Dear TI support,
I have a small 2.1 channel Home Theater (HTIB), that uses two TAS5611 chips. One for the front channels and the other for the sub. They work in master/slave mode that is the front amp is the master and the sub is the slave.
The problem is the whistle (teapot like) sound that is occasionally heard from the sub when playing music. The whistling stops after few sec when the music stops. It sounds like an AM interference, but not sure.
The front chip uses a 10K resistor on the FREQ_ADJ leg, but there is no whistling on the main channels. How can I avoid the whistling on the sub in this configuration ? Change the resistors to 20K for lower freq would help ?
Thanks,
Zoltan
Hi Zoltan,
Is your audio source AM? If it is AM interference, the noise should go away if you change to a different AM band.
Are you using PBTL for the sub? Does it have good differential pair like layout for the inputs to minimize the common mode noise? Maure sure that there is sufficient decoupling caps for VDD.
reg,
Paul Chen
Audio Applications Engineer
Dallas TX USA.
Hi Paul,
No there is no AM source on the HTIB. It has multiple inputs, analogue and digital (spdif) as well, but it doesn't matter which input is used. The amps are in 2ch BTL mode (M1,M2 and M3 grounded). The sub amp works in BTL as well and I just wonder why they did not use the PBTL mode for higher power. Anyway the second channel on the sub amp is unused, but it's powered up.
The whistling doesn't always present, usually appears after few minutes of playing music from any source. I can email the schematics, but since it's a commercial design, I wouldn't post it here.
Zoltan,
We can work this issue through email if you share your email address.
Paul.
I have sent you the amplifier schematics to your public e-mail and have also updated my profile with the e-mail info. Please let me know if there is anything that can be improved or changed to the original design and if you need more drawings from other parts (Power Supply, board layout, etc.).
In the schematic the main amp is the same as the sub execpt the C and D channels are not used in the sub. I wonder if the noise is due to the degradation of the clock signal to the slave sub. I would suggest the following -
1. Observe GVDD and PVDD voltages on the scope and see if they are still stable when the noise is occuring.
2. Make sure that the noise is not from the sub audio source (SW+/SW-). Anyway to feed the main amp audio source to the sub for testing?
3. Configure the sub as Master and main amp as slave. I'm not sure if it is possible for you to cut the trace that ties the FREQ_ADJ pin to VREG.
4. If (3) can not be done, observe the clock signal at OSC_IO of the sub. See if the clock is degraded when the noise is occuring.
5. Some of the filter component values are not specified in the schematic. RC snubbers on the output pin side are not needed (e.g. R746/C773).
6. Have you tested it using a pure sine tone rather than music?
Looks like I really need to get (borrow) a scope to look for the source of the noise...
Re. 3, I would like to keep the board intact as much as possible, but changing R751 (10Kohm) wouldn't hurt, I think. In addition on some other models I have seen a little cicuit that switches between 10 and 20K on FREQ_ADJ. I guess it was designed to avoid unwanted interference, but not exactly sure how does it work (what detects if there is noise on the output).
Anyway, will get back with my findings once I got a scope.
Thanks again,