5V operation with internal pull-up resistors (STM32F205)

[schneider2]

Posted on April 12, 2012 at 09:32

I'm currently developing a ''plug-in adapter'' using the STM32F205 that is to be used in a variety of legacy designs. These old designs all rely on 5V components and at least some of them require bidirectional communication over an open-drain line.

Because I cannot foresee in which applications the adapter will be used and how the other components will be configured, I would like to use an open-drain output stage for all I/O pins and also enable the internal pull-up resistors (external resistors are not an option because they won't fit onto the adapter). How does a configuration like this work in a 5V environment? From the data sheet (Doc ID 15818 Rev 7) I was able to gather the following information: - Injected current (Iinj) on 5V-tolerant pin: -5/+0 mA (Ch. 5.2, Table 8) ----- Positive injection (Vin > Vdd) is not possible on these pins. Iinj must never be exceeded (Footnote 3 referring to 5V-tolerant pins) - Maximum input voltage (Vin) on 5V-tolerant pin: Vdd + 4.0V (Ch. 5.2, table 7) ----- VIN maximum value must always be respected. Refer to Table 8 for the values of the maximum allowed injected current (Footnote 2, referring to 5V-tolerant pins) How am I to understand these values? Vin would be in the permitted range [0V...3,3V+4V]. According to Table 8, positive current injection is not possible, so the output configuration should not matter as long as the input voltage stays below 7,7V. However, wouldn't there still be a current going through the pull-up resistor (thereby violating the allowed injected current)? Can you help me understand these specifications? Would my application be possible under these circumstances? Thanks Patrick

[Tesla DeLorean]

Posted on April 12, 2012 at 11:36

You'll have to use external pull ups on the FT (5V tolerance) pins, as the IO ring is not at that level. The internal pull-up won't get you above the main supply, and I wouldn't want to be back feeding them. Is 3.3V not sufficient for VIH threshold? In the IO cell design, FT basically means the ESD protection diodes aren't clamping to 3.3V, but to a level that can float higher. At least that's how I understand what ST is doing.

Can you honestly tell us the you can't fit a few 0201 resistors on the front or back side of the board? They are pretty tiny, could always park on top of a via.

[schneider2]

Posted on April 13, 2012 at 09:00

I don't need the pins to output more than 3.3V. It's just that most of the I/O pins are routed off the main board and they *could* be interfaced to components that use open drain bidirectional communication where the idle voltage is 5V.

That in itself is not the problem as the STM32 can be configured to use open drain outputs. But I also want a defined voltage level on these pins, even if they are not actively driven (hence the need for pull-up resistors). Since there are a lot of pins that can be used in this manner, I'd look at many, many resistors (which I'd like to avoid). Meanwhile, I've found another footnote: ''Pull-up and pull-down resistors are designed with a true resistance in series with a switchable PMOS/NMOS'' So unless these FETs are designed with an embedded diode to block current in the ''reverse'' direction, there really would be feedback into the 3V3 supply... I guess I'll have to rethink my approach (or else use 40 resistors, however tiny they may be). Thank you for your input.