As promised, here is the summary of comments I received regarding the
Varian VAST and robot autosampler. I have simply copied the six
responses below. I believe I have removed all identifying info from
them.
Additionally, I will make some comments about my recent demo trip to
see these two devices in operation in the real world. I observed the
operation of a SMS 50 sample robot autosampler and Gilson/VAST with HCN
flow probe at a well known chemical company and the Gilson/VAST with a
HCN flow probe at a biomed company. These companies have no
significant complaints about either of these devices. They both work
well and dependably. The only problem experienced with the robot has
been sample breakage BY THE CHEMISTS INSERTING THE TUBE IN THE SPINNER.
The only problems found with the Gilson handler were corrected by
adopting an appropriate protocol, as follows: Residual wash solvent
CHCL3 causes a large solvent peak. Solution; part of the sample line
was not being subsequently washed with CDCL3. Solvent evaporation
causes degradation of lineshape toward the end of the 96 well sample
rack. Solution; cover rack with AL foil. The chemical company has
experienced no clogs in the sample lines. One should not change
solvents during a given run, eg CDCL3 followed by H2O will surely
precipitate something in the sample line. The biomed company handles
blood serum samples. They did experience clogs initially, but have
elimintated them by using an appropriate wash protocol. Both companies
report no detectable carryover of one sample into another. Both have
developed their own software interface for acquiring spectra because
Varian's earlier versions (eg VNMR 4.x) were inadequate. However,
VNMR 6.1B is supposed to be much more versatile in this regard. Both
companies are happy with the performance of these two devices and
readily recommended them to me.
*****************
I have two carousels, one on an Inova-500 and one on an Inova-400.
Neither worked well right out of the box. However, with some effort
involving some tweaking of parameters, a visit from an installer, and
some phone help we finally got things going pretty well. They both work
flawlessly.
We purchased the carousels because we didn't need/want the expense
associated with the larger sample changer and VAST was not a viable
alternative in our lab (yet). In my experience they work well to fill
up night time slots and so we use them for that purpose. On my 500, I
can load up a number of similar samples and run them from my home.
The carousel is less useful in a walk-up environment where there may be
the potential for using the carousel for some things and managing
samples individually. Indeed, I would say this is a significant
disadvantage of the carousel.
*****************
I have two types of sample handlers here. For most systems we use a
Zymark robot and 5 mm NMR tubes. These systems have been extrememly
reliable. I do not use Varian's software but have written my own (We
actually were using Zymarks robots about 1 year before Varian started
to use them).
I also have a Gilson with a flow probe that runs samples for the
combinatorial chemistry group. We use only one solvent (25/75
CDCl3/DMSO-d6). We cover the plates with a thin sheet latex and use a
modified plate cover to hold this in place. We do have occasional
problems with clogging, samples must be carefully prepared. The worse
part is there is no automatic detection of a clogged system and
automation continues if not stop manually. In my experience I found
that trying to use multiple solvents can be a big problem. The transfer
tubing seems to absorb some of just about any solvent you use and when
switching to a different solvent the original solvent still shows up.
This will not be very noticable if you are using deuterate solvents,
but if you use protonated solvents it could be a problem.
*******************
We've had pretty good luck with our Zymark robot system on the Unity+,
which I realize is not the same as the two systems you mention, but
hopefully the newer carousel is even better. The robot functions
well with about zero maintenance for the past two years. The
automations setup is like a lot of other Varian stuff--you can do
almost anything but figuring it out can be tricky. We don't
challenge the system much either: we never do VT work (but often
do high temperature work where all of the automation samples are at the
same T) and we always load the system with samples having the same
solvent and composition. Under these conditions, I would say we
get about a 90% success rate for generation of adequate spectra.
Some things that go wrong: autolock doesn't get a strong enough
signal so the sample is rejected. Sample fails to spin so the
sample is rejected. Air flow for sample ejection isn't quite
enough, so the robot doesn't pick up the sample and quits. Robot
picks up sample, but due to low eject air, picks up sample too high on
tube and breaks tube while trying to rerack it. Automations run
fails due to syntax error in sample entry file (a problem if somebody
manually edits the file--it's very picky about things like
whitespace). Robot has a location set and somebody places a
sample in that spot (it's not obvious at the console where the robot
wants to put the sample). Almost all of these problems can be
fixed with the proper care in operation.
