0
\$\begingroup\$

I designed a power converter PCB for up to 24 VAC input and +5 VDC output. I assembled three boards and tested them with a +12 VDC bench power supply, but it always outputs the input voltage minus the diode drop voltage.

The layout should be pretty decent with a solid GND plane, short HF loops etc. according to the datasheet.

Also the schematic looks like the one recommended for +5 VDC in the datasheet and I use exactly the recommended inductor. I really don't know what is going on. I chose capacitors with a high enough voltage rating, so that derating is compensated.

What I measure on the different inputs of the buck converter:

  • FB = 1.7 VDC @ 11.3 VDC input voltage
  • INTVcc = 3.45 VDC constant
  • SW: Gives sometimes a low pulse, but is nearly always high, that is the reason for the same output voltage as the input voltage.
  • PG: is low, so indicating a power error

(Note, that I had a pretty bad GND loop with the probe for this measurement, since it was hard to reach it, so ignore the overshoots.)

enter image description here

I really don't know where the mistake is.

Important: R2 and C7 are not placed in my schematic.

Here is the datasheet and the reference schematic from the datasheet:

enter image description here

Here is my schematic:

enter image description here

Here is my layout:

enter image description here

enter image description here

enter image description here

enter image description here

\$\endgroup\$
5
  • \$\begingroup\$ What is the exact part that you installed on the board? Note that the parts ending in -5 or -3.3 need the output voltage on the feedback pin, rather that via resistor divider \$\endgroup\$
    – Ferrybig
    Commented Jan 31, 2022 at 10:30
  • \$\begingroup\$ Ok, you might be right, it is the LT8609AEDDM-5#TRPBF \$\endgroup\$ Commented Jan 31, 2022 at 10:32
  • \$\begingroup\$ Could you bridge out R5 with s short for testing purposes? \$\endgroup\$
    – Ferrybig
    Commented Jan 31, 2022 at 10:33
  • \$\begingroup\$ You guys are totally right, I choose the wrong one without the FB path. So I need only a direct FB and no resistor divider network. Thanks \$\endgroup\$ Commented Jan 31, 2022 at 10:39
  • \$\begingroup\$ Ok, it works perfectly after shorting the FB path. I really did not see that in the datasheet, tought there were no fixed voltages, only adjustable. Thanks \$\endgroup\$ Commented Jan 31, 2022 at 11:07

1 Answer 1

3
\$\begingroup\$

Switch mode converter chips generally come in 2 different types, adjustable and fixed voltage ones.

Adjustable regulate the voltage on their feedback pin to their reference voltage, and the feedback pin is an high impedance input

Fixed voltage have a lower impedance on the feedback pin, and expect different voltages on the feedback pin. They are harder to combine with extra resistors, as the values of the internal resistive divider is not always mentioned in the datasheet.

You designed your circuit for an LT8609AEDDM, but actually populated it with an LT8609AEDDM-5. The chip wants 5V on is feedback pin, but is only getting 1.7V, so it reaches near 100% duty cycle, and keeps its SW pin mostly directly connected t the positive pin.

In your circuit, the fix is simple, remove R4, R5 and C8, and place 0ohm resistor (if using an assembly service) or a solder blob between the pins of R5

\$\endgroup\$

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.