Page 1 of 1

Buck converter for CVDD?

Posted: Thu 2019-07-11 19:02
by msat

I'm starting on the design of my PCB using the VS1005, and have been thinking about potentially using a buck DC-DC converter for CVDD, as that seems to be where the greatest power efficiency gain would be over a linear regulator. Maybe I missed it, but I have not seen mention of it in any documents or this forum, so I figured I'd ask before I waste time on it. First off, is it ok to supply CVDD1-3 with an external rather than the internal regulator? Is there anything I need to do to the internal CVDD regulator output to prevent bad things from happening? Is there any specific power-up sequencing I need to adhere to? Anything else I need to be aware of?


As long as there's no deal breakers with externally supplied regulated power, if it comes down to it, I suspect I could put some dummy load on the internal CVDD output if instabilities might result otherwise, at least until the core is up and running so that I can shut off the internal CVDD regulator via the REGU_CF bits.

Re: Buck converter for CVDD?

Posted: Fri 2019-07-12 8:41
by Hannu

This is totally untested, so try and measure.

It's true that best efficiency can be gained on CVDD. However this might have bad effects on audio performance if the switcher noise gets through.

The recipe to investigate this is to set core voltage some working minimum voltage on given clock frequency with setclock.
Then you enable the switcher and drive externally CVDD 0.1 V above what you set it internally.
Don't exceed max CVDD or the chip gets toasted.

Internal regulators don't mind if you drive higher voltage outside. They just start regulating VHIGH to CVDD if the voltage drops below the set value. So you get backup and normally current would flow from the switcher.

I would add crowbars to keep the voltage below max CVDD, double the decoupling capacitors values and drive the switcher's enable pin from VS1005 so that if it is total disaster, I can disable it. Also adding filters on input side of the switcher would bee good idea. That would keep the VHIGH clean and usually it is good idea.

Re: Buck converter for CVDD?

Posted: Fri 2019-07-12 12:57
by pasi
Are you going to use USB?

If you don't require USB support (and don't need 1.65V reference), you could drop AVDD and IOVDD to 2.8V, and VHIGH to around 3.0V. The difference between regulating and DC-DC for CVDD should narrow quite a lot.
As for a clean power-up with a DC-DC CVDD:

You could use IOVDD to enable your CVDD DC-DC. Like Hannu said, as long as a higher voltage is supplied from outside, the internal regulator won't need to supply power. You can even shut down the CVDD regulator from your software for a bit of extra saving.

Re: Buck converter for CVDD?

Posted: Fri 2019-07-12 17:36
by msat
Thanks for the replies!


The potentially bad effects that a buck converter might have on the analog section has crossed my mind. Testing would be in order. I was certainly expecting to have to beef up all the decoupling caps.

It hadn't crossed my mind to tie the CVDD inputs to both the internal and external regulator/converter. If the DC-DC added non-negligible noise that couldn't be practically filtered out, it could potentially give the user the option to select between a power efficiency mode, and a high audio performance mode. I'd rather not, but the option could be there just in case.


I will be using USB, though I have considered bucking VHIGH as you suggested. It would let me eliminate a separate regulator for the SD card, but I was kind of worried about the amount of effort I'd have to expend on ensuring good audio performance, not just within the analog subsection of the VS1005, but some additional analog hardware in my circuit.

Ok, so it sounds like I probably wont let out the magic smoke by using an external regulator as long as I don't exceed the maximum ratings for the subsection I'm powering. Next order of business is a little testing on my HiRes board.