University of Ottawa NMR Facility Web Site

Please feel free to make suggestions for future posts by emailing Glenn Facey.

Monday, April 14, 2008

Tuning and Matching an NMR Probe

The circuits in NMR probes are essentially band pass filters. The properties of the band pass filter are determined by the values of capacitance for the capacitors in the circuit and the inductance of the sample coil. One is usually able to vary the capacitance of the tuning and matching capacitors to optimize the probe for the experiment at hand. The adjustments are typically made by turning rods that extend from the bottom of the NMR probe. These rods are often labelled "T" for tune amd "M" for match. Changing the tuning capacitor will shift the band of the bandpass filter and is therefore used to tune the probe from on frequency (or nucleus) to another. This is much like tuning your radio to a particular radio station. The tuning control of a radio is also a variable capacitor. Changing the matching capacitor changes the efficiency of the band pass filter. These two capacitors are not independent and they must be adjusted iteratively to optimize the NMR probe. Proper tuning and matching are essential in getting high quality NMR results. The figure below shows the response on an oscilloscope for an NMR probe connected to a sweep generator. One can think of these plots as frequency on the horizontal axis vs efficiency (or match to 50 Ohms) on the vertical axis. The horizontal line represents a 50 Ohm reference. The first panel shows a probe that is well matched and tuned. The second panel shows a probe that is properly matched but poorly tuned and the third panel shows a probe that is properly tuned but poorly matched.


Anonymous said...

excellant nmr web

umayalrakesh said...


nice notes... the vertical axis in the scope is the efficiency of the power rite? i couldnt understand the 50 ohm reference.. wats the relation between efficiency, resistance and inductance ?

Glenn Facey said...

Dear Umayalrakesh,

Thank you for your comment. NMR probes are matched to 50 Ohms at their resonant frequency. The plot on the scope is created by taking a swept RF output from a sweep generator through port 1 of a "magic tee". ports 2 and 3 are connected to a 50 Ohm reference load and the NMR probe, respectively. The output at port 4 is connected to an rf detector on the sweep generator. An X and Y signal from the sweep generator is sent to an oscilloscope which is used as an X Y recorder. The X axis is frequency sweep and the Y axis is the rf output. When ports 2 and 3 are impedence matched (i.e. both 50 Ohms), the output is zero. When ports 2 and 3 are impedence mismatched then port 4 sees an output signal. The XY plot therfore resembles an absorption which can be used to tune and match an NMR probe.

See Fukushim and Roeder, "Experimental Pulse NMR: A Nuts and Bolts Approach" (Addison-Wesley, 1981) Chapter V section C.7 p,392.

ടിറ്റോ said...

My name is Titto. could you please tell me why in HMBC experiment, some of the correlations are not seen (some time i dont even see 2 bond correlations). i think it is connected with the sensitivity of the NMR. tuning has any effect? I am using Bruker 300 NMR, and HMBC experiment programme is . Please answer my dobt.

Glenn Facey said...


Thank you for your comment. If your NMR probe is tuned well and the HMBC is set up correctly for the long range coupling you are interested in, then the main reason correlations may be missing is if the couplings are very close to zero.

ടിറ്റോ said...

hi Glen,
thank you for the advice. sometimes, even the coupling with 8Hz which is seen in COSY is missing. i tried to change CNST value to 8Hz, and 2 Hz. the tunning is fairly good (tune 1 LED, matching 3 LED, bruker 300 MHz). could you tell me which could be the best HMBC programme?

Glenn Facey said...


The 8Hz 1JHH you observe in a COSY spectrum has nothing to do with the (2-4)JCH you measure in an HMBC spectrum. I don't think your issue is due to probe tuning. It is more likely due to sensitivity. HMBC experiments use fairly long delays to accomodate small couplings. During this delay there are significant T2 losses. Is the correlation you cannot find between a broad proton resonance and a carbon?

Catherine said...


I have a question about keeping your probe tuned during an experiment. I typically use high power for my experiments and I need to keep the tuning the same for the different samples that I run. I notice that my reflected voltage will increase 40 mV (or more) than what I started with at the beginning of the experiment and when I hook an spectrum analyzer I see the tuning change a little. How do I prevent this from happening? I thought maybe using a longer last delay (pulse delay) because I thought things were getting heated up. Maybe teflon tape? Not sure what to do.

Any advice would help. Thanks.

Catherine Hedge

Glenn Facey said...


I have noticed similar behavior as well while doing solids experiments. The way I have handled it is as follows: Tune the probe after putting in the sample. Run you pulse sequence for about 5 minutes. Halt the acquisition and immediately retune the probe. Restart the acquisition.


Anonymous said...

Hi Glen,

I am Lin. Previously, I used Bruker SS NMR. when I need to tune, I will do the sam e thing like u said, and will try to get the best graph from the pc screen. Is taht mean my probe was connected to sweep genarator and oscilloscope? and now, I am using JEOL, when I need to tune, i will do the same as Bruker, but it is only monitored by the relection meter.

and one of my friend asked me if i used external equipment to tune the probe. This question makes me a little bit curious. I would liek to know what exactly she meant by extra equipment?

Do you have any idea? or may be she is refering to the osciloscope thing?

thank you for your time.


Glenn Facey said...

Hi Lin,

Thank you for your questions.
When you tune your probe on a Bruker spectrometer you are using the spectrometer's electronics to tune the probe. Bruker spectrometers produce output very similar to that of the oscilloscope and sweep generator I showed in the post. If you are measuring the reflected power on your JOEL spectrometer, you are trying to minimize the power reflecting back to the amplifier. When the reflected power is minimized, the probe is well tuned and matched. You are not using any external equipment. One can however tune and match a probe using a seperate sweep generator and oscilloscope (as I have in the post) or a self contained device like the one provided by Morris Instruments. One could also use a spectrum analyzer.


Anonymous said...

Hi Glenn,
My name is Jerry, and I currently have issues with tuning my 500Mhz proton channel on oscilloscope.

Its showing two signals instead of the one in the center. I know my question may be silly, but what would be the reason for this ? Possibly bad insulation ?

Thanks very much!

Glenn Facey said...

Hi Jerry,

I have seen this before on NMR probes as well. It could be that the proton and X channels of the probe are not sufficiently isolated from one another. Try changing the X tuning while looking at the proton tuning on your oscilloscope to see if the situation improves. You could also take the cover off the probe and make sure that nothing is being shorted out against the probe cover. Is this a new problem?


Josh said...


I have a question regarding tuning using an oscilloscope. Why is it that the reflected power is so large at the edges of the pulse, and does the disparity in reflected power at the pulse edge versus the center of the pulse effect the final pulse shape?

Glenn Facey said...

Hi Josh,

I don't know the answer to your question about why the reflected power is higher at the edges of the pulse. You can measure the pulse shape directly with an oscilloscope to see if the shape is affected. See this post:


Unknown said...

Hi gelenn
I have been reading your your posters, from which I can always find answers to my doubts.

Here I got a naive question--- the Mtune and tuning interface, which is more trustworthy when the outputs are not the same. Sometimes, the display of Mtune does not reach lowest point when the tuning interface shows Zero