USB and MSA power sequencing issues

This contains the scottyspectrumanalyzer yahoo group backup
Post Reply
spyker49
Posts: 254
Joined: Sat Apr 28, 2018 3:29 pm

USB and MSA power sequencing issues

Post by spyker49 » Sat Mar 30, 2013 2:15 am

Hello Group

The problem of my USB-MSA not working some times (MSA does not respond) has been difficult to reproduce. Here are my interim findings: Happy to be corrected.



My work around:

A seperate hardwired common ground must exist between all the units, PC and USB board and MSA, before anything is powered up.

Power up sequence:

(1) MSA power is OFF.

(2) USB cable to PC must be UNplugged.

(3) USB MSA cable plugged in.

(4) MSA is powered ON.

(5) USB cable is plugged into the PC.

(6) Start USB_MSA Liberty software.



What is the problem?

In the USB-MSA combo there is a mixture of unsequenced and independent power sources, so we face the drama of the classical multi-DC rail digital and analog signals, each with their own sneak paths, clamping and latching up.



*Rule*: Input pins should not be taken high when the inputs own IC supply is lower than the input signal level.

*Rule*: Outputs driving high should not drive inputs which are not powered up.

*Caveat*: Unless the IC inputs are specially designed to operate safely/unclamped in the unpowered mode. There are ICs that do this.



Background:

Most digital logic and analog hybrids have inputs diodes to 0V-GND for negative transients-- That is usually good.



Most digital logic and analog hybrids also have inputs diodes to their supply VCC for positive (over supply) transients-- That is mainly good, but is REALLY BAD if the IC is not powered up.



If the IC is not powered up then any high external voltage source fed to the IC's input pin is routed via the VCC diode to the target board's VCC rail.

This results in the high input signal trying to power up the whole remote unpowered board! Plus that input IC's own OUTPUTS to yet other inputs!!



Essentially all the remote unpowered input pins are a DEAD SHORT to any external output high source.



I suspect a powered up USB CYPRESS can not tell that all the logic in the MSA is still powered down, as it does what it is progammed to do. Can this be looked at?



Failure Mode Hypothesis: The total USB 3.3V drain by the CPU I/O ports (highs) driving into unpowered MSA input shorts droops the 3.3V VCC and/or causes the CPU internals to misbehave and latches up until powered down.



POSSIBLE SOLUTIONS:

(1) Power the whole USB board (+5V) only from the MSA +5V. Is there enough MSA +5V power available?

(2) [better] Use a MSA power ON signal to switch, via a small relay or FET, the USB +5V power feed. Intercept RED wire in the USB cable or cut up the USB board +5V power tracks.

(3) [best] Design a full composite USB-MSA control PCB running only from the MSA power system.



Happy to be corrected. Regards Len Spyker

Post Reply