Dear Friends,
I have received a couple of replied e-mails. Now summarize them below. You
may see that not much information yet at this moment, because it's a quite
new issue and it's just starting this year. Sincere thanks to Drs. Robosky,
Webb, Daelen, Turner, Strahan and Kurutz.
The new console shows the most improvement in the solids and imaging areas.
Due to the new consoles improvement in speed and dynamic range it would be
wise to buy the new console. Varian is still working out some software bugs
but its worth putting up with those.
We installed our new system a couple of months ago. You can check out the
installation here
(<
http://www.nanuc.ca/facilities/console.php>
http://www.nanuc.ca/facilities/console.php)
We will be using the system for 1D high throughput ans so have not
investigated it's full capabilities visa-vi 2d, 3d or solids. One big item
is you must use VNMRJ.
With new receiver RXAD, Varian has no more quadrature detection but still
pts synthetiser.
Regards,
Yuyang
The following (copied from Dr. Webb) is only for those interested in the
new system.
---------------------------------------------------------------------------------------------------
The New Console
Frequency Synthesis
Frequency synthesis is performed using PTS synthesizers, but the AP bus
delays present in INOVA are gone.
Transmitters
RF transmitters are similar to those in INOVA, but the phase-shift
resolution has been improved to 0.04439 degrees (360/8192) from 0.25 degrees.
Digital Controllers
The RF Control Cards of INOVA have been completely replaced by digital
controllers. These are fed control programs over Ethernet via the new
Ethernet router from instructions generated by the host computer. Each
controller has 64Mb of memory. All RF controllers are identical and are
interchangeable. Each controllers activity is determined by its dedicated
interface board that is mounted on the backplane of the card cage (which
channel, for example). RF, lock, gradient, master and receiver controllers
are based on the same board. The receiver has the ADC board mounted and
corresponding signal input via a BNC. The lock controller has an output BNC
to deliver the lock frequency to the lock RF transmitter/receiver and a
connector to receiver the detected lock signal. This is digitized on the
board and can be displayed by the host computer during the experiment at
any time. The controllers receive an initializing signal from the master
controller, also mounted in the same card cage. Each performs its own
control program independently and only for the channel or function
specified. All controllers generate shaped pulses and/or waveform spinlocks
or decoupling as part of the pulse program, in real-time. No separate
hardware, such as a waveform generator is needed. The speed of the device
is adequate to generate arbitrary waveforms since no communication to or
from another device is needed. This also eliminates all AP bus delays so
that pulse programming may be made simpler and more reliable (fewer coding
errors from coding calculations to correct for AP delays). All
communications are based on the 80MHz clock.
In INOVA, the host computer delivers its instructions to the acquisition
computer which sends out instructions either via dedicated wires for
high-speed control, or the AP bus that interjects (deterministic) delays
for communication. All devices (RF channels and gradients) are controlled
by this single pathway. In order to perform complicated RF excitation,
waveform generators are used to store and deliver waveforms. The use of two
devices engenders additional AP delays.
Receiver
The receiver is completely redone and is now positioned in the digital card
cage because it is primarily a digital device. It has its own Ethernet
control line and CPU, along with 64Mbytes of memory. The control program is
loaded over the Ethernet line directly from the host computer. This board
receives an initial timing synchronization signal from the master
controller (along with all of the other controllers). It performs the
signal detection directly at the IF frequency. No signal amplifiers are
present on the board. A small mezzanine board contains the 80MHz ADC (14
bit) that is used to directly digitize the 20MHz IF frequency obtained from
the mixer. The output of this is fed into a DSP chip that performs digital
filtering and down-sampling to a 5MHz bandwidth. The output of this goes to
a second DSP chip that performs more digital filtering and down-sampling
direct to the spectral window (sw) with an effective 20-bit resolution.
This direct IF detection removes the need for quadrature detection, since
signal frequencies are all positive and there is no confusion as to which
side of the carrier frequency is correct.
In INOVA and UNITYplus the receiver is located in the analog card cage. It
receives the 20MHz IF from the mixer and then further amplifies and
down-converts the signal to dc. This would alias positive and negative
signals in the 0 to sw window, so quadrature detection was used to
differentiate positive from negative signals.
The analog filters limited the maximum spectral range to 500 kHz. The
hardware digital filters were capable of filtering and downsampling down to
baseband, but the speed of the processing limited the simultaneous
optimization of passband (optimally a top-hat function) and baseline
quality, resulting in a compromise solution. Newer requirements such as
providing good baselines for high Q probes having long ring-down times had
to be dealt with by digital filtering in the host computer.
The new receiver can deal with homonuclear decoupling by proper data point
sampling while the RF is gated on and off. This capability is general and
can be used with any pulse sequence. In INOVA, single-channel
homo-decoupling requires specific statements involving explicit acquisition
to be coded. The automatic nature of this capability makes the use of a
second channel for homo-decoupling unnecessary. The direct control of
receiver gating gives more flexibility and higher quality data since there
are no free-running clocks governing the sampling period.
Shim Power Supply
A new, more compact, shim power supply has replaced the INOVA power
supply.
