1
\$\begingroup\$
  • The MAX1771(always in DCM, has no oscillator) was chosen as controller for minimum component count: no error amplifier compensation, no slope compensation, no current feedback RC filter, no external oscillator.

  • The input bulk cap is low ESR, the inductor is shielded. Q1: STB41N40DM6AG. D2: GB02SLT12-214.

  • 12 V in, 170 V out, 10 mA load. The frequency is very low and inside hearing range.

  • Schematic: enter image description here

  • MAX1771 block diagram enter image description here

  • I lost the PCB file, but the layout is similar to this.

  • Gate node:

enter image description here

  • SW node (x10 probe):

enter image description here

  • Output: (x10 probe):

enter image description here

  • Current feedback:

enter image description here

Some DCM equations:

$$I_{max}=\frac{V_i D}{f L}$$

$$\frac{V_o}{V_i}=\frac{I_{max} D}{2I_o}$$

However MAX1771 changes frequency and duty at the same time, which makes it unclear how frequency can be increased.

Suggested method to increase switching frequency:

  • Reduce max inductor current.
  • Reduce inductance.
\$\endgroup\$
10
  • 3
    \$\begingroup\$ Dang! 14 times single stage boost. You may be hitting the minimum on-time of the controller and going into skip mode or loose regulation. \$\endgroup\$
    – winny
    Commented Dec 18, 2021 at 17:37
  • \$\begingroup\$ @Andyaka yes, acoustic noise, probably from inductor. Anything can be done to prevent it ? \$\endgroup\$
    – 7E10FC9A
    Commented Dec 18, 2021 at 17:55
  • \$\begingroup\$ What operating frequency did you mean to achieve. At the moment it looks about 7 kHz. \$\endgroup\$
    – Andy aka
    Commented Dec 18, 2021 at 17:56
  • \$\begingroup\$ I have no direct control over switching frequency because MAX1771 use internal SR latch as oscillator, and the datasheet is a bit vague. My intention is 12V to 170V boost convertor with max 50mA load. \$\endgroup\$
    – 7E10FC9A
    Commented Dec 18, 2021 at 18:27
  • \$\begingroup\$ @winny The chip always operate in DCM. Is there anyway to force frequency up to increase gain ratio ? \$\endgroup\$
    – 7E10FC9A
    Commented Dec 18, 2021 at 18:36

1 Answer 1

2
\$\begingroup\$

From the scope shots, period seems to be around 110µs, or 9 kHz. So it will make noise at this frequency, either from the ceramic caps, magnetostriction in the inductor, or thru-hole components vibrating due to magnetic field.

These whines are annoying... To get rid of it, you can increase frequency to ultrasound like 25kHz, or higher if you don't want to bother dogs and cats. Since this is a variable frequency converter, you can do this by decreasing peak inductor current. For example you could use a higher value for R1.

Losses will increase slightly.

Frequency will also depend on load current. So make sure to test it with the actual load it is supposed to power. If this draws more current than the load you used for testing, then maybe you won't hear any whine.

\$\endgroup\$
4
  • \$\begingroup\$ It's tested on actual load about 10mA. maybe this deserve another question, but can you explain how the SR latch, voltage and current feedback control the frequency ? \$\endgroup\$
    – 7E10FC9A
    Commented Dec 18, 2021 at 22:44
  • \$\begingroup\$ What this chip does is simple: when output voltage is too low, turn on the MOSFET, when current in sense resistor is too high, turn off the MOSFET, then repeat. So peak inductor current is always the same, on-time depends only on inductor and input voltage, each pulse sends one unit of energy from input to output, and off-time depends on load current. So frequency will be proportional to output current. \$\endgroup\$
    – bobflux
    Commented Dec 18, 2021 at 23:21
  • \$\begingroup\$ Thanks. If I choose a smaller inductor, the peak inductor current will be reached quicker right ? since $$I_{max} = \frac{ V_iD}{fL}$$ \$\endgroup\$
    – 7E10FC9A
    Commented Dec 20, 2021 at 12:50
  • \$\begingroup\$ That is correct. It will also increase frequency. \$\endgroup\$
    – bobflux
    Commented Dec 20, 2021 at 17:09

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.