0
\$\begingroup\$

I'm planning to use STA540 in a guitar combo, because it's one of a few chip amps that can give me about 15W with 12V supply and 4ohm speaker.

My problem is that the chip has four amplifiers and I will be using only two of them (bridged). What should I do with the other two?

My ideas:

  1. Ground the inputs and leave the outputs hanging.
  2. Use the second pair in parallel with the first one. This probably won't help me push more power to the speaker because I think I will be limited by input voltage and not the IC itself and I don't see any other advantages of this
  3. Find a smaller IC -- but I already tried that and didn't succeed....
  4. Use them to drive headphones. Is this realistic, or would the headphones explode?
\$\endgroup\$
  • \$\begingroup\$ This is a dual or quad piece. Take a look in the datasheet for dual st.com/web/en/resource/technical/document/datasheet/… \$\endgroup\$ – mcmiln Jan 11 '16 at 20:14
  • \$\begingroup\$ I'm planning to use it in the bridged mode (figure 5 in the datasheet), but i only need a mono amplifier -- in_r and out_r will be unused in my case. \$\endgroup\$ – cube Jan 11 '16 at 21:30
  • \$\begingroup\$ Use two 8 ohm speakers. \$\endgroup\$ – Andy aka Jan 11 '16 at 21:40
  • \$\begingroup\$ I this case, the only thing connecting each input is the SVR, which means you could either ground the other inputs or just leave them floating. No big either way. \$\endgroup\$ – mcmiln Jan 11 '16 at 21:44
  • 3
    \$\begingroup\$ Would recommend NOT leaving unused inputs floating. \$\endgroup\$ – Brian Drummond Jan 11 '16 at 23:51
0
\$\begingroup\$

Since the inputs are designed for capacitive coupling, floating the inputs (and outputs) is okay because the chip provides internal resistors that establish the DC biasing point at the inputs.

Grounding the inputs will drive the unused outputs to the supply rails, which may increase (or decrease) supply current. Grounding inputs will likely trip the clipping detection (pin 10) even when the active channels are not distorted. Grounding inputs may also shift the SVR voltage, which could undesirably shift the quiescent DC voltage of the active outputs.

You could try both floating and grounding the inputs, and measure which results in the lowest supply current (to minimize wasted energy). If grounding inputs, check for significant shifting of the DC output voltage.

Paralleling outputs likely won't work here because the voltage gains of the amplifier channels will not exactly match. Because the active feedback of each amplifier tries to correct any output-voltage error, the amplifier channels will just fight each other causing excessive power consumption and likely distortion.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Thanks for the answer. Do you think I could use the outputs to drive headphones? \$\endgroup\$ – cube Jan 12 '16 at 9:07
  • 1
    \$\begingroup\$ You absolutely should not leave the inputs unconnected, because it will cause oscillation and other undesirable effects. Just build a resistor feedback network around the unused channel, essentially a dummy channel, and connect the input parallel to the real channel. \$\endgroup\$ – ilkhd Jan 12 '16 at 11:37
  • \$\begingroup\$ @cube: Headphones have much higher impedance than speakers (30 to 60 ohms is typical), so an amp that can drive a 4-ohm speaker at 15 watts would drive headphones at 1 to 2 watts. This level is okay for many headphones, but almost certainly too high for your ears. To reduce the watts by a factor of four, you could drive the headphones from just one amplifier output through a coupling capacitor instead of connecting the headphones across the full bridge voltage (that the speakers use). \$\endgroup\$ – Fred Schleifer Jan 12 '16 at 17:52
  • \$\begingroup\$ @cube: Many of the comments and answers here warn against floating the inputs. This is absolutely true for standard op amps, but the integrated amplifier you are considering already includes internal feedback and biasing resistors. These resistors should prevent any problems with floating inputs. There is a very small chance that the unused inverting channel (A2 or A4) will oscillate without an AC input load, but tying a small 100 nF capacitor from the input to ground would solve that problem. But this chip looks quite bulletproof, so I suspect the designers allowed for open inputs. \$\endgroup\$ – Fred Schleifer Jan 12 '16 at 18:09
1
\$\begingroup\$

No, floating the op amps is not a good idea! Random noise can cause the op amp to oscillate between high and low at high frequencies if the inputs are left floating. Attaching the output to VCC or GND can cause excessive current draw and stress the op amp. Driving both inputs to high/low places stress on the amplifiers as well, as it will be driven as strongly as possible.

You'll want to place the op amp in voltage follower mode, attaching the positive input to a divided voltage from VCC.

Refer to this Maxim app note that explains very well how to minimize the impact of uncommitted op amps in your system: https://www.maximintegrated.com/en/app-notes/index.mvp/id/1957

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ If I look at figure 1 in the amplifier's datasheet and imagine that IN3 and OUT3 is floating, it looks very much like figure1 in the article you linked, only the gain is not unity, but some fixed default value. For IN4 I don't see what would happen there, but I assume it will behave similarly. Am I wrong? \$\endgroup\$ – cube Jan 12 '16 at 10:33
  • \$\begingroup\$ @cube Be careful, many audio amplifiers are not unity-gain stale and will oscillate if connected as followers. \$\endgroup\$ – ilkhd Jan 12 '16 at 11:43

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.