Yet another funny Chip-Amp: The Deviant
Hi there,
I have already described a number of amplifiers based on chips not initially intended for audio, mainly voltage regulators:
http://www.diyaudio.com/forums/chip-amps/176052-now-regulator-chip-jlh-amp.html
http://www.diyaudio.com/forums/chip-amps/192934-se-class-regulator-chip-amp-madness.html
http://www.diyaudio.com/forums/chip...oss-tringlotron-regulator-chip-amplifier.html
http://www.diyaudio.com/forums/chip-amps/193214-class-chip-amp-now-complementary-version.html
http://www.diyaudio.com/forums/chip...ator-chip-amp-family-welcomes-new-member.html
http://www.diyaudio.com/forums/chip-amps/175457-just-fun-regulator-chip-amplifier.html
This time, I set my sights on another category of circuits, vertical deflection chips for CRT TV's.
These circuits contain a number of function blocks, including a power amplifier.
In its normal application, it is rather far off from an audio amplifier, being current output and featuring flyback handling circuitry, but it can very easily be converted to an audio amplifier.
To that end, the flyback generator is left apart, and the PA supply is directly connected to the power supply.
The inverting input is available at pin 10, and is connected to the feedback divider formed by R3 and R4.
The DC output level is set by R5.
There is a compensation cap, C3 connected to an intermediate stage of the amplifier. There is also a 220K resistor, but it does seem to be useful for audio; it probably helps control the circuit during the flyback period, and its only apparent effect is to reduce the loop gain.
The inverting configuration is slightly inconvenient, because of the lowish input impedance, and the fact that the amp goes into unity gain when the source happens to be disconnected.
To avoid instabilities in this scenario, C5 provides an HF path to the ground.
Note that the feedback resistors could be made significantly larger without problem: the typical input bias current is only 100nA.
Now, the $1,000 question: does it actually work?
Yes!!!!, and rather well: it provides a good clean output, is well behaved even with a 10KHz squarewave and has a power bandwidth extending to 25KHz, quite OK for audio.
With a 34V supply, the output power reaches 13W at the onset of clipping.
To cope with this power, I have improved on the original heatsinking method: I have made a U-shape copper heatsink link by bending a power bus strip.
It is soldered to the tabs, very close to the case to minimize the thermal resistance, and the underside is pressed against the bottom of the IC, with thermal compound in-between.
In those conditions, the chip has no problem delivering 13W sinus into 8Ω. It wouldn't be prudent to go lower, as the maximum peak current is already exceeded.
Could other function blocks be reused in an audio context?
The oscillator is probably ruled out, because of the positive feedback.
It would be tempting to use the flyback circuit to boost the supply voltage, like in Philips chips, but bits are missing, and it requires an inductive load to operate.
The internal regulator (pin 7) can be used to supply other circuits, like a preamp or a tone control.
The buffer, between 12 and 1 is probably perfectly usable, I will certainly try it.
One thing to be noted: this circuit does not include a short-circuit protection, and I had the occasion to verify it the hard way: an alligator clip went loose, brushed very shortly on something else, and it was gone.
Funny thing, it didn't go dead short, unlike 99.9% of semiconductors: instead, it went open circuit, something must have played the role of a fuse, either a bonding wire or a metallization.
This means that I didn't have time to carry out more detailed measurements, like THD for example, but anyway, it looked quite decent.
In short, a good little chip-amp requiring very little passive components and working surprisingly well.
Note that the TDA1170 is a very old circuit, modern IC's can certainly do better regarding power, bandwidth, and features like short-circuit or thermal protection