GPIB equipment inconsistent behavior

[shuttlefan]

Inherited a spaghetti code that controls two yokogawa GS610 power supplies. 
It has the VISA resource selection on the front panel and the lines directly used by VISA write/reads.  Not using the initializing VI from the NI Instr drivers.  Code works to drive both units. 

Modernizing the code, instead of front panel, it's reading the ID# from an ini file and then looking up the results from VISA find to generate the VISA resource.

Up to here, no issues.  The VISA resource is wrapped in a powersupply class.  I have a "constructor" that just calls the initialize VI from the driver.  There is a difference here between the old and new code.  VISA open is explicitly called where as it was not before.

So here's where the problem shows up when I try to initialize my objects.

 The starting order is Neg PS, followed by Pos PS.  No particular reason for the order.

When pos_addr = GPIB0::1, neg_addr = GPIB0::2, then Neg PS starts fine.  Pos PS hangs

I've swapped out a spare unit for pos_addr and same behavior.  

So then I tried pos_addr = GPIB0::3, neg_addr = GPIB0::2, this time, the Pos PS works, but Neg PS hangs

So it seems the higher ID works but not the lower one.  Is there some GPIB order/timing that needs to be followed?  It's the same object code so I'm not sure why it's behaving differently