DC-DC converters to power C6748

Hi Dean,

Assigning this question to the appropriate team to get you more information on what DC-DC converters might fit your design.

Best,

Emily

Hi Dean,

Thank you for your question.

Do you have any specific constraints for Efficiency, VOUT Accuracy or VOUT ripple for any of the three rails?

If C6748 doesn´t have any particular constraints, I think that  TLV62568A could be a good compromise, since it is surely a cost effective device, with maximum output current of 1A and the forced PWM operation that guarantees a low output ripple even at low current (at the expense of the low current efficiency).

Just a comment; TLV62568A accepts a recommended maximum VIN of 5.5V (absolute maximum of 6V). Please assure that your VIN doesn´t exceed this value for any operation conditions (for example rail startup, voltage spike coupled externally , ...).  

Please ask again if you need further support.

Thanks,

Stefano

All information in this correspondence and in any related correspondence is provided “as is” and “with all faults”, and is subject to TI’s Important Notice (http://www.ti.com/corp/docs/legal/important-notice.shtml).

Hi Stefano,

If it wasn't clear, the C6748 is a TI DSP.  Full part number is:  TMS320C6748.  So your question about specific constraints is better answered by another TI expert.  I realize this question will span multiple departments within TI, so you'll have to ask around internally to get the correct answer.

I appreciate the comment on VIN for the TLV62568A.  I should have specified in my original post that we have a 6VDC 3A power supply feeding our product, and there is a reverse polarity idiot diode on the front end.  Therefore, VIN for the DC-DC converters will never exceed 6V, and will be running in the 5.2V to 5.4V range.

It will be difficult to kludge a new DC-DC converter solution into our existing PCB.  Therefore, I'll need a schematic review by TI to minimize any mistakes on the next board spin.

I'm looking for a DC-DC converter solution that will safely power the C6748, has low EMI, and low cost.  If the TLV62568A works, then great.  If not, why?  And what other DC-DC converter solution is recommended?  I realize I could simply use the solution provided in the application note.  But, as I've said that solution is 10+ years old.  Surely TI has developed better parts (EMI, cost, efficiency, output ripple, etc) over the past years, and that application report no longer has the "ideal" solution.

Thanks, Dean

Hi Dean,

Thank you very much for the clarification.

I will check internally and let you know which are the recommended Buck converters.

Thanks,

Stefano

All information in this correspondence and in any related correspondence is provided “as is” and “with all faults”, and is subject to TI’s Important Notice (http://www.ti.com/corp/docs/legal/important-notice.shtml).

Hi Dean,

Thank you very much for your patience.

We didn't find any particular constraints on C6748 voltage rails, thus I think that TLV62568A should be a good cost effective solution for your design.

Thanks again,

Stefano

All information in this correspondence and in any related correspondence is provided “as is” and “with all faults”, and is subject to TI’s Important Notice (http://www.ti.com/corp/docs/legal/important-notice.shtml).

Hi Stefano,

Thanks for you answers so far.  Glad to hear the TLV62568A can power the C6748 DSP without an issue.  My new problem is Vin max.  We have redesigned our power protection front end, so the new Vin is 6.2V to 5.5V.  Obviously this will violate Vin Max for the TLV62568A.  So, I'd like to add a DC-DC in front of the TLV62568A parts to regulate the max voltage they see, plus it provides a few extra benefits.  Anyway, I'd like to set the output of this DC-DC to 4.8V.  What I'm finding is some DC-DC converters (not all?) don't like Vout close to Vin.  This is an obvious design parameter when using LDO, but DC-DC converters data sheets don't come out and say "for this part, Vin must be xx above Vout".  I've played around on Webench with a few DC-DC converters that I thought would work, only to find out I'm violating the minimum Vin.

I think I've found a part that will work, but I'm not sure.  Please double check to see if TPS54202 will work for the following requirements.

Vin nominal:  5.5V to 6.2V

Vin absolute max:  6.8V

Iout:  2.1A

Vout:  4.8V

Thanks, Dean

Hi Dean,

Thanks again for your inputs.

Unfortunately I don't have experience with TPS54202, I will forward your request to the appropriate team.

Thanks,

Stefano

All information in this correspondence and in any related correspondence is provided “as is” and “with all faults”, and is subject to TI’s Important Notice (http://www.ti.com/corp/docs/legal/important-notice.shtml).

Hi Dean,

TPS54202 is a 28Vin part with 2A current capability.

For your application, I recommend the TPS563201. Our colleague will comment if he has more suitable device.

Thanks,

Lishuang

The TPS563201 will not work (at least the way I understand things) due to the requirement in the datasheet that ratio of Vout to Vin must be 75% or less.  In our application, Vin minimum is 4.8V.  4.8V / 5.6V = 85.7%.  If I'm wrong, please let me know ASAP, as this part meets every other requirement.

Using the Webench design tool, it selected TPS56340 and TPS563249 that would meet my requirements.  However, when I deep dive into the datasheets, I find the same problem:  Vout to Vin ratio for both of these parts is roughly 60%.  Not sure why the Webench design tool is recommending them, unless my concern about Vout to Vin is wrong.

As I continue to search through TI's offerings, I've found three parts that may work.  Please check if any of these will work for ALL of my requirements (please refer to my previous post for the requirements).  The parts are:  TPS563211, TPS56339, and TPS54302.

Thanks, Dean

Hi Dean,

Due to minimum off time limitation, there is always an allowed maximum duty cycle given a fixed switching frequency. For example,  the minimum off time of TPS563201 is 220ns(which you can find from datasheet), the maximum duty cycle can be approximated by: 1- 580kHz*220ns=83%. Taking considerations of power losses with heavy load operation, the maximum duty cycle should be much lower. That's why in TPS563201 datasheet we recommend the maximum duty cycle should not be more than 75%. TPS563240 and TPS563249 have similar constraints with TP563201. 

For your application, I think you are finding a DC/DC with frequency foldback scheme, which allows switching frequency decreases once longer duty cycle is needed under low VIN conditions. As you mentioned, TPS563211, TPS56339, and TPS54302 have frequency foldback scheme and could work for all of your requirements. As TPS563211 has not been released so far,  you can make a choice between TPS56339 and TPS54302.

Regards,

Jared.