Replacing the power supply on the xCore-200 MC Audio B

Sub forums for various specialist XMOS applications. e.g. USB audio, motor control and robotics.
User avatar
xlordofpainx
Active Member
Posts: 62
Joined: Thu Mar 08, 2018 2:44 pm

Replacing the power supply on the xCore-200 MC Audio B

Postby xlordofpainx » Thu May 17, 2018 9:09 am

I am interested in switching the ST1S06xx http://www.st.com/content/ccc/resource/ ... 100211.pdf

with another power supply that is easier to hand-solder
http://www.analog.com/media/en/technica ... 6_1087.pdf For example.

Can someone confirm that it would be possible?
I am looking at the requirements, but I am not exactly sure

VDDIO and OTP_VCC supply is within specification before the VDD
(core) supply is turned on. Specifically, the VDDIO and OTP_VCC supply
is within specification before VDD (core) reaches 0.4V (Section 13).
The VDD (core) supply ramps monotonically (rises constantly) from 0V
to its final value (0.95V - 1.05V) within 10ms (Section 13).




Image This is how the new power supply sheet is - kept the brownout and delay circuit, but changed the 3v3 and 1v Bucks for slightly larger ones.
User avatar
mon2
XCore Legend
Posts: 1158
Joined: Thu Jun 10, 2010 11:43 am
Contact:

Postby mon2 » Thu May 17, 2018 10:24 am

Hi. The Analog part is a voltage supervisor and not a power supply regulator. Search for switching buck regulators in SOT-23 package. Review the input voltage range for such parts to confirm if it is suitable.

Where will the 1v8 rail connect to in your design? That voltage rail is not being sequenced and is not correct as drawn in the schematic.
User avatar
xlordofpainx
Active Member
Posts: 62
Joined: Thu Mar 08, 2018 2:44 pm

Postby xlordofpainx » Thu May 17, 2018 11:42 am

So the 1.8 is for another chip, but it just happened to "snip-photo-bomb" the buck-converters that I am talking about. The brown-out,sequencer and buffer are not important for this question. The buck converter I chose to replace the standard ones used in the xmos audio board design (https://www.diodes.com/assets/Datasheets/AP3419.pdf)
has enough current and I've regulated the circuit so that i outputs the right voltage, but I was worried that I haven't taken something into account as this is my first project of such scale. I am unsure, for example, about the 10ms maximum time for the core to reach 1V. The data sheet says
Image And it looks like it can manage in less than 5ms, but I still wanted that I ask in the forum for a second opinion.
User avatar
mon2
XCore Legend
Posts: 1158
Joined: Thu Jun 10, 2010 11:43 am
Contact:

Postby mon2 » Thu May 17, 2018 12:56 pm

What is the value of C10 (U6) in your power supply supervisor circuit? Leave the room for a delay cap but it should not be required. Actually may be held against you if you delay the PG output. Do not stuff C10 for initial testing unless you have a specific reason to do this.

Why not replace the linear LM317 regulator with another very efficient AP3419 for the 1v8 leg of the power supply? Otherwise, you may find that the LM317 will get warm to hot to the touch and require a heat sink varying with the expected load of this voltage rail. Just remember that in linear regulators, you will be burning the 5 volts down to the target value of 1v8 so the efficiency is quite poor and it will show in the amount of heat produced by this very old school component.

D7 may or may not light up as expected. Review the datasheet for the LED you will be using. A small trick you can use here is to insert an inverter at the output of the low voltage rail. Then at the output of the inverter, arrange to have the cathode (-) of the LED -> the current limit resistor of 390 ohms or similar -> then the anode of the LED to +5 volts.

When the 1v8 rail is LOW (ie. 0 volts or close to), the inverter output will be HIGH and the LED will remain OFF. When the 1v8 rail is not LOW, the inverter output will be LOW and cause the LED to be lit. Without this extra single gate inverter or similar, the D7 led will be most likely dim, if it works at all.

Use a single gate inverter like this one:

https://www.diodes.com/products/logic/s ... /74LVC1G04

and power it from the +5 volt rail for the LED.


Image
Attachments
ap3419_1v8.png
(73.53 KiB) Not downloaded yet
ap3419_1v8.png
(73.53 KiB) Not downloaded yet

Who is online

Users browsing this forum: maqu67 and 7 guests