*************************
We bought an INOVA400 with a robotic arm (Zymark) sample changer last
year. We had some initial problems with alignment and positioning. The
problems were mainly due to the fact that the magnet was on
antivibration legs (old fashion doughnut type) and it's position was
changing based on the LN2 and LHe content which changed it's weight.
After adjusting the legs to minimize this position change, we have not
had any more problems. The robot control system (software) locks up
occasionally, and we need to reboot the system to take care of it. The
operational software for the robot on the VNMR has a few glitches and
problems, but non are very serious. As far as CDCl3 evaporation, since
most of the samples for which we use the auto sampler are for H-1
observe that take only a few minutes per sample, we don't have to worry
too much about solvent evaporation.
*************************
I exclusively use the Gilson with a VAST-type environment. Have used it
for about a year. No major cough-ups this far.
>>Do they have frequent mechanical problems?<<<<
None this far; a syringe leaked, swapped out, trivial.
>>And - when they work do they work properly? <<<<
Yep.
>>Does the Gilson device:
make errors in sampling:
Nope, although Varian's VAST automation software is very limited in
sampling patterns that are allowed. for e.g., sample #1 must be in top
right or top left corner and other such idiosyncracies which
combi-chemists can live without. Given a sampling pattern, I have never
seen an error.
>>suck air into the line and have associated problems, get clogs in the
line,
get
encrusted with deposits:
well, this depends on lots of other co-factors; I would say the Gilson
is the least trouble-making of the various components that make up
VAST. The source of clogs is (besides a junky sample, of course)
transfer lines, probe inlet port, injection port/ valve on the Gilson,
which, if all one does is direct injection NMR (not LC), one does not
even need. There is a fair bit of redundancy with the default Varian
setup. Make sure you get a newer style probe which rinses better in 1
cycle. We have had lots of problems with the older probe design.
Apparently, the newer design gives higher S/N as well as wider inlet
ports. I hear good things about the Nalorac probe as well. If you do a
spec test at Varian, drive it to the edge with DMSO. If DMSO is fine,
all else will be fine.
>>need disassembly for cleaning <<<< Yes, it does, but if you can write
that
very cool article on NMR specs, you can clean out a clogged Gilson!
>>How do you prevent evaporation of volatile solvents such as CDCl3 or
acetone in a 96 vial sample tray? <<<<
All plates are always run covered, at least in Al-foil, or polyprop
mats which are pierced with a specific Gilson needle. Interestingly, I
have not had a problem with volatile solvents. Remember, WET-supression
on CHCl3 at 7.27 will obliterate and/or distort baseline in the
aromatic region which can really piss chemists off!
Finally, I found VAST to be a very learn-as-you-go deal. I will warn
you that there are exactly 2.5 individuals at Varian that can answer
VAST-related questions; this is a real problem especially if you are
running a facility all by yourself, like I am.
Varian's VAST automation software is a disaster since it is a
hodge-podge of legacy Zymark-type stuff, old, old VNMR4.x stuff and new
Tcl scripts. Tech-Support/advice is very thin since only 2.5 folks what
what the heck VAST is all about. Be prepared to write a bunch of Tcl,
!sh, and such.
All in all, a pleasurable experience, once you have trouble-shot the
darned thing. At that point, you may never want to return to tubes.
****************************
I have an autosample carrousel (9 samples) in our Gemini 2000 (300
MHz). I'm very happy with it. It has only little problems, some times
it hangs up with no apparently problem, but it has never had mechanical
problems. From time to time I must adjust the air presure.
In two or three months I will install a 50 autosampler handling in our
Gemini 2000 (200 MHz), if you get news about this robot, I will be very
grateful if you send it to me.
Dr. Joseph Vaughn
Associate Director, NMR Facility
Chemistry Department
Florida State University
Tallahassee, FL 32306-4390
Phone 850-644-3334
Fax 850-644-8281
email vaughn@pro.sb.fsu.edu