TDA car-audio chip amp: purpose of output caps

[macboy]

speakers with a weird crossover network that has a cap in series with the woofer anyways, which is why removing the cap from the amp works?

It would be odd for a woofer to have a serial capacitor (only). If the serial capacitor is in parallel with an inductor, then the same problem persists.

Serial capacitor (only) for a tweeter makes sense, however.

I don't know enough about passive crossovers but I've always seen them with a LC lowpass filter with the cap in parallel at the woofer.

The speaker would defeat the capacitor in parallel.

Speaker voice coils have (commonly) 8 ohm impedance, which for a woofer, really does measure as roughly 8 ohms resistance. The purpose of the parallel capacitor (after the inductor) is to have a lower than 8 ohm impedance above the crossover frequency, reducing the current that goes into the speaker, thus providing an additional pole to the filter.

[wiss]

speakers with a weird crossover network that has a cap in series with the woofer anyways, which is why removing the cap from the amp works?

It would be odd for a woofer to have a serial capacitor (only). If the serial capacitor is in parallel with an inductor, then the same problem persists.

Serial capacitor (only) for a tweeter makes sense, however.

I don't know enough about passive crossovers but I've always seen them with a LC lowpass filter with the cap in parallel at the woofer.

The speaker would defeat the capacitor in parallel.

Speaker voice coils have (commonly) 8 ohm impedance, which for a woofer, really does measure as roughly 8 ohms resistance. The purpose of the parallel capacitor (after the inductor) is to have a lower than 8 ohm impedance above the crossover frequency, reducing the current that goes into the speaker, thus providing an additional pole to the filter.

And to ensure that the driver always sees a low impedance, this will reduce out of band distortion compared to a single series inductor.

[Zero999]

mikerj  urinated ...

You are quite correct,

 
TRYIN' TO PATRONIZE? ..... PFFFFTTTTTT!!!     

...
the output of the amplifier itself does not need to swing negative, but the signal applied to the speaker absolutely must swing negative; the net DC component applied to the speaker needs to be kept to very small values, preferably zero.  It is the job of your "emergency"  caps to remove the DC bias from a conventional power amp run from a single supply rail.

I appreciate you are finding this entire topic very difficult to understand, but the capacitor effectively converts a 0 to Vcc swing to a -Vcc/2 to +Vcc/2 swing.

pssst ...the caps are there for emergency DC protection (the amp/speaker will work, with single supply, w/o those caps)[/size] ....the caps are there for emergency DC protection (the amp/speaker will work, with single supply, w/o those caps) ... the caps are there for emergency DC protection (the amp/speaker will work, with single supply, w/o those caps).

No, the capacitors are necessary. Take a look at the schematic when there's no audio signal. The outputs of the TDA7273 are biased at half the supply voltage. Without the output capacitors, the current through a 3 Ohm speaker will be a steady 2A, causing it to overheat. Hopefully, the IC's internal protection circularity will kick in and shut it down before any damage is done to either the speakers or amplifier.

[Alex Eisenhut]

speakers with a weird crossover network that has a cap in series with the woofer anyways, which is why removing the cap from the amp works?

It would be odd for a woofer to have a serial capacitor (only). If the serial capacitor is in parallel with an inductor, then the same problem persists.

Serial capacitor (only) for a tweeter makes sense, however.

I don't know enough about passive crossovers but I've always seen them with a LC lowpass filter with the cap in parallel at the woofer.

The speaker would defeat the capacitor in parallel.

Speaker voice coils have (commonly) 8 ohm impedance, which for a woofer, really does measure as roughly 8 ohms resistance. The purpose of the parallel capacitor (after the inductor) is to have a lower than 8 ohm impedance above the crossover frequency, reducing the current that goes into the speaker, thus providing an additional pole to the filter.

This is true! I looked up some woofer datasheets and checked the wiki for the T&S parameters, turns out voice coils are designed to have a DC resistance (Re) of 80% of nominal impedance. How little I know!

Call it 6 ohms, so with a 12V TDA sitting at 6V, without that "emergency" cap, you'd get about 1A at 6V = 6W static power dissipation, which is just about the limit of that chip anyways. So while not quite a "short circuit", it also means there's no useful headroom for actual musical signal there, which is pretty much a practical short circuit. And you get a hot speaker for no audio output.

This has been a weird and wonderful thread.

[David_AVD]

This has been a weird and wonderful thread.

The troll is probably safely back under his bridge now. 

[anman]

A good board TDA7293 for car audio.

[Richard Crowley]

A good board TDA7293 for car audio.

Sure, if you use a PAIR of chips for each speaker, you can use them in "bridged" mode. Where BOTH sides of the speaker float at 6V during quiescence.  But that was not the original question.

[salbayeng]

What Richard said, and
I don't think it is necessarily a "good car audio board"

(a) They have fitted smaller electros than the footprint on the PCB, so they are probably cheap 85C 1000hr caps, that will deteriorate in a year or so with a  typical car parked in the sun. They should  really be 105C 5000hr rated, and operating at < 60% of rated voltage.

(b) They have 250V rated film caps , wtf! , this is to appease the golden ears brigade, seriously a 63v polypropylene cap would suffice, but doesn't have that valve audio look!.

(c) The TO-220 regulators? have no mechanical support, so typical car vibrations would fatigue the legs and they would fall off.

(d) This board requires a split power rail of +/-35V  so not useful with a car's single 12v supply. You would need some kind of auxiliary power supply.

(e) to get the purported 350W out you need several 1000uF caps scattered around, that need to have ripple current ratings of several amps each, the caps on the board don't look big enough to do the job.

(f) The green connector with the output (in middle of PCB) is labelled "IN GND" ,  not "SPKR+ SPKR-"  , if they can't get the simple things right, the rest of the design is suspect.

[Tac Eht Xilef]

if they can't get the simple things right, the rest of the design is suspect.

What, you expect good design sense from someone who re-awakens a necro-thread with a thinly-disguised Aliexpress referrer link? That's a spammer...