|
|
Karbon-CL
is the world's first four-camera PCI Express
frame grabber. It can simultaneously acquire
from up to four Base CL cameras or two Full
CL cameras (including 10-tap CL). It is built
on top of BitFlow's FlowThru technology, which
provides zero latency access to data, super
low CPU usage, and unlimited DMA destination
size. The Karbon-CL is the first member of BitFlow's
Karbon family, a platform that will host a wide
variety of virtual frame grabbers. These virtual
frame grabbers can be customized to meet your
specific needs.
The Karbon-CL has been designed
with two main applications in mind. First, in
situations where more than one camera is needed,
the Karbon-CL can reduce both the system cost
and the hardware footprint by its ability to
acquire from up to four cameras. Second, in
situations where extremely high data rates and/or
frame rates are required, the Karbon-CL has
been designed to acqcuire up to 160 bits at
85 MHz pixel clock rate and DMA at data rates
up to 2.0 GB/S. For example, the Karbon-CL can
acquire simultaneously from two 10-tap CL cameras.
Half-Size x8 PCI Express Board
Up to 160 bits input at 85 MHz
Acquire from four independent
Base CL cameras
Acquire from two independent
Medium/Full CL cameras
Acquire from two 10-tap CL cameras
FlowThru technology eliminates
the need for on-board frame buffers,
even with the fastest cameras
Hardware Bayer matrix decoding
(optional)
Multi-tap cameras rasterized
on the fly
Highly configurable acquisition
engine
DMA at data rates up to 2.0 GB/S
Supports images up to 256K x
128K
No frame rate limit
Independent trigger and encoder
inputs for each camera
Independent CCs outputs for each
camera
General purpose I/O
Appears to software as four independently
programmable frame grabbers
Supported by the BitFlow SDK
on 32 and 64-bit Windows
Acquire image sequences well
beyond the 4 GB barrier
RoHS Complaint
One of the exciting new features
of the Karbon-CL is the support for virtual
frame grabbers. A virtual frame grabber, from
the software and interface point of view, is
identical to a normal frame grabber. The difference
is that the Karbon architecture can support
a wide variety of different virtual frame grabbers
with no change in the hardware. This design
can also support different numbers of frame
grabbers on the same platform. For example,
the Karbon-CL can be configured as two full
Camera Link frame grabbers, or it can be configured
as four base Camera Link frame grabbers. The
software will see two or four (respectively)
completely independent frame grabbers. The Karbon-CL
could even look like one "double full"
frame grabber (160 bits), if and when the camera
manufacturers come up with one. Further, the
board become an interface for the completely
non-standard device that you might be designing
in your lab, just lets us know your requirements.
Switching between the different virtual grabbers
only requires a press of a button in our configuration
tool. As the Karbon family expands, you'll see
more new and exciting uses of this virtual frame
grabber architecture.
For the last 10 years, BitFlow's
camera interface products have been built around
our revolutionary FlowThru architecture. Comprised
of a user-programmable Video Pipeline, a flexible
Camera Control Unit, efficient high-speed video
FIFOs and a highly-optimized scatter/gather
DMA engine, the FlowThru architecture allows
the Karbon-CL to control, acquire, reformat
and transfer video data directly into the user's
application at camera speeds with zero latency
or CPU usage. Our FlowThru architecture has
been continuously optimized and enhanced to
support a wide variety of imaging applications
such as document/package processing, semiconductor,
continuous web inspection, sequence capture
and motion analysis and can easily be adapted
to the specific needs of your application.
The Karbon-CL uses a x8 PCI Express
bus interface. The PCI Express bus offers huge
increases in DMA performance over the PCI bus.
However, what is less well known is that the
PCI Express bus is always peer to peer. This
means the the Karbon-CL does not share the bus
with any other devices. In most motherboard
architectures, it will talk directly to the
PCI chipset that is on the memory bus. This
direct connection equates to higher sustained
DMA bandwidths regardless of system load. Also,
most motherboards support concurrent full DMA
speed on all of their PCI Express slots. The
board will work in any slot that it fits in.
This means not only x16 and x8 slots, but also,
as is becoming the trend, x4 and x1 slots that
use x16 connectors. Performance will be degraded
in x1 and x4 slots, but the board will work
fine in applications that don't require maximum
data rate.
Supported by a GUI camera file
editing utility (CamEd), the Karbon can acquire
fixed or variable size images and features a
programmable ROI (Region Of Interest) sub-windowing
capability. The board provides a full set of
camera control signals (CC1, CC2, CC3, CC4)
and sync inputs (LVAL, FVAL, PCLK, trigger and
encoder) for each camera connected to the board
(up to four cameras can be connected). These
signals are completely independent, although
there are provision for driving all cameras
from a signal set up encoder/trigger signals.
There is also a large number of programmable
general purpose outputs and inputs that are
not tied to the camera's timing. The Karbon-CL
board, as with our past interface products,
supports not only simple triggering modes but
also complicated, application-specific triggering
and control interactions with your hardware
environment.
|