a mosfet, and is meant to drive a mosfet provided by you. Furthermore, the datasheet and/or application notes should help you out with design equations and important considerations to make it work and be robust. \$\endgroup\$ Mar 17 at 20:36
  • \$\begingroup\$ Thanks for your reply. Actually this is what I was looking for - driving an external MOSFET uisng a controller that has internal MOSFETs. So your reply addresses my problem. \$\endgroup\$
    – Tencent
    Mar 18 at 4:50
  1. Your output cap is reverse polarity.
  2. It's a PMOS so it's required state is inverted relative to the output of the IC. If it were an NMOS requires a high-side floating drive instead.
  3. MOSFETs require actively pulling it LO or shorting the gate to source to turn off. Simply disconnecting the drive signal leaves the gate-source capacitance charged thus leaving the MOSFET on.

With the amount of effort you're already going to, just build one from scratch using an appropriate buck converter IC. At 50A, you might find a single-phase converter lacking anyways.

Except that this isn't a power supply. It's a charger. And 50A is probably fast charging which means it's also probably a lithium battery which means if you mess up you're going to burn your house down. Don't mess around with charging lithium batteries with half-assed methods that just act like a power supply and not like an actual battery charger. And especially if fast-charging them.

So don't go with a buck converter IC, go for an actual charger IC instead. Or you can add the appropriate high-side drive to your little circuit above, add on a current sensor, possibly replace the diode with a synchronous NMOS, and have an MCU control them to make a fancy MCU controlled charger. And of course, you'd have to read up on the charging algorithms and very carefully test things under safe conditions and close supervision until you know the code is correct, lest you burn your house down.

  • 2
    \$\begingroup\$ I agree. Modifying a buck converter with internal power FETs to work with external power FETs is rather dubious. Now, using a buck controller which needs external power FETs anyway might be the correct path. Except that it will be a power supply and power supplies are not battery chargers and power supplies cannot be connected to charge batteries, so if the point is to charge batteries, a power supply is a completely wrong for this. \$\endgroup\$
    – Justme
    Mar 17 at 20:35
  • 1
    \$\begingroup\$ That's a PMOS meaning the 2596 output is inverted compared to the necessary drive waveform. Which doesn't affect your conclusion ... this is not a good idea. \$\endgroup\$ Mar 17 at 20:47
  • \$\begingroup\$ @BrianDrummond Good catch. Corrected. \$\endgroup\$
    – DKNguyen
    Mar 17 at 21:24
  • \$\begingroup\$ Thanks for your reply. I agree about all your points 1-3. It's my bad, actually I didn't designed or draw this diagram myself. I just presented it as a topology. And yeah, I understand what you say about battery chargers. CC controlling has to be implemented on top of this, which I'll be doing as a next step. But achieving CV control is the key to everything else, which I was focusing here. \$\endgroup\$
    – Tencent
    Mar 18 at 5:04
  • \$\begingroup\$ Problem with dedicated charger chips is, most of them only use internal switching and feedbak controlling and are limited to currents less than 10A and are not widely available either. However, I think now I should go with a dedicated buck controller chip that supports external switching MOSFETs \$\endgroup\$
    – Tencent
    Mar 18 at 5:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.