0
\$\begingroup\$

The last time I asked a question on this forum you were all incredibly helpful thank you for that: First PCB (raspberry pi hat)... can't find short circuit, I have an extension on this previous question.

Problem: The LM3489 chip (supposedly rated for max current of 4A) used previously became very hot to the touch (drawing around 500mA... chip does not get hot at all under no load) and while all of the components worked for > 30 min I am worried about the continuous use of this overheating chip. I understand that the tracing on the previous board was horrible and I did not follow the recommended layout at the bottom of the data sheet, but the only example circuits that I found demonstrating the LM3489 show between 100-500mA outputs, nowhere near the 3A that I need. https://www.ti.com/lit/ds/symlink/lm3489.pdf?ts=1612195295134&ref_url=https%253A%252F%252Fwww.google.com%252F

Working (but hot) Schematic for LM3489: enter image description here Current Bad Layout: enter image description here

Question: Why was the LM3489 getting so hot (tracing just that bad / chip can't actually run pulling 3A)? Will fixing the layout to match the recommended on the data sheet fix the heat?

I would really like to avoid another costly round of boards not working, so any help at all that you can provide would be immensely appreciated.

What I have now tried: Switched out C2 with a lower capacitor 100nF to 1nF with no success (board did not produce 5V out and became really hot under no load).

If any parts of this post are confusing or if I should provide any other helpful information, please let me know and I will update the information as soon as I can.

\$\endgroup\$
10
  • \$\begingroup\$ It would appear that the LM3489 controller can regulate 4A with an external FET while without is limited by the 6 ~8 Ohm internal FETs to about 400mA with some duty cycle . \$\endgroup\$ Commented Feb 2, 2021 at 5:28
  • \$\begingroup\$ In the schematic linked in the previous post I attempted to use a "P-Channel -60V 3A 3V @ 250uA 105mΩ @ 3A,10V 1.6W SOT-23-6 MOSFET RoHS" as the external PFET which would allow for 3A or did I implement that incorrectly? \$\endgroup\$ Commented Feb 2, 2021 at 5:36
  • 1
    \$\begingroup\$ Would you mind reposting the schematic of your actual board here? I know it’s linked from your other SE post but I had to scratch my head for a minute to find which of these inline schematics was the right one for your LM3489 solution. \$\endgroup\$
    – pion
    Commented Feb 2, 2021 at 5:36
  • \$\begingroup\$ Added "Working (but hot) Schematic for LM3489" under Problem: \$\endgroup\$ Commented Feb 2, 2021 at 5:43
  • \$\begingroup\$ C2 100nF conducts all the high frequency noise to the FB pin which makes the driver switch more often trying to reduce ripple that it by this cap is creating from hysteresis as there is not fixed clock inside. The faster it switches, the more energy is dissipated. Although claimed this "compensation cap" is not necessary, one example shows 100pF in the feedback divider. \$\endgroup\$ Commented Feb 2, 2021 at 6:18

2 Answers 2

3
\$\begingroup\$

Problem: The LM3489 chip (supposedly rated for max current of 4A)...

The LM3489 is a PFET Buck Controller. Maximum current is determined by which external FET, inductor and flyback diode are used.

the only example circuits that I found demonstrating the LM3489 show between 100-500mA outputs, nowhere near the 3A that I need.

You can use the LM3489 in a higher current design, but you will have to upgrade the external components to handle it.

The FDC5614P supposedly has a maximum current rating of 3 A when when mounted on a 1in2 pad of 2oz copper on FR-4 board. In practice you would not want to do that because the FET would run extremely hot and have a high chance of burning out, as well as heating up other components around it. In your circuit the FET should only be conducting for a part of each cycle so you might get way with it, but for safety I would choose a device with much lower RDSON.

If a 22 μH inductor is appropriate for 0.5 A then for 3 A you may need a smaller value to avoid saturation (4.7 μH is a typical value used in regulators with similar sized inductors).

When converting from 24 V to 5 V the flyback diode will have to pass ~80% of the average output current, ie. 2.4 A at 3 A. The MBRS140 is only rated for 1 A.

To determine the best components for your application I suggest you either simulate the circuit, or copy an existing design with similar specs. Ti's reference design PMP7771 uses a FET with 0.021 Ω RDSON, a 5.6 μH inductor, and a 3 A Schottky diode.

\$\endgroup\$
0
\$\begingroup\$

An immediate thought is that the current limiter may be kicking in. Section 8.2.2.4 of the datasheet says that you can disable the current limiter by tying ADJ to GND and ISENSE to VIN. As a quick experiment, can you try doing this and see if your max current increases? (This will involve desoldering three passives and adding a couple of external wires, or making a breadboard version of your circuit.)

\$\endgroup\$
3
  • \$\begingroup\$ Unfortunately I cannot reach the boards until tomorrow, but I think the ADJ and ISENSE are connect as per data sheet recommendation (see updated post with previous schematic). I will see about running this test and update as soon as I can get to them. \$\endgroup\$ Commented Feb 2, 2021 at 5:45
  • \$\begingroup\$ Could you elaborate on the test you want me to run, with the current configuration of ADJ to VIN and ISENSE to the FET gate I am not sure what to do? \$\endgroup\$ Commented Feb 2, 2021 at 19:10
  • \$\begingroup\$ I’m simply suggesting to try temporarily disabling the current limiter according to the datasheet. Referring to your schematic, you will have to desolder R1, R2, and C4, then solder a wire from the ADJ pin to GND and solder another wire from the ISENSE pin to VIN. \$\endgroup\$
    – pion
    Commented Feb 3, 2021 at 6:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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