TAS5613/TAS5630 - 70 Volt Audo

Our immediate need is to deliver audio for a 25VAC speaker loop.  In our loop of speakers you will find each speaker with an autotransformer on the front end.  All speakers are eight ohms but the audio level of each speaker is attenuated with its own transformer that is tapped for a volume that is acceptable for a particular acoustic environment.  What this really means is that under full load, we need to deliver 35.4 volts peak to peak into a forty ohm load.  This results with 16 watts being delivered in the extreme case.  In order to accomplish this we have to keep in mind that amplifiers are unable to go rail to rail and the power supply must be set to a level, allowing audio to be delivered without clipping or distortion.  Efficiency is also a main concern.  The environment for the amplifier would be a closed box without the ability to bring in cooler air.  We would also like to limit the size and weight of a heat sink.  In an ideal world, delivering 16W of power without a heat sink would be preferred.

We were attracted to the product offering of Texas Instruments because they had amplifiers that worked with a single rail supply.  Our product works with a 24V system but a boost power supply can be designed delivering the required power in an efficient manner.

The optimum design would result with our load being connected for a single BTL (bridge tied load) operation.  With a load configured in this way, we should be able to achieve better efficiency and omit the requirement for coupling capacitors.  In summary, the optimum analog amplifier would have the following characteristics:

1)  Be able to deliver 25VAC (RMS) @ 16W (with a 40 ohm load)

2)  Operation with a single rail power supply with a minimum voltage rating of 50 VDC

3)  Output able to provide for a single BTL load (essentially being a mono amplifier driving a full

      bridge output)

4)  A single channel audio input will be provided

We currently have two demo boards that we are evaluating, each missing our target in a different way:

1) TPA3123D2EVM

     a)  The TPA123 has power supply limitations, limiting the voltage swing of the delivered audio

     b)  The amplifier runs efficiently so a heat sink is not required for this power level

2) TAS5630DKD2EVM

     a) The TAS5630 has more output circuitry than what is required.

     b) The sweet spot for efficiency is not within our power range

     c) Requires heat sink when delivering 16W

Regards,

Anthony

There was a mistake made in my last posting.  The 25VAC (RMS) converts to 70.7 Volts peak to peak. 

Magic is  going on in our lab.  The TAS5630 was configured in a full bridge configuration with the power supply level set at 45 VDC.  (The output was configured for a single BTL load.)  Audio was observed with the oscilloscope at the 70 volt peak to peak level.

What’s worth noting is that the minimum voltage for our power supply was 45 VDC.  The temperature of the IC increased as we went beyond this level while keeping the audio input constant.  When the power supply voltage was increased to 50 VDC, the IC temperature increased 40 degrees Centigrade followed by an over-temperature shutdown of the TAS5630.  These observations were made after the heat sink was removed.

Regards,

Anthony

Hi, Anthony,

In these types of applications, engineers typically use a step-up transformer on the output of the amplifier, and then a step-down transformer on the loads. It looks like you are trying to eliminate the step-up transformer. I can't see an obvious problem that won't let this work, but it is a little different than "normal."

You are going to need to use a heat sink of some sort in this application. As you raise VDD, the switching losses increase, reducing efficiency, and increasing power dissipation.

We also depend on the grounding of the heat sink to ground the thermal pad on the IC which is connected to the silicon die inside.

But, it does indeed sound like magic is happening!

Just don't let the magic smoke out - the TAS5630 has a prescribed start-up sequence that you MUST follow, otherwise, you will let the magic smoke out.

-d2

Don:

Thanks for your response.  I found it enlightening with a humorous twist.

What I was really hoping for was a topology recommendation.  As stated previously, we found some characteristics of the TAS5630 to be unacceptable. . .

2) TAS5630DKD2EVM

     a) The TAS5630 has more output circuitry than what is required.

     b) The sweet spot for efficiency is not within our power range

     c) Requires heat sink when delivering 16W

In my search for alternate choices within the TI family, I found the following possibilities. . .

1)       TAS5111A

2)     TAS5611A

3)     TAS5261

My main concern was finding a device that would permit a power supply rating of 50 VDC.  I wouldn’t be surprised if a heat sink of some sort was required to provide the required convection cooling.  We can only dream of using an etch ground plane to dissipate heat.  We would have to increase efficiency dramatically, however.

I would hope that these devices would perform better from a thermal perspective with our target output parameters. . .

1)       Be able to deliver 25VAC (RMS) @ 16W (with a 40 ohm load)

2)       Operation with a single rail power supply with a minimum voltage rating of 50 VDC

3)       Able to provide for a single BTL load (essentially being a mono amplifier driving a full bridge output

4)       A single channel audio input will be provided

My selection of devices all required a PWM input, however.  My current dilemma can be summarized with the following questions:

1)       Are any of my three candidates a better choice to replace the TAS5630?

2)       Can we obtain better efficiency for our application?

3)       If these devices are better candidates to minimize hardware, do you have a recommendation for a device that would convert an audio input to the required PWM protocol.

Regards,

Anthony

The TAS5615 may also be a viable candidate.  This device takes an analog input so in my situation there would be no need to provide PWM conversion.

Could I expect better performance with this part?  This part appears to be more geared for our intentions.

Anthony

Hi, Anthony,

TAS5615 is NRND - Not Recommended for New Designs - probably not a good choice.

I think you've boxed yourself in, and TAS5630 is your only way out.

While the power stages might appear appealing, with these, you will need to add an I2S to PWM conversion IC (like TAS5086) which requires a micro to set-up. And, assuming your input signal is analog, then you'll need one of our A/D converters (PCM1808?) to take analog to I2S. It starts getting ugly really fast.

Heatsink or transformer. Pay your money and take your choice...

-d2