TPS62740: VSEL4 behaviour when unterminated

Hi Barry,

I don't quite understand your question. Are you asking if the behavior of the VSEL4 pin is defined when it is left floating? No, it is not. It must be terminated, per the D/S.

Hi Chris;

It’s not intended to operate it with VSEL4 floating, its intended to see IF VSEL4 is floating... from some tests on the 6740 EVM, I observed this behaviour of VSEL4 following VSEL3 state, during power-up, and a short time following, less than a second, out to a couple of seconds...as I thought that I mentioned, I am trying to see if  VSEL4 is terminated, and the VSEL4 pin/pad is not accessible to check/confirm the connection. Therefore, as a test to catch an unterminated VSEL4, I can „wiggle“ VSEL3 during power-on, and, if VSEL4 is connected,  about will be stable/work as expected... if VSEL4 is not terminated, about will change by 800 mV more than expected when VSEL3 is changed. This is ONLY at power-on, ONLY as a test. I’d be interested if you could see/confirm this behaviour? My next level is to try with just the device, no EVM, and Probe with a wafer probing station:-)

Barry

Power-up sequences

- turn on Vin, EN = Vin,

-  All control lines that are at Hi state are tied to Vin.

Here's a scope shot, kind of nasty resolution, but it's no R&S or Tek scope ;-)

it's 'cold start' condition, Vin at 0, going to 3.6 V (i think)

VSEL3 is tied to Vin to start with, then disconnected and reconnected to GND (by hand, as fast as possible... did not have any SPDT switches close to hand ...

VSEL3 is loaded by the scope probe, 1 M, 12 pF. Which does, as I saw in other tests, affect the VSEL3 change of state when it's otherwise 'floating'

Anyway, here, one can see that Vout is high, to begin with, and then drops a bit when VSEL3 changes to Low, and a bit later, about 600 ms, Vout shows that VSEL4 has made it to the low state.

now a couple of other scope images, two cold-starts, in a row, with VSEL3=GND in one case, VSEL3=Vin in the other case, showing that VSEL4 does seem to 'follow' VSEL3 at cold start (but, not forever ;-) ) This shows that Vout is consistent with VSEL4 initially at the same 'state' as VSEL3...

Thanks. So all these tests were on the EVM with VSEL4 floating?

In the second waveform, Vout goes to 3.3V which indicates that both VSEL3/4 are high.

In the last waveform, Vout starts at 3.3V (both VSEL3/4 = high), then goes to 2.9V (VSEL3 = low, VSEL4 = high), then goes to 2.1V (both VSEL3/4 = low).

In the second and third waveforms, is VSEL3 changing by hand? I used a function generator in my setup but otherwise, they should be the same.

Yes, TPS62740 EVM, VSEL4 floating.

VSEL3 changed 'manually'; I ain't got no stinkin' function generator... ;-)

If you contact me directly, I can tell you a bit more about my application, but can't say much more in the forum mode.

In the case where we go from 3.3 V to 2.9 V, then drop further to 2.1 V, the elapsed time from the VSEL3 state change seems to be the VSEL4 'discharge' time, which does have quite a bit of variation across variations in Vdd OFF time, and repetitions.

Good to hear that you are seeing similar results, confirms that I'm not too  crazy ;-)

I do see similar results, but only when I power up Vin.  Toggling VSEL3 during operation did not seem to couple to VSEL4.

Same setup as your scope.  Green is VSEL3, yellow is Vin, red is Vout.  Vout first goes to 3.3V, before dying down to 2.5V.