FROM MATT NOY 02/06/09
======================
Alone, the USB_DAQ can interface to a wide range of single ended and 
differential logic signals. Unfortunately, due to modern silicon 
processing, the inputs are not 5V tolerant. This means that standard TTL 
signals may _not_ be used, or damage to the device will occur. Other 
than that, you can individually jumper the banks from 2.5V to 3.3V, and 
by changing a resistor somewhere, should even be able to change the 2.5V 
plane on the PCB to some other standard, though I've never tested this 
feature of the design. (The 3.3V plane also powers the USB interface, so 
shouldn't be changed).

To cope with 5V signaling, I put some level translators on the daughter 
card, these are brought out on some 0.1" pitch IDC connectors labeled 
Laser 0 and Laser 1, just next to where the PMTs would plug. Each 
translator has 8 bits and, in the current firmware, I fixed their 
directions to be:

Laser 0: output from the USB_DAQ
Laser 1: input to the USB_DAQ.

They also have the capability to be tri-stated, but are currently not.

It's relatively simple to change this, but would require a rebuild of 
the firmware.

As the names suggest, these were originally intended for use in 
synchronising the spill with the RAL laser. In the end, I think it was 
only ever me who used them when dealing with the laser at IC.


In terms of the trigger signals, I store 8 bits per BX, 2 of these come 
from the PMTs and 6 from the pins. The PMTs generally produce quite 
short signals (~5ns) so I use a level sensitive latch, the other 6 are 
latched on a 40MHz clock. These are all reset at the end of the BX when 
the data byte is written into the FIFO buffer.

The 6 non-pmt trigger inputs are divided into two groups, trig_bus and 
the original laser interface inputs (to the USBDAQ). In practice, these 
are simply labels. Looking at the pin constraints, I see that I 
connected the Laser 1 inputs as follows:

Bit: function
0: lamp_sync
1: qswitch_sync
2: trig_bus(0)
3: trig_bus(1)
4: trig_bus(2)
5: trig_bus(3)
6: unallocated
7: hw_trig_in

So, 0->5 should go in with the PMT record, provided that the signal is 
longer than 25ns (or falls in the right place by luck...). Bit 7 is the 
input I connected to the IC laser output when we were doing the tests at 
IC on TPAC1.0. This is latched using a similar level sensiitive 
mechanism to the PMTs and will start the spill operation when the 
correct spill mode bit is set.

The outputs from Laser 0 are allocated as follows:

Bit: function
0: lamp_trigger
1: qswitch_trigger
2: unallocated
3: unallocated
4: unallocated
5: test_trig_output (you set when you want this to start and stop)
6: in-spill indicator: high during BT
7: spi_indicator. Jamie wanted this for something...