22
\$\begingroup\$

Summary: I'm setting up a USB-powered Li-ion charging circuit and would of course like to draw the maximum charge current possible situationally -- but also want to ensure I don't violate the USB specification on current draw. While I have been able to satisfy one/two of the USB conditions individually, I am having a little trouble thinking of a smart way to satisfy them all. Here is what's happening...

Four USB spec rules (three regarding Current draw maximums, and one regarding Inrush current limiting) pertinent to my particular situation are as follows: (Note: I'm showing different time-stages in this diagram:)

Rules

The good news is...

I happen to be using an FTDI USB-UART IC, the FT232R (datasheet), which takes care of the enumeration and also has outputs that indicate Enumeration status and Suspend status, which would be useful for setting when to draw what level of current.

Also helpful is that I'm using a Li-ion charger IC, the MCP73871 (datasheet), which has ChargeEnable pin, as well as Prog2 pin and Prog1 pin which allow setting Charge current limits.

The functions of these useful pins for both above ICs are summarized in my (rough) attempted layout below. It's not yet complete, as indicated by the Question-marks at certain connections:

My layout

My layout

Current-draw specs

Finally, here are the situational current-draw specs for the two ICs. Of course, these current draws also have to be counted into the 0.5 / 100 / 500 mA current-draw limits on the USB supply: Current draw

QUESTION: There seems to be a solution to adjust my layout so as to satisfy all the four rules/conditions at once, but I'm not seeing it; Any ideas?

As you could see in my layout earlier, I've set up an unconnected circuit (RC combination on MOSFET) for the soft start problem; and I've got the limit-setting resistor on the PROG1 pin, which could perhaps be used as part of a divider for partially solving the 500mA vs 100mA vs 0.5mA problem (based partly on FTDI's strategy in this app note). But that's as far as I was able to reach.

\$\endgroup\$
0

2 Answers 2

11
+250
\$\begingroup\$

"The MCP73871 device specifically adheres to the current drawn limits governed by the USB specification." -MCP73871

You don't need to current limit anything. Your MOSFET solution would only be useful to current limit inrush due to a capacitive load but your VDD pin is not capacitive nor would that circuit limit once it's already on. All you need to do is figure out how to use #PWREN and #SLEEP to switch in different resistor values. The MCP73871 does the rest.

  1. Just hook #SLEEP to #CE to take care of "standby".
  2. Add an inverter between #PWREN and #PROG2 but make sure floating input defaults the inverter output to LOW.
  3. Place a 2.2k pulldown on #PROG1

Note: make sure you have R1 because #PWREN will drive high and you need to limit the current into the NPN.

enter image description here

\$\endgroup\$
2
  • \$\begingroup\$ This sounds correct; I am now about to look into this method, in case there is any case it doesn't satisfy, and will comment as soon as I check it \$\endgroup\$
    – boardbite
    Sep 22, 2012 at 9:42
  • \$\begingroup\$ let me know if you need help with resistor values \$\endgroup\$ Oct 3, 2012 at 19:45
1
\$\begingroup\$

Perhaps a reset controller would be useful here; connect its output to the RC network for startup. Then take SLEEP, invert it, and it with the reset signal, and feed it to CE. For PWREN, create a voltage divider at about 1.5V or so. Put it through an inverting comparator against 1V, and feed that to PROG2.

\$\endgroup\$
1
  • \$\begingroup\$ Can you draw a schematic, it would be much easier to follow then a text explanation and you could then use text to explain the schematic. \$\endgroup\$
    – Kortuk
    Sep 21, 2012 at 10:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.