2
\$\begingroup\$

I'm using the following circuitry to protect my microcontroller input pin from the dangers of the bad and ugly outside world (a trigger signal coming from an unknown source).

The signal comes in from the right, goes through some preliminary filtering/clamping, and then enters the MAX366CSA (a signal-line protector IC, see datasheet). The output pin of the MAX366 goes straight to the microcontroller pin (Atmega2560).

enter image description here

For the measurement I'm using a 10x/500MHz/8pF, properly compensated scope probe. The blue curve is measured at the MAX366 output pin, The yellow curve shows the 4.5V square wave that's being fed into the TRIG IN connector.

enter image description here

As you can see, the rising edge is fine, but the falling edge shows a strange double-decay characteristic. After a fast drop to about 1.6V, a very slow decay with a time constant of around 500µs follows. From the MAX366 datasheet I couldn't find an obvious reason for that behaviour, i.e. their application examples never show any special requirement for the output, like a pull-down to discharge the output or something. The signal at the MAX366 input is fine and has a fall time comparable to the rise time.

Can anyone shed light onto this behaviour? Is there a design flaw on my side, or something obvious I've overlooked in the datasheet?

To improve the falling edge, would you suggest a ~20k pulldown to do the job?

\$\endgroup\$
2
  • \$\begingroup\$ By convention, signals go from left to right, please. \$\endgroup\$ Aug 27, 2014 at 23:13
  • \$\begingroup\$ @SpehroPefhany: I know, sorry. This is a small cut-out of a larger schematic where it sort of made sense to have it that way around (the microcontroller is to the left of that cut-out)... \$\endgroup\$
    – DerManu
    Aug 29, 2014 at 14:35

1 Answer 1

4
\$\begingroup\$

Quoting the MAX366 datasheet:-

"When signal voltages exceed or are within approximately 1.5V of either power-supply voltage (including when power is off), the two-terminal resistance increases dramatically, limiting fault current as well as output voltage to sensitive circuits"

The slow edge starts when you get within 1.5V of the negative rail, so the part is behaving as it should. Without more knowledge of what is on the other side it's hard to tell what would help.

If you want to use the MAX366 you could use a comparator with a 2.5V reference, then you wouldn't care about what happens below 1.5V or above 3.5V. You could even have some hysteresis so slow rise/fall would give a clean trigger pulse (a few comparators have hysteresis built in). Or just use a NC7WZ14P6X, which I think would give you enough margin (but check).

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Thanks! I hadn't wrapped my head around the fact that by using a 0V as V-, I was in a "fault condition" with low-level input, according to their terminology. Also thanks for the comparator suggestion. For simplicity I'll first try a pull-down. Though this might go wrong since reading the datasheet again suggests that also in high-level the MAX366 output becomes highZ -- the pull-down would then make the pin go low right after reaching 3.5V. I'll report back. \$\endgroup\$
    – DerManu
    Aug 28, 2014 at 7:00
  • \$\begingroup\$ The ST inverter is a mid ground, tiny and only a few cents. \$\endgroup\$ Aug 28, 2014 at 12:56
  • \$\begingroup\$ For protocol, I've tested it with a 18k pull-down on the MAX366 output pin and it works fine. It reduces the top voltage from 4.1V to 3.7V and reduces the fall-time from 500µs to about 2µs. (However, since I've already ordered comparators, I'll throw them in anyway, as per Spehro Pefhany's suggestion) \$\endgroup\$
    – DerManu
    Aug 29, 2014 at 14:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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