First board design - Power supply

Technical discussions around xCORE processors (e.g. xcore-200 & xcore.ai).
basil4j
Member
Posts: 9
Joined: Tue Aug 17, 2010 11:21 pm

First board design - Power supply

Post by basil4j »

Hi Guys,

I should have really posted a ‘hello world’ but figured i’d get right into it.

I have been programming the propeller for years (and notice a few familiar forum names) but have decided to look at XMOS because I have a concept which needs more MIPS and I think the architecture would be more suited to this particular application.

Anywho, I’m working on a design for the power supply and have read all the PSU posts I can find on xmos.com, xcore.com and xlinkers archives.

I have questions on some of the designs I have seen:
LTC3417 (Reference docs) – Physically too large for my requirements
NCP1521 + FAN2558 (Reference docs) – 180mA on the 3V3 side if too low for my requirements unless I put in an extra 3V3 for the other devices (not ideal)
NCP1532 (Xcore.com forums I think?) – I like this but won’t be able to solder it 


I would hope to get >= 800mA at 3V3 and approx 600mA at 1V0 (double the requirement).
I like the idea of a single DC-DC converter but am open to suggestions.

Have anyone used the MAX1775? http://search.digikey.com/scripts/DkSea ... 5EEE%2B-ND... looks nice.


bearcat
Respected Member
Posts: 283
Joined: Fri Mar 19, 2010 4:49 am

Post by bearcat »

The bigger brother of the NCP1521 is the NCP1529. It's rated to 1A. Same circuit. Good for either 1V or 3.3, but you would need two supplies. It's tough to beat this series because of the low part count and no diode. The package is reasonable to solder by hand. One issue with all the small switchers I looked at is the low max voltage, usually 5.5V maximum.

Lot's of choices for power supplies.
basil4j
Member
Posts: 9
Joined: Tue Aug 17, 2010 11:21 pm

Post by basil4j »

That looks like the bees knees, thanks :)

I'll be adding a beefier 24V PSU which can supply 5V out to the smaller regulators so the low input is not an issue.

Is it better to go 5V>3V3>1V, or 5V>3V3 & 5V>1V?


EDIT: Forgot to ask. If I use 2 of these (1 for 3V3 and 1 for 1V), what it the best (most reliable) way to enable the 1V0 once the 3V3 reaches a stable voltage?
bearcat
Respected Member
Posts: 283
Joined: Fri Mar 19, 2010 4:49 am

Post by bearcat »

I don't think you can get the currents you want by chaining the supplies. But be sure to review the design checklist by XMOS as you need to sequence the power suplly startup.
User avatar
boeserbaer
Active Member
Posts: 51
Joined: Fri Jan 29, 2010 4:36 pm

Post by boeserbaer »

For my design I used an LTC 3573 isolated switcher (works from 3V ->24V) 5V on the secondary. I used 2 ncp1521 switchers for the 5-> 1.0 and 5->3.3. Works great.

I did have to use a power supervisor on the 3.3V. I used an ADCMP354 due to availability.
I have had no problems with power on reset or other issues issues related to power.
basil4j
Member
Posts: 9
Joined: Tue Aug 17, 2010 11:21 pm

Post by basil4j »

bearcat wrote:I don't think you can get the currents you want by chaining the supplies. But be sure to review the design checklist by XMOS as you need to sequence the power suplly startup.
Working on the sequencing now, hence the edit in my last post :) Using the NCP1529 for both 3V3 and 1V0 will give me enough power for what I need. After doing more research and thinking, I have changed my design a little and any higher current devices I use will get their own PSU.

Thanks for the help guys! Any suggestions for a 'power good' input to the 1V0 NCP1529?
User avatar
boeserbaer
Active Member
Posts: 51
Joined: Fri Jan 29, 2010 4:36 pm

Post by boeserbaer »

regarding power good from 3.3V, I used an Analog devices ADCMP354 (0.6V Vref) with a voltage divider on the input of 383k/100k. Once the 3.3V power is over 2.89V, then the adcmp 354 asserts the proper enable to the V1p0 switcher.
powerCropped.pdf
(174.46 KiB) Downloaded 275 times
powerCropped.pdf
(174.46 KiB) Downloaded 275 times
I also included a power on reset which drives both the rst and trst pins.
powerOnReset.pdf
(508.97 KiB) Downloaded 277 times
powerOnReset.pdf
(508.97 KiB) Downloaded 277 times
basil4j
Member
Posts: 9
Joined: Tue Aug 17, 2010 11:21 pm

Post by basil4j »

boeserbaer wrote:regarding power good from 3.3V, I used an Analog devices ADCMP354 (0.6V Vref) with a voltage divider on the input of 383k/100k. Once the 3.3V power is over 2.89V, then the adcmp 354 asserts the proper enable to the V1p0 switcher.
Thanks :) Ill look into that.
boeserbaer wrote:I also included a power on reset which drives both the rst and trst pins.
Thats similar to what I have (which I based on the reference docs). Good to see i'm on the right track at least in 1 area!
PSU.png
(44.88 KiB) Not downloaded yet
PSU.png
(44.88 KiB) Not downloaded yet
BUF.png
BUF.png (27.67 KiB) Viewed 6003 times
BUF.png
BUF.png (27.67 KiB) Viewed 6003 times
(I wonder why my buffer schematic will display but not the PSU one...hmmm)
User avatar
segher
XCore Expert
Posts: 844
Joined: Sun Jul 11, 2010 1:31 am
Contact:

Post by segher »

It looks like you have mixed up #RST and #TRST on the buffer?
basil4j
Member
Posts: 9
Joined: Tue Aug 17, 2010 11:21 pm

Post by basil4j »

segher wrote:It looks like you have mixed up #RST and #TRST on the buffer?
Good spotting! My eyes have become immune to that sort of thing hehe thanks
Post Reply