Some thoughts on the DSS/UET front panel
----------------------------------------
(RK 4/10/97)
(rev. 4/22/97)
(4/22/97: Click hereTo see comments on the draft
front panel design.)
(4/17/97: Mike has provided the first draft of a front panel
design. Click here to see it(small version or big version).
It is also attached at the end of this document.)
(4/16/97: It has been suggested that maybe the remote panel
should just be a laptop PC. A low-end laptop with color display
can be purchased for <$1000. The main benefits are that it should
cut development time, be more flexible and possibly even be cheaper.
many more parameters could be displayed, even graphics.
The down-side is that most laptops have just a keyboard for input,
joysticks and knobs would have to be added separately, plus the
added flexibilty may lead to complexity for simple tasks).
(Addendum: 4/15/97: Following another meeting on 4/15/97, the
new consensus is to put most of the front panel (LCD) displays
and knobs/joysticks/buttons on the remote panel, and to leave
the cabling for the main (fixed) front panel. This makes the
remote panel an integral and necessary part of the design, but
eliminates the need to duplicate displays and knobs, and puts the
important stuff closer to the experimenter where it belongs.
Mike has prepared an initial drawing of what the front and remote
may look like, this will be available soon),
(The following is a follow-up of the meeting of 4/9/97 between
WSR, MGR and RK).
These are some thoughts on the front panel design for the
new DSS (DSS-3) being constructed by Mike. These are all preliminary
and suggestions and ideas for change are welcome.
(1) It seems generally agreed that there should be a single
front panel that serves both the UET and DSS (and A/D). This may
change later as Mike finalizes the physical layout of the boxes and
the need to be close to the host CPU etc. There may be unforseen
advantages to separating the UET and DSS (and A/D) front panels.
It is also generally agreed that there should be the option for
a remote panel, which should contain an LCD/LED display and some knobs
and switches for control of essential parameters of the DSS (such as FREQ
or ATTN). This remote panel would be connected to the main (fixed) panel
via a cable. It would be used mainly while searching for units, and with
some knobs and buttons could be used for data input in psychophysical
experiments.
(2) What should be on the (main) front panel ?
The following are some items that would be considered either
essential or desirable:
(a) 16 UET inputs (BNC) + 16 LED's (one for each input)
(i.e. upto 16 units or events could be timed simultaneously).
(b) 16 TTL outputs (BNC) + 16 LED's (one for each output)
(these can be used for a variety of purposes such as
turning lights on/off, driving stepper motors etc.)
(c) A power switch with LED indicator.
(d) An LCD or LED panel capable of Alphanumeric character display.
A desirable size seems to be 4 rows of 40 chars each. Since
this LCD panel will be shared between the UET, DSS and A/D,
an even larger panel maybe better. This panel should be
readable under light or dark conditions.
(e) Sync. OUT for two (or 4 ?) DSS channels. (BNC)
(this is a level that always goes high when the DSS is producing
sound, and goes low when it is not. It's purpose is to allow
an oscilloscope or other instrument to be triggered at the
same time the stimulus is turned on)
(f) Monitor OUT for four DSS channels (BNC).
Note: This is not the attenuator output. These are typically
connected to a speaker or oscilloscope.
(g) DSS (attenuator) OUT for 2 or 4 channels. This is what would
go to the animal/subject. (This from the 25-bit DAC).
(In some units the 25-bit DAC may be omitted to save cost, in
which case the attenutaor input will be from the 16-bit DACs)
(h) A switch to toggle between manual and computer control +
an associated LED. A push-button switch may be better than
a toggle switch.
(i) A/D inputs : Sixteen single-ended inputs (BNC) for the PC44,
plus 2 differential inputs for the PC32 A/D converters.
(The PC44 has 2 onboard A/D converters, each with 8 multiplexed
inputs. Net throughput is about 500,000 samples/sec per A/D)
Note: The PC32 and PC44 each have onboard A/D's, so we need
to provide separate inputs for each. The two sets of inputs
should be physically separate and/or color coded to avoid
confusion.
(j) DAC outputs for both the PC32 and PC44 onboard DAC's. (BNC)
(k) Knobs and/or joysticks for manual control of DSS parameters.
