1
\$\begingroup\$

My prototype 5 W audio amp board based around a (pulled) Rohm BA532 amplifier IC (datasheet) worked just fine. It was built on a section of strip board and was based on the following schematic:

enter image description here

This is essentially this circuit with some slight component value mods based on what I had to hand at the time. The most significant is the value between pin 7 and 8 (47 pF instead of 220 pF) and pin 3, which uses 100 nF instead of 10 nF. I see other schematics which use values close to 47 pF for 7,8 and leave pin 3 unconnected. Sadly, there is no reference schematic in the datasheet.

Satisfied with the results, I then created a PCB using exactly the same components. But now the amp is motor-boating at a low frequency when the input isn't floating.

So, with nothing connected to the amp, it's fine. However, if I connect a POT and ground the signal input or just ground the input directly, I get the instability. I double checked the connections and everything checks out and the amp actually works, it's just unstable.

So, to my question: looking at the datasheet IC circuit diagram, and in particular the compensation pin (3) and the phase (margin?) pins (7,8), how should I change the values to increase stability?

enter image description here

My marginally educated guess is that increasing the capacitance at pin 3 increases the high frequency roll-off into the power amp section, so that's probably the way to go there. And increasing the capacitance between 7 and 8 increases the NFB from power to VAS stage, so maybe that's the way to go there. But what do I know.

Any advice?

\$\endgroup\$
3
  • \$\begingroup\$ Motorboating is a low-frequency problem, while compensation suggests a high-frequency problem. Sometimes motorboating or squealing is a layout problem related to speaker current paths. Other times it is a power supply problem that might be solved by vastly increasing that 220uF VCC bypass capacitor - just be careful where its ground-side is connected - try a few different GND points. \$\endgroup\$
    – glen_geek
    Commented Mar 2, 2023 at 15:05
  • \$\begingroup\$ The circuit you copied uses a huge 220 Farads (maybe?) capacitor but your capacitor is much smaller. Why not use a Name Brand IC (Texas Instruments) that has a datasheet and a recommended circuit? I think Motorboating is because the supply voltage has weak current and its voltage is bouncing up and down. \$\endgroup\$
    – Audioguru
    Commented Mar 2, 2023 at 16:24
  • \$\begingroup\$ @Audioguru Why not use a Name Brand IC (Texas Instruments) that has a datasheet and a recommended circuit? Because that would be too easy:-) \$\endgroup\$
    – Buck8pe
    Commented Mar 2, 2023 at 17:17

1 Answer 1

2
\$\begingroup\$

now the amp is motor-boating at a low frequency when the input isn't floating

It's quite likely that the "motor-boating" is caused by an inadequate PCB layout where the main power feed current is flowing in tracks that the input circuits use. Power feeds should go straight to the main powers pins (10 and 2), with pin 2 (GND) being absolutely critical. Any other connection to GND (such as from the inputs) must tee-off from pin 2 and not anywhere else.

This is called star-pointing and here's an example of poor star-pointing in an LM386: -

enter image description here

Image from here.

If there is a spurious connection to GND through equipment earthing, this is also a big no-no.

how should I change the values to increase stability?

More than 50% of power amp instability problems come from bad grounding and not usually component values.

\$\endgroup\$
5
  • \$\begingroup\$ Also try adding some power bypass capacitors near the chip; 0.1µF and 10µF may help. \$\endgroup\$
    – rdtsc
    Commented Mar 2, 2023 at 12:25
  • \$\begingroup\$ Unfortunately, you may be right. I had a look at my pcb layout and even though it uses a ground fill across the board, the connection to pin 2 (Gnd) is certainly not a star! \$\endgroup\$
    – Buck8pe
    Commented Mar 2, 2023 at 13:11
  • \$\begingroup\$ You may be able to prove this by directly attaching your 0 volts power feed AND your return speaker wire directly (as close as possible) to pin 2 @Buck8pe \$\endgroup\$
    – Andy aka
    Commented Mar 2, 2023 at 13:26
  • \$\begingroup\$ Thanks @Andy aka, I'll give that a shot later. Re-spinning the board is sadly not trivial. I'll include rdtsc's low esr decoupling strategy also (but not together, obviously). \$\endgroup\$
    – Buck8pe
    Commented Mar 2, 2023 at 14:03
  • 1
    \$\begingroup\$ Improving ground connections fixed the problem. Thanks for all the advice. \$\endgroup\$
    – Buck8pe
    Commented Mar 6, 2023 at 8:11

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.