series.doc>>12/15/99
Acquiring a Series of Spectra
(ARX Spectrometers)
I. Xwinnmr contains a utility that will run a series of experiments. To use it, all you have to do after you have the acquisition parameters set up appropriately is to type multizg instead of zg. You will have to specify how many experiments you want to run. There is a variety of reasons to start the acquisition with multizg rather than zg, and several are described below. First it is important to understand what multizg actually does.
A. Multizg starts the acquisition that is set up in the current data set you are displaying when you type the command. It will finish that acquisition, then proceed to the next higher experiment number (exp. no.), start that acquisition, then go to the next higher exp. no., etc. The process number (proc. no.) will not change. For example, if you are displaying a data set identified as sucrose exp. no. 6 process no. 2, and you tell it to run 4 experiments, you will run:
name exp. no. proc. no.
sucrose 6 2
sucrose 7 2
sucrose 8 2
sucrose 9 2
The total amount of time required for these experiments will simply be the sum of the amount of time that would be required for each. Type expt to find out how long it will take to acquire the data set currently displayed.
B. Multizg will behave slightly differently depending on whether the following exp. no.'s already exist before you start the multizg.
1. If they did not already exist, multizg will copy all of the parameters of the experiment you started it from into the following exp. no.'s before it starts the acquisition. This is a very useful way to run a series of identical spectra. (Unfortunately, there is a stupid feature that when you start multizg, you may get error messages that say that you do not have permission to write to /tmp/expt. If you just select OK for all these messages, it will run anyway.)
2. If they did already exist, multizg will just run whatever is already set up in those exp. no.'s. This is a very useful way to run several different experiments on the same sample, e.g. a COSY, followed by a NOESY, followed by an HMQC, etc. Of course, you will have to set up all of the acquisition parameters properly before you start the multizg.
II. Examples of reasons to run multizg:
A. When signal averaging for a long period of time, it can be advantageous to acquire data for shorter periods of time in separate files, then add the files later. If something happens to the spectrometer or the sample during a long run, at least part of the time spent acquiring data could be useful. In this case, you would set up only the first exp. no. in the series and make sure that following exp. no.'s do not exist before you start. Type expt once all the parameters are set, then multiply by how many experiments you plan to do and you will know how long the entire run will take to complete.
B. An obvious extension of case A. is a kinetics experiment where you monitor a change in the sample. The major difference is that you would use a slightly different pulse program from the one in the default files that has an extra delay inserted after ns scans have been completed. (See NMR lab personnel for guidance on which pulse program to use for your application.) In this way, you can set an interval between the acquisition of the data sets. Obviously, in this case you would not add the files later.
C. If you need to run several different experiments on the same sample, you will not have to come to the lab in between if you set up all the acquisition parameter sets you need in increasing exp. no.'s before you start. This can be a much more efficient use of spectrometer time compared to running each experiment separately. Type expt while displaying each data set and you can calculate the total time the series of experiments will require.
III. Other considerations:
A. To halt a multizg before completion of the number of experiments that you specified, first type halt, wait for the message in the bottom line of xwinnmr that the data have been stored (usually it just says checklockshift, but that's it), then type kill and select the entry containing multizg.
IV. Data Processing
B. It is never a bad idea to look at each individual file after it has been processed even if you plan to add them up. The easiest way to step through them quickly is with the dual display. Type dual and you will probably get a message that says the 2nd data set is not defined. Click OK then you will be in the edc2 menu. Input the correct specification for the 2nd data set. (Usually it will have the wrong exp. no. and proc. no. by default.) You don't have to input anything for the 3rd data set. Select SAVE. Both the original data set and the one you specified as the 2nd data set will be displayed together. An additional window will identify the 2nd data set. The buttons on the left can be used to manipulate both spectra together or just the 2nd data set. Click on the button labeled e to change to the next higher exp. no. Continue to click on e to step through the rest of the series.
C. To add a series of files containing raw data (fid's), read in the first file in the series then type fidadd. You will have to enter information to identify the first data set in the series, but if you typed the command from that data set, the default values will be correct. You will also be asked to specify the total number of fid's to be added. The summed data will be placed in the first unused exp. no. For example, if you had a series of files from 1 through 20, the sum will be placed in exp. no. 21. (Even if there are no data in a parameter set, if the exp. no. exists, the summed data will be placed after all those exp. no.'s that existed previously.) The total number of scans will be correct in the summed fid file. Once you are sure that the summed file is all that you need in the series, delete the individual files used to produce the sum. If you acquired the series on disk y, be sure to copy the summed data elsewhere or it will be deleted within 7 days. (See III. A. above.)