CY7C65632 Fail to Enumerate, odd behaviour on TEST and PWR# pins

[Anonymous]

We are debugging a new design which uses the CY7C65632-28. This is the second time we have used this chip - the first time we had no problems at all. 

Compared to the previous design, all we have changed is to move to a bus powered design: 
1) tie SELFPWR to GND, and 
2) Use the internal 3.3V regulator to supply the VDD pins. (Feeding 5V into VCC).

The chip powers up OK. It checks for an eeprom - the signals look the same as the old design. 

But when connecting to a host the behaviour is different: 
1) The Host fails to enumerate. VID and PID are both 0x000. 
2) The TEST pin and PWR# pin both oscillate at 1ms. The signal on both pins are the same. That is both TEST and PWR# go high for 1ms, then both go low for 1ms. This repeats for approx 1 second. 

The signals on RESET, the 5V and 3.3V rails, and the clock all look normal while this is happening. It seems like the pins are toggling in time with the USB packets? 

It doesn't seem to be trying to access an EEPROM because the pins toggle exactly the same. 

I thought this might be a test mode - but when I hold the test pin high the behaviour of the chip is different. 

I'm at a loss! Is this a known behaviour?

[Anonymous]

No replies to this... but here is an update that might help someone else.

   

The enumeration problem was I had D+ and D- reversed - Doh!  Once that was fixed the hub enumerates just fine.

       

However I still have this strange 500Hz square wave on the PWR and TEST pins. And I just discovered the GANG pin does it as well. And it happens when the port is not in standby.

       

So when the port is in standby, the GANG pin is low.

   

When the port is active, the GANG pin oscillates at 500Hz, and so does the PWR and TEST pins.