As if there were not enough already posted here about this chip . . . well, maybe there isn't.
Has anyone prototyped using the leaded, through-hole version of this part, i.e., the TDA8920BJ, rather than the SMD version?
Although the stand-up, SIP version is readily available, NXP tech support says they never built a board with it, so they have no data. They regarded the probable inductance between the chip's power pads and any through-hole board's bypass caps to be unacceptably large, so they rejected the idea, a priori.
The anticipated effect is degraded THD, which I suppose must be based on NXP's simulation data. This seems plausible, but since nothing was built and tested, there is no spectral data on file at NXP to quantify it. NXP says its data for the SMD eval board indicate that the chip's prominent harmonics are all odd-numbered, with the third and fifth dominating, and the THD figures are in the 0.02% range, 0.05% max. If the SMD/SIP difference is large enough to goose these same harmonics into audibility, then yuck. FWIW, the increase in EMI is expected to be in the 6 to 10 dB range, still well within acceptability.
So, although it may seem like it was an interesting move to put the chip on the market in a second package that the company's own applications guys say is worthless, maybe there was a sensible rationale, and it's not as bad as they expect.
Thanks, all.
- David
Has anyone prototyped using the leaded, through-hole version of this part, i.e., the TDA8920BJ, rather than the SMD version?
Although the stand-up, SIP version is readily available, NXP tech support says they never built a board with it, so they have no data. They regarded the probable inductance between the chip's power pads and any through-hole board's bypass caps to be unacceptably large, so they rejected the idea, a priori.
The anticipated effect is degraded THD, which I suppose must be based on NXP's simulation data. This seems plausible, but since nothing was built and tested, there is no spectral data on file at NXP to quantify it. NXP says its data for the SMD eval board indicate that the chip's prominent harmonics are all odd-numbered, with the third and fifth dominating, and the THD figures are in the 0.02% range, 0.05% max. If the SMD/SIP difference is large enough to goose these same harmonics into audibility, then yuck. FWIW, the increase in EMI is expected to be in the 6 to 10 dB range, still well within acceptability.
So, although it may seem like it was an interesting move to put the chip on the market in a second package that the company's own applications guys say is worthless, maybe there was a sensible rationale, and it's not as bad as they expect.
Thanks, all.
- David