The parameters used most often are Attenuation and
Frequency, but it would be nice if the parameters associated
with a particular knob or joystick were programmable. If
possible we should have 4 knobs and 2 joysticks. The minimum
would be two knobs and one joystick. A nice touch here would
be if there was a small (4 to 12 char) LED or LCD panel above
each knob/joystick, which could show the parameter being
currently varied. A slider has also been suggested.
(l) Rate Generator output(s) (BNC).
(m) A reset switch ?
(n) External start inputs for the DSS, UET and A/D's (BNC).
(o) Some front panel space should be left in reserve for future
expansion.
In general we should avoid placing things on the front panel that
don't need to be changed between experiments. Some of these could be on
a "back panel". Examples are: trigger inputs for A/D start, clock sync.
cable, etc.
(3) What should be on the (optional) remote panel ?
The following are some items that would be considered either
essential or desirable:
(a) An LCD or LED panel capable of Alphanumeric character display.
A desirable size seems to be 4 rows of 40 chars each. Since
this LCD panel will be shared between the UET, DSS and A/D,
an even larger panel maybe better. This panel could be smaller
than the one on the fixed front panel, but for consistency and
ease of use it may be desirable to have the same parameters on
both panels.
(b) A switch to toggle between manual and computer control +
an associated LED. A push-button switch may be better than
a toggle switch. (Note: the host computer should always be
able to override either front panel).
(c) Knobs and/or joysticks for manual control of DSS parameters.
The parameters used most often are Attenuation and
Frequency, but it would be nice if the parameters associated
with a particular knob or joystick were programmable. If
possible we should have 4 knobs and 2 joysticks. The minimum
would be two knobs and one joystick. A nice touch here would
be if there was a small (4 to 12 char) LED or LCD panel above
each knob/joystick, which could show the parameter being
currently varied. A slider control has also been suggested.
The exact nature and placement of these knobs/joysticks/buttons
on the remote panel esp. needs feedback from users of the
lab.
(d) Two or more push buttons that could provide feedback in
psychophysical experiments. These could perhaps be timed
by the UET ?
(d) Two or more LED's that could provide feedback in
psychophysical experiments.
(4) What should be displayed on the LCD/LED panel ?
This should not be a fixed format, but different parameters could
displayed to suit the experiment. The following is a partial list of
some possible parameters that could be displayed:
FREQ, SPL, ATTN, FCARR, FMOD, DMOD, DELAY, ITD, ILD,
RTIME, FTIME, SDUR, REPINT, NREPS, Time Base, Click Rate,
DSID, File Name, Current Rep #, Success rate, Spike Rate, etc. etc.
The important thing is that the list of parameters actually displayed
is programmable, and may be added to later as new experiments are devised.
Many of the parameters do not change during stimulus presentation, but
some (most notably, Current REP #) should be updated in near real time
Consistent formatting of the LCD display is important to avoid
confusion during experiments.
Another important function of the front panel is to provide feedback
about the state of the UET/DSS. For example, it could display "IDLE"
while in it's idle loop, or "MANUAL" when in manual mode, or "FIFO
OVERFLOW" or "ILLEGAL FREQ" etc. An extension if this is to build in
some self-diagnostics that could run at each power-up (e.g. a memory test,
attenuator test etc.) and the results could be displayed on the LCD panel.
(5) How should the front panel be controlled ?
This is largely upto Mike, but the general consensus was that
a low-cost microprocessor should be dedicated to control of the front
panel. The same processor would control both the fixed and remote panels.
(6) How should the remote panel be connected ?
The specifics of this depend on what is most convenient for Mike,
but the initial thought is to use a serial port on the PC32 (or PC44, or
both ?) for all communications. Maybe a power line for the remote panel
and a few data lines for the buttons and lights on the remote panels
could be run along the same bundle as the serial line.
(7) Who's in charge ?
The host computer should be able to override either front panel
at any time. More than one control program could run on the PC32/PC44
for different experiments to provide varying levels of manual control.
The following is a reduced size sketch of the draft version of the
display/control panel.