Why the diode in a wall wart hum buster?

You could make the tiny effort to work it out, too, but since we have to spell it out for you: A typical regulator has a noise density in the several hundred nV/rtHz, whereas a cap multiplier can be in the single-digit nV/rtHz ballpark.

Sometimes it matters.

Jeroen Belleman

Besides the noise level, as George and Jeroen pointed out, the PSR of a good cap multiplier is much, much better than a regulator's, especially at SMPS frequencies.

For instance, an LM7815's typical PSR is less than 60 dB even at 120 Hz, and much worse at 150 kHz. A good cap multiplier can get you > 140 dB at SMPS frequencies (>10 kHz, say). You have to watch out for C-E capacitance and Early effect to get it that good, but it isn't hard to do.

Cap multipliers are super-important with discrete circuits, which typically have poor PSR on their own. They're a good match for op amps too--op amp PSR is excellent at low frequencies and poor at high frequencies, so the combination is good at all frequencies.

Cheers

Phil Hobbs

A supply with significant ripple is unregulated.

NT

You saw right through my sugar coating from the get go, didn't you? The

You made me curious enough to open the wall wart with my dremel. Inside, (as you predicted) it contains a transformer, a full wave rectifier, and a reserve cap. It's guts are strikingly similar to:

FWIW, the 470 uF and 0.1 uF cap combo gets mentioned further down in that article. Some people in the group jump to the wrong conclusions, so allow me to embellish my schematic to make things clearer.

o----->|--------o------------o---o Vcc | 1N4002 | | | | | X--o + | | stereo op-amp

A heat shield is permanently attached to the top of the stereo op-amp so its part number is unknown. The stereo speakers contain an IBM SKU, which is 02K0322.

So we now arrive at the interesting part of my followup, which requires some sleuthing. Why did IBM put a diode in the positive supply path?

Did IBM put it there for RPP? It strains my suspension of disbelief that company that's too cheap to put a regulator in their wall wart decides to splurge in other areas. But, you never know. There's a lot of divisions in IBM and each one "plays jazz" its own way.

The diode actually comes in handy for my Altoid jelly bean mixer, which shares wall wart power with the speakers. But IBM certainly didn't put it there for that reason.

There's one more piece to the puzzle. In additional to the wall wart, a power cable with 2.1x5.5mm male plugs on each end comes with the speaker. Perhaps IBM intends for the power cable to connect the speakers to an optional, ?old school?, common power source? For that particular case, the diode makes sense.

Thank you,

You saw right through my sugar coating from the get go, didn't you? The

You made me curious enough to open the wall wart with my dremel. Inside, (as you predicted) it contains a transformer, a full wave rectifier, and a reserve cap. It's guts are strikingly similar to:

FWIW, the 470 uF and 0.1 uF cap combo gets mentioned further down in that article. Some people in the group jump to the wrong conclusions, so allow me to embellish my schematic to make things clearer.

o----->|--------o------------o---o Vcc | 1N4002 | | | | | X--o + | | stereo op-amp

A heat shield is permanently attached to the top of the stereo op-amp so its part number is unknown. The stereo speakers contain an IBM SKU, which is 02K0322.

So we now arrive at the interesting part of my followup, which requires some sleuthing. Why did IBM put a diode in the positive supply path?

Did IBM put it there for RPP? It strains my suspension of disbelief that company that's too cheap to put a regulator in their wall wart decides to splurge in other areas. But, you never know. There's a lot of divisions in IBM and each one "plays jazz" its own way.

The diode actually comes in handy for my Altoid jelly bean mixer, which shares wall wart power with the speakers. But IBM certainly didn't put it there for that reason.

There's one more piece to the puzzle. In additional to the wall wart, a power cable with 2.1x5.5mm male plugs on each end comes with the speaker. Perhaps IBM intends for the power cable to connect the speakers to an optional, ?old school?, common power source? For that particular case, the diode makes sense.

Thank you,

Nothing wrong with 140 dB of rejection, but what good is it? We are talking about an audio amplifier in a ham radio. Do you expect there to be 140 dB of headroom in the signal path up to the audio amp?

I haven't read about the Early effect. Is that like the miller effect, but between the CE pair?

"Super-important"? If the op amp has 80 dB of PSRR across the audio band, I would be more than happy with that for a typical audio circuit.

I may have missed the context of this circuit. Is this a part of a powered speaker? I assume this circuit is built into the speaker and powers the built-in amplifier?

Yes, it's part of a powered speaker. My schematic captures the front end of the power circuit from the amplifier built into a pair of IBM SKU

Thank you,

So many of those statements I've no idea what they mean.

NT

What is RPP?

It's what makes the collector curves slope downwards towards zero. It looks like a resistor between collector and emitter, which is what causes the leakage.

Not all of us build audio circuits. A nanovolt of noise on a photodiode bias circuit makes a big difference. Same with some PLLs.

Cheers

Phil Hobbs

Ok, thanks.

Sure, context makes all the difference.

Reverse Polarity Protection. It protects the circuit when you replace the wall wart that came bundled with the speakers with a different wall wart that just happens to have its polarity reversed.

Until now, RPP was one of two proffered theories for the diode. The other theory proposes that IBM put the diode into the circuit to prevent backfeed into yet another device connected to a common power bus.

Then there's my own latest theory. My own theory has it that IBM put the diode into the power filter section of their circuit to actually, you know, filter out noise. LOL.

My ears hear a difference. Background noise becomes louder when the diode is removed. Of course, right proper noise measurements must now be recorded to defend my theory from any allegations of audiophoolery being leveled against me. :0)

Thank you,

Perhaps the amplifier gets noisier when supply voltage is higher?

piglet

That seems the most likely scenario. :0) As mentioned earlier in the thread, the speakers share the power bus with a newly minted device of mine known as the "Altoid jelly bean mixer." So, if nothing else, the diode comes in handy to prevent backfeed from the mixer into the speaker stereo amp and vice versa.

There's also a power cable bundled with the speakers that lends weight to the backfeed prevention argument. The cable contains a 5.5mm X 2.1mm male plug on each end, which hints at a shared power bus.

Thank you,

You may have missed the wall wart input to the circuit. The diode will "see" ~20V or less.

Ed