My Power On Reset circuit, sometimes goes high for almost 10ms when power is applied!
Green is power being applied. Blue is the POR going high before going into the reset condition.
Well that’s interesting…
Some thoughts:
That looks like 60Hz wobble at the left. I suppose it’s possible for the probed nodes to look like a high impedance under minimal applied potential and it might not mean anything, but in many cases it’s appearance has been a signal that a silly mistake is in progress…
That transient looks to be most of 5V in amplitude. Curious, seeing as how the driven opto’s input diode clamps that node to something less than ~1.5v after the reset expires. I don’t see much but that 10K resistor driving the node high, so the best explanation would seem to be that your TP3 node is high-impedance for some period of time at the outset.
Also, if you’re stuffing R1 with the 10K shown, the opto’s getting substantially less drive current than suggested in the datasheet. Perhaps one can get away with that if on-state resistance isn’t a concern, but the potential for inconsistency stemming from production and temperature variance seems risky.
1 Like
The problem is that it allows the relay to be on during those 10ms. The reason for the relay is that the data line is also used for power. When connected without the relay we get a bounce of up to 9V. This tends to destroy parts connected to it. This is an old design that I am trying to fix with putting in the solid state relay. When to POR works correctly everything is wonderful.
I have used a 299 ohm resistor. Same thing happens. In my mind the POR needs to stay low until over 100ms after the power is applied.
This would reinforce my suspicion that there’s something askew with the grounding somewhere, either in the apparatus, the measurement, or their combination. The SSR shouldn’t be on without current flow through the input diode, and if that’s happening, to measure 5V across a diode with a rated Vf of ~1.5 max means that it’s either being absurdly over-driven or (more likely) that one is not measuring what they think they are measuring.
That said, while I could be overlooking something, I see nothing in the '6328’s datasheet indicating that the output is guaranteed to be logically valid and glitch-free over the entire startup process from Vcc=0. For the logic-based systems it contemplates, a 10ms glitch on startup is unlikely to be a problem, because the part goes on to issue a valid reset signal which would be used to reset the system, fixing the effects of it’s own and any other startup anomalies occurring throughout.
Though I still suspect there’s an overlooked grounding/measurement issue somewhere, I think you may have a larger issue in that the chosen device may be inappropriate for the task.
1 Like
I agree with @rick_1976 that there’s likely a measuring or grounding issue in play.
Just looking at datasheets of other similar parts with open-drain outputs, the typical behavior is that the voltage measured between Gnd and Output is undefined until Vcc reaches the reset threshold voltage, Vth, at which time it will pull the output low for the power-on delay time.
However, that voltage is still well below the nearly 5V you are showing.
Separately, the value of R1 would ideally be in the range of 1k to 2k for a 5V Vcc in order to provide enough current for the CPC1004 input diode yet stay well below the 20mA max sink current of the MAX6328.
Giving it a solid continuous GND works wonders, unfortunately in real life I can not do that. Both GND and 5V are hot plugged.
HI @ethompson ,
As DigiKey gentlemen Rick and David already pointed out, this is clearly a grounding issue. The problem lies outside the shown circuit. Likely the “ROV GND” floats during the 10ms period (and your scope is grounded to a ground that does not float, but is a reference to the +5V supply). Can you show an expanded view of your setup (illustrating the circuits involved in generating the +5V and the hot-plugging arrangement).
Cheers, heke