I am trying to thread a needle here and realize that I will likely not be within worst case prameters for a limited run of parts. I am designing a circuit that will be connected to USB Host ports that are already in the field. It will only be connected to our USB port, not any port in general. The question is whether I should set the circuit to run slightly above the minimum current limit threshold or slightly above the minimum overcurrent pulse width. As presently designed the worst case load will be 636 mA baseline with peaks to ~1.1 amp for 1.1 msec at 5 msec intervals. By making adjustements in the power circuitry I can get the baseline current below 600 mA, but the overcurrent pulse gets higher in amplitude up to 1.5 amps. The TPS2074 sepcifies 1 msec as its minimum delay and 600 mA as its lowest current limit. The TUSB2046 indicates that it has "noise filtering" but has no specificaiton on it.
I have measured 2 msec delay of the TUSB2046 on three units. If that was the value for every unit in the field I would be OK, but don't want to
depend on that test alone for making more units. I would like to have a specification I
can work from. Is there a range of filter times designed into the TUSB2046B that will reject pulses of say
1 to 4 msec or is it 2 +/-0.2 msec or is there a specification? On a related part,
is there a distribution of expected current levels for the overcurrent trip
point of the TPS2074? I tested three and found it to be 1.1A. The specified range is 600
mA to 1.2A (typical 900 mA). Is the 600 mA level only reached at extreme
temperature? Is there a range at 40 degC that I could expect? Is there any
quality data that might indicate the likelihood of a 600 mA threshold part
being encountered?
Please advise
Thanks
Ed