Amplifiers
The system uses the same RF amplifiers as used for INOVA. Blanking
control is automatically blanked on all amplifiers for the nuclei and H2,
ensuring that no noise is injected by unused channels or from a channel set
to H2 (eliminates noise in the lock circuit). This eliminates the need for
ampmode selection in most cases.
Gradients
The system uses a new (L650) gradient amplifier for single-axis operation.
It is also more compact. Three-axis operation uses the same amplifier as
used for INOVA.
Gradient shaping is a standard capability, not requiring a waveform generator.
The Magnet Leg
The magnet leg has been completely redone, removing all pneumatics
functions. The cabinet is under its own switched power. Its sole job is
to provide RF to the probe, to amplify the signals at the NMR frequency and
then to mix down these signals to the 20 MHz IF frequency. It also supplies
probe tuning functions.
In INOVA and UNITYplus the magnet leg contained pneumatic control for upper
barrel and VT gas, as well as RF and signal detection hardware. Preamps and
T/R switches were separately housed near the probe. Probe tuning required
moving cables.
Preamps
Multiple preamps and T/R switches are now possible in addition to the
standard two. The architecture is general and in parallel. Each preamp T/R
switch module serves for excitation and observe functions for all
experiments so that calibrations can be made once for any experiment,
either direct or indirect observe. Since Channel 1 is always for 1H, a pw90
calibration is the same for both modes. When 13C is pulsed it is via
Channel 2, for both modes. Channel 3, if configured with a preamp module
would serve for X pulses, decoupling and observe. Up to four preamp modules
may be mounted within the cabinet. INOVA has up to two preamp modules in a
separate assembly
Mixer
A universal observe mixer is present, receiving inputs from all preamp
modules. The preamps provide some gain and the remainder of gain is handled
in the mixer. Cold probes have internal cryogenically-cooled preamps and
their output is supplied to the mixer via a separate connector. No gain
functionality is present in the receiver (DDR) board. In INOVA, the preamp,
mixer and receiver board all have gain stages. Gain distribution has been
carefully adjusted to maximize dynamic range and minimize distortions.
2H Diplexer
The 2H diplexer box has been moved into the cabinet and is connected to
both the channel amplifier used for 2H decoupling and to the lock RF. The
RF is then supplied to the lock connector on the probe. 2H can be observed
directly in this manner (with tn=lk) for 2H gradient shimming, for example.
2H can be pulsed by the decoupling channel also, and pulse sequence
capability is present to have simultaneous 2H pulsing on the decoupling
channel and observe with tn=lk (overriding the default either/or
condition). INOVA usage requires swapping a cable to perform this activity
(useful for calibration of the decoupling channel by direct 2H observe).
Tune Box
Probe tuning is far more convenient and accurate. Each RF input to the
probe goes through a dedicated directional coupler. This device gives a
feedback that is used to produce a meter reading on the tune box. This
small box is attached to a long cable that permits convenient tuning at the
probe.
Pneumatics
Pneumatics has been re-organized into a more logical package. All
gas/air control devices are mounted in this one cabinet, including separate
pressure control valves for purge of probe, shims and upper barrel, VT
pressure and flow, anti-vibration legs, pre-conditioning unit (e.g. FTS)
and an accessory . The cabinet may be mounted on a wall near the gas
supply. Full solids operation and nano-probe operation is integrated. The
unit has its own power switch. Error detection is integrated and
communicated to the acquisition system.
Digital Control
VT gas control is controlled from the host computer via a liters-per-minute
control panel. The flow rate is detected electronically and high/low set
points can be set by the user for fault detection.
Safety
Considerable care has been taken to anticipate error conditions. Bypass
routes have been designed so that if a failure happens gas still flows
safely over the sample. If an FTS conditioner is used for low temperature,
operation above VTC sets up a small bleed stream through the FTS while a
direct stream goes into the probe. Similar protections are designed for the
bucket operation.
The Host Computer
The host computer would be a Sun or PC ( under Linux) using VnmrJ 2.1A
Software Version
The host computer (Sun or PC running Linux) uses VnmrJ 2.1A. The VnmrJ java
interface is standard (no classic interface). A number of new features have
been added since the 1.1D release
PSG
Previous INOVA pulse sequences will run without modification, even though
the underlying hardware is completely different. The go program assembles
independent control programs for each controller and distributes them over
the Ethernet connection. All BioPack, for example, run without
modification. Pulse sequences having corrections for AP bus delays will
still run, the corrections will just involve subtracting zero delays. New
sequences may be written without the need for these corrections.
No waveform generators are present, but waveform generation tools such as
Pbox, etc. are still used to create waveforms. These are used by the
standard PSG elements such as decshaped_pulse, obsprgon, etc. The shape
text files are still stored in shapelib. The go program will store the
instructions for executing the shape in the controller memory for the
specific channel used and these instructions will be stuffed into the FIFO
on the controller as needed. Simultaneous pulses or events of any kind on
multiple channels are handled as parallel events and are executed at the
proper time with only one synchronization signal given by the master
controller at the beginning of the experiment. The 64MB memory on the
controller provides a large resource for any shaped pulse or shaped
decoupling experiment.
Auto-phase incrementation has been designed into the controller hardware so
that shapes of any kind can be digitally frequency offset (SLP) without
preparation of a shapelib file.
Received on Mon Aug 15 2005 - 09:10:28 MST