Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter

[TheShaman]

Forgot to mention passive battery management board has been tested and works great in standard configuration. Thanks Ian!

[Fridrik]

Originally Posted by qusp

fridrik, i'm not sure these regulators are that suitable for an IV stage, LDOs in priciple and across the board generally rely heavily on their output capacitors to provide for transient current demands ie. their transient response is directly proportional to the output capacitance; by themselves while the noise is good and low, the transient response is usually pretty ordinary.

this is fine for things that have low, or stable current demands that can be supplied by lowZ decoupling at point of load, but IV stages though not having huge current demand (most of them), do have transient current demands that may be enough to deplete local decoupling.

what do you think Ian?

Thanks for your comment qups. I will take this in consideration, But be aware that I'm trying to remplace a LM337T-ND of the original design which is also supposed to be some kind of LDO ;-)

For the moment I will just use the TPS7A4700 to power the FIFO and the SPDIF adapter. Ian do you suggest 2 LDO board to power Fifo+SPDIF ?

[iancanada]

I'm afraid I run out all of the TPS7A4700 PCB. Thank you so much for your very kind donations. I appreciate!

Will combine the shipment for those who has a GBIII order number. PCB shipped today to all of those without a GBIII order number.

Pretty sure I'll supply this PCB again with the Si570 board.

Thanks again

My best regards.

Ian

[iancanada]

Originally Posted by TheShaman

Forgot to mention passive battery management board has been tested and works great in standard configuration. Thanks Ian!

Hi TheShaman,

Good to know that, you did great job! board looks nice, good picture.
Please let me know if you need any support.

Ian

[iancanada]

Originally Posted by qusp

fridrik, i'm not sure these regulators are that suitable for an IV stage, LDOs in priciple and across the board generally rely heavily on their output capacitors to provide for transient current demands ie. their transient response is directly proportional to the output capacitance; by themselves while the noise is good and low, the transient response is usually pretty ordinary.

this is fine for things that have low, or stable current demands that can be supplied by lowZ decoupling at point of load, but IV stages though not having huge current demand (most of them), do have transient current demands that may be enough to deplete local decoupling.

what do you think Ian?

Although 4700 has very low noise density, but it optimized to VCO/XO/RF power supply with very good high frequency performance and can drive low ESR MLCC output capacitors. I still think a well designed shunt reg is better for a I/V. But I might be wrong. Hope somebody can give 4700 a try for analog applications.

Ian

[qusp]

if you look at te step response it confirms what i'm saying and the psrr could be better too (-60 -70dB through the audio frequency range), better than any other LDO i've seen however. its perfectly suitable for how you are using it Ian, it has amazingly low noise, but its not really the thing supplying the load current in your designs, its supplying the capacitors which are supplying the load, so the caps are/should be chosen with this in mind

have you exposed the sense pin on your PCB Ian? or have you just tied it to the output voltage?

nvm i'll look at your schematic....

[iancanada]

Originally Posted by qusp

if you look at te step response it confirms what i'm saying and the psrr could be better too (-60 -70dB through the audio frequency range), better than any other LDO i've seen however. its perfectly suitable for how you are using it Ian, it has amazingly low noise, but its not really the thing supplying the load current in your designs, its supplying the capacitors which are supplying the load, so the caps are/should be chosen with this in mind

have you exposed the sense pin on your PCB Ian? or have you just tied it to the output voltage?

Agree. I have been trying different caps for it in the past weeks. It seems MLCC plays batter then TAN, through it's more sensitive to mechanical vibration. But which MLCC is the best? I'm still wondering and looking for.

Of course, the sense line should always be independent from the output

Regards,

Ian

[qusp]

many designs, particularly if the reg is on the same board just tie it directly to the output to save trouble and assuming that there will not be significant voltage drop over a short trace, but external designs like yours that can and will have unknown loads its better to have it available, so glad you did also its pretty unique in that most adjustable regulators do not have a sense pin, because the pin that is usually used for sense is changed for an ADJ pin, sensing the voltage across the setting resistors. As a result its often only fixed voltage LDOs that have a sense pin.


as for Caps, I would use a mix of Panasonic SP-CAP (SX variant) (I use these whenever I can, Acko uses them too, great caps) and MLCC and I mightve looked at using some AVX Accu-P thin film caps for the HF Si570 decoupling directly on the pins

[hochopeper]

I've been wondering if one of those tiny cap arrays would be useful for Si570 decoupling ...

DK - 490-3446-1-ND (only rated for 6.3V) - GNM314R70J105MA01L Murata Electronics North America | 490-3446-1-ND | DigiKey

Some datasheets - http://search.murata.co.jp/Ceramy/im...70J105MA01.pdf

http://www.mouser.com/catalog/specsh...M%20Series.pdf

An AVX version - http://www.avx.com/docs/Catalogs/w2lw3l.pdf

[iancanada]

Originally Posted by qusp

many designs, particularly if the reg is on the same board just tie it directly to the output to save trouble and assuming that there will not be significant voltage drop over a short trace, but external designs like yours that can and will have unknown loads its better to have it available, so glad you did also its pretty unique in that most adjustable regulators do not have a sense pin, because the pin that is usually used for sense is changed for an ADJ pin, sensing the voltage across the setting resistors. As a result its often only fixed voltage LDOs that have a sense pin.


as for Caps, I would use a mix of Panasonic SP-CAP (SX variant) (I use these whenever I can, Acko uses them too, great caps) and MLCC and I mightve looked at using some AVX Accu-P thin film caps for the HF Si570 decoupling directly on the pins

Originally Posted by hochopeper

I've been wondering if one of those tiny cap arrays would be useful for Si570 decoupling ...

DK - 490-3446-1-ND (only rated for 6.3V) - GNM314R70J105MA01L Murata Electronics North America | 490-3446-1-ND | DigiKey

Some datasheets - http://search.murata.co.jp/Ceramy/im...70J105MA01.pdf

http://www.mouser.com/catalog/specsh...M%20Series.pdf

An AVX version - http://www.avx.com/docs/Catalogs/w2lw3l.pdf

Good ideas. Based on my testing result, capacitors are very very important for a clock power supply.

Ian