Usb device fixes

[jayich]

We implemented a couple of changes (described in the commits) to get our only two 'USB' GPIB devices working.

We have found that when using the latest pyvisa that '::INSTR' is included.

>>> import visa
>>> visa.__version__
'1.7'
>>> rm = visa.ResourceManager()
>>> rm.list_resources()
....

We also have had problems with devices needing some time between write and read. This solution looks like a hack. If you guys have a better solution that would be great.

Credit to @aransfor and @gregllong for both of these fixes.

[maffoo]

didn't know you could pass a tuple to startswith. nice.

[aransfor]

I think this is a hack, it's kind of annoying that a specific device server should have to have some write/read delay when in principle should have some standardized protocol that the GPIB server deals with. I would chalk this up to one poor device but now there is both a Rigol signal generator and some Agilent thing @jayich uses that hangs on GPIB reads

[aransfor]

I think the GPIB server should at least timeout instead of hanging on a bad read operation

[maffoo]

@aransfor, the server is supposed to time out. By default the read timeout is just 1 second, but you can set this per context, if you want to do something for a particular device. I think the read and read_raw methods might behave differently in this regard, however, so that could be the issue, because I see that the new USB code uses read.