TAS5719: TAS5719 OSC_res value selection

Hi John

18k 1% resistor is requested in our datasheet. 

this chip do be sensitive to the value of osc-res.

have you take a measure for the actual value of the resistance?

thanks.

jesse

Jesse,

We are using 1%, 18.2k resistors, and they measure within spec. Will you please comment on the range of the chosen resistor?  Will the 16k resistor cause other problems?  Either providing a range or a chart indicating the tradeoffs based on the resistor value would be appreciated.

Please explain the purpose of the resistor in the oscillator.  Thank you

John

hi John

we don't have any study about how the device performance with 16kohm resistor.

I just curious about if using 16kohm with other devices, not the 10 parts, can they work normally or not?

the resistor is for oscillator trimming which is used for PLL circuit.

have you check with ABA test, then we can know whether it is caused by chip or the board.

you can take a good board, and soldering 1pcs from those 10pcs to check whether it can work normally or not.

ps. please pay attention to the solder situation under the chip, for bad soldering under the chip can also lead to this problem.

thanks.

jesse

Jesse,

The 18.2k resistor has been used on all manufactured products, and up until we started to notice the audio issue, there were no issues at all.  The 16k was substituted onto the boards having an issue, and the audio intermittency stopped. During the initial investigation, I removed an IC from a non-production  PCB and observed no solder bonding the thermal pad to the thermal land, which caused me to remove more parts on one of the intermittent boards.  X-rays revealed small solder deposits under the part, but insufficient compared to what TI recommends in the Thermal Pad application note.  This lead me to think the lack of solder resulted in a bad grounding of the part, which resulted in the oscillator not functioning properly.

By the way, this issue is observed is when a HDMI signal is fed into the board, and a speaker is attached to either of the speaker terminals.  The audio is intermittent OR doesn't appear unless you touch the osc_res with a finger or jeweler screwdriver, or unless you touch the trace (with solder mask) leading to the resistor with the same screwdriver, or you touch the IC body near pin 16. Once you remove power from the PCB and reapply it, you need to taouch the board again.

John

Hi John

when soldering is not enough or Pseudo Soldering between this problem do happened before at other customer's side.

this is because all the gnd of the chip should be connected to the power pad so that all GND inside of chip can be connected together.

thanks.

jesse

Hi Jesse,

Thanks for your response.

Please let me restate how I understand what you wrote, "When there is not enough solder (or there are cold solder joints), this problem can occur. " Please clarify "before at other customer's side". I do not understand your first sentence. Do you mean this has occurred at other TI customer sites?

If all of the grounds inside the IC are connected to the PowerPad, then I can understand how insufficient solder can lead to less than optimal grounding of the part.  But is using a 16k resistor to "fix" the problem a good long-term solution (i.e., how low can this value be reduced without affecting the performance of the PLL?). Since I do not know the circuits inside the IC, I do not understand how decreasing the resistor from 18.2k to 16k can compensate for a grounding issue.

Thanks for your response.

Regards,

John

Hi John

usually, we recommend soldering should cover above 50% of the power pad. less than 50% would generate this kind of problem easily.

about the 16kohm, we don't recommend using this resistance for it hasn't been verified, so we cannot say it is also ok for normal case.

About the internal circuit, for this is an old chip, we need more time to confirm how it work inside.

thanks.

jesse

Hi Jesse,

Thanks for your response. This is an x-ray image of the solder deposits under the PowerPad.  X-ray does not indicate whether a bond is cold or if it actually connects the IC to the PCB. If we assume there is a good bond between the two, then a quick drawing indicates there is nowhere near 50% coverage

You commented that you will need more time to confirm how the PLL inside the IC works.  Will you be willing to determine this and report back to us?  As I mentioned, we have placed a 16k osc_res resistor, and it "fixed" the problem.  But we need TI to tell us whether this is acceptable from a long-term functionality perspective.  Thank you

Hi John

I think maybe my answer confused you. Please allow me clear this.

We don't recommend using 16kOhm as your method to fix this problem. 

As you can see in our datasheet, 18kOhm 1% is our recommended value.

I just can help you check the internal circuit, can to see whether any clue can be found.

X-ray can't show actual connection.

why not taking a try re-soldering the chip and check whether the issue still exist or not?

thanks.

jesse

Hi Jess,

Unfortunately, the boards are no longer available to experiment.  As I mentioned to you, our manufacturing partner did remove the TAS5719 on one board, where there clearly was inadequate solder.  They added more solder and reattached the IC to the PCBA.  They stated this did not resolve the intermittent audio issue.  Unfortunately, they only tried this on one of the six boards.  How many times can a part like this (i.e., with thermal pad) be removed and reheated before causing damage?

Also, for future purposes, will you please recommend a removal technique for this particular type of IC?  We have a hot air tool and a preheater.

You mentioned TI does not recommend the 16k resistor. However, is there any design notes for this part that indicates how the 18k resistor was chosen and the recommended range, and possibly the tradeoffs?

Curiously,

John