TDA1543 chips broken?


2005-08-24 10:37 am
Please help me with debugging my TDA1543 DAC. I'm getting some really strange results. I suspect my chips are broken.

This is what I did: I created a tower of 8 chips with glued on heat sinks and used passive I/V. The results where a really distorted sound. I couldn't get rid of it by tweaking the Vref and Vout resistors.

Then I tried a single chip with the same result. No matter how I tweaked the I/V en Vref resistors always distorted sound (tried with supply voltages between 6 and 8 Volts).

So I made a really simple test rig with a 7806 and a chip socket to easily swap I/V en Vref resistors. I took 3 chips out of the tower and measured.

The results the same for all chips:
with Rout = 820 and Rref = 1K: Vref about 1V, Vout = about 1V.
with Rref = infinity (no R): Vref about 2.3V, Vout just under 1 V.
with Rout = 2k7 and Rref = infinty, Vout just under 1 V.

According to this post: I should be aiming for a Vout of about 3.3V.

So I tried a few more values with the same results until I arrived at this combination:

With Rout = 2k7 and Rref = 1k I got some varying results with one chip giving Vout of about 2.3V (and Vref = 1V). The other 2 chips gave large differences for Vout between left and right channel like (3.9v and 2.4V; 3.6 and 2.2V). I decided to continue with the chip without channel imbalance.

Finally Vout above 1V!

By doing the math In the post mentioned above I calculated that Rout = 1k74 and Rref = 4k4 should be about right. Measured results with Rout = 1k8 and
Rref = 4k7: Vref = 2.11V and Vout = 1.02V ??? What is going on? This doesn't seem right ...
Ok, then trial and error again: Rref = 1K. Vout is 3.5V and 2.5V (Rref = 1V) ??? Channel imbalance???

I've not a clue what is going on. Am I doing the wrong measurements? Are the chips broken somehow (I know I did apply 5V as I2S level. This seems high as the datasheet specifies 2V is enough for a 'high'. The datasheet doesn't specify a max I2S voltage)? Did something else wrong?

Please help. I've spent many many hours and still don't have a clue.

Thank you, MArco