XS1-L4A-64-TQ48 Test board

[AtomSoft]

How about this:


When 3.3v is hit it enables the 1v buck and then when that hits 1v (0.9 trigger) it triggers the reset. This way all voltages are monitored and with only 1 added part to the original schematic. So the XMOS wont be enabled until both supplies are nominal.

I might also add a 1uf close to the APX and BU48 VCC pins to ensure they get a smooth signal and will activate when ready and not before due to a spike.

[AtomSoft]

New Schematic...

Changes include a better look (easier to understand it), 1v reset circuit... other stuff... cant remember :) (oh yeah new osc OE tied to 3v3)

I do love the tps3808 idea, but it adds more complexity to a already very limited space board. But Ill look into it anyway. It seems to be calling me in to read the datasheet better. :)

[mon2]

Pin 1 (RST) of each of these power supervisor devices are open drain. This means that the pin will remain LOW till the threshold(s) are met and then tri-state. Therefore, you will need to add pull-up resistors to each of these devices to define the logic high state. Really do not believe you need the APX803 due to the design of the SC189 device. Also, do check on the availability of the ROHM device (BU48) and their respective footprint.

[AtomSoft]

Ah i see ok.. wow i must have been tired when i read that before. Ill remove the APX and BU part (Digi-Key Stock: 2,463 and i made the footprint myself which was easy (ssop5)), but ill remove them both for the TPS3808G09 (TPS3808G09DBVR Digi-Key Stock: 23,230).

Ill get on that now.

[AtomSoft]

Ok hows this look. MR is tied high because its not needed here... even tho there is a internal pullup, i assume its best practice to ensure its high... No cap is on CT thus producing a ~20ms delay which i think is good... or should i shoot for a longer delay like 200ms?

[Folknology]

Surely APX803xx are much more cost effective than TPS3808Gxx ?

regards
Al

[AtomSoft]

yeah but they dont offer the simple delay option. Its a simpler solution. Its price over complexity... I rather pay that extra to avoid adding 2-3 more parts on the board. Also its not like im making a hundred of these :) even if i was it would lower the cost of that one part by about 50% which is always good.

Preview attached

[Folknology]

The AXP803 have a built in delay BTW.

Looking at the SC189c datasheet the EN pin triggers at 1.2V (not 3v3 PG) that might be a bit early perhaps?

regards
Al

[AtomSoft]

The AXP803 have a built in delay BTW.

Looking at the SC189c datasheet the EN pin triggers at 1.2V (not 3v3 PG) that might be a bit early perhaps?

regards
Al

Look at the Under-Voltage Lockout

I assume no power will come out unless that requirement is met regardless of the EN pin.

[AtomSoft]

Page 14: