Fun with a Philips TDA1517P chip?

[AdamThorne]

Hey guys. I'm going really low-fi on this one, and hoping to learn a bit along the way. I've come across some cheapo computer speakers, and after breaking them open found the amp for the unit. The central bit is a Philips chip - TDA1517P, a datasheet for the chip is found trivially with google.

I was thinking I might build up an amp around this chip from the ground up as an exercise. I figure replacing the power supply alone will help a bunch (12v 1a wallwart stock). But apart from that about all I know is how to follow the reference circuit. Can any of you chip amp guys suggest improved implementation ideas? Or available resources for this type of thing beyond the datasheet I found?

Thanks, all!

[DigitalJunkie]

Well,you can use a nice stiff (regulated?) power supply with lots of filtering capacitance(cap multiplier?),attach it to a decent heatsink,and use decent quality components(no need to go overboard though).

I built a simple stereo amp with a pair of TDA7350 chips.Just a simple P2P setup,circuit straight from the datasheet (bridged config.) with some 'salvaged' film caps,and lots (30,000+uf!) of filtering capacitance.It runs from a single +12-15V supply. (currently using my 12V battery bank+solar panel )
The sound is actually fairly good,for such a "quick N dirty" project.
The bass response is impressive for such a low-power amp,probably due to the large amount of local filtering capacitance.
TDA7350 12V chipamp

[AdamThorne]

Ok. So I'm looking at my reference design on the datasheet. There are caps on the output. 1000uf.

Hmmm, 1000uf is more capacitance than the fancypants audio caps offer. And it's carrying the amplified signal, so I'm going to need ... uh... I don't know exactly what spec for max voltage. The same as the voltage supplied to the chip? That makes sense, but things I've thought made since have turned out wrong in the past... Also the data sheet gives an amp cutoff at like 45 hz and says it's externally fixed, so I'm thinking of upsizing the input and output caps. So I need to choose a good output cap.

Parts Express has a 500uf, 100v non-polarized part. I could double (or triple) it up...

Digi-Key has Mica and Teflon caps like part number 338-1325-ND. But I don't know what the audio charecteristics of mica / teflon are.

Digikey also has panasonic electrolytics that I understand are good in power supplies...

Also, the ref design has a little cap (100 nF) on the power input, and then looks like it has a big stiffening cap (2200 uF) optionally there with it. Does implementing the big cap mean I can skip the little one? Or does the little one continue to give some sort of bypass benefit?

Thanks for help!

DigitalJunkie - Your TDA7350 project looks like just the sort of thing I was thinking of accomplishing. Except you started with a nicer chip. The 7350 datasheet looks to be much more informative than the 1517 sheet. Perhaps if this project turns out and I start thinking I would like more than 2 x 6 watts
I'll go looking for some of those 7350 chips... Nice work!

Adam

[Photonicube]

The 100nF cap is ceramic it is there for better High frequency ripple rejection. The 2200uF is electrolytic and is the primary bulk capacitance.

[AdamThorne]

I see. Thank you for the tip!

-A

[jives11]

Hi Adam , I'm following the path you created here. I opened up a pair of cheap PC powered speakers and extracted the drivers, which may go into a pair of needles when I get around to making them. I also thought about recycling the chip amp board that I found inside. Mine has a TDA1716P In what looks like the sample design implementation. However before it is an LM324N chip and a lot of green mica caps. I'm wondering if some frequency shaping is being done here to try to reduce the averse effects of the plastic enclosures ?

Did you reuse your chip amp ?

[AdamThorne]

In that case, my best advice is take care extracting the chip! I wired mine up according to the spec sheet, and upon applying power it went kablooie. I figure I probably melted something important trying to remove the chip, as the package had a set of pins for heat dissipation and they were all suck in a big solder blob. I had to heat the whole thing up pretty good to get it out. On the other hand, maybe I just mis-wired it.

The implementation that I removed mine from also had an op-amp chip (I forget which one) in there. The whole thing was more complex than the reference design, it wouldn't suprise me if they were doing some sort of response shaping. Or maybe also some buffering around the volume knob. I couldn't say for sure if you'll want to save it for use in your next implementation b/c like I said, soon as I went to try mine I let the magic smoke out.

Good luck, let me know how yours turns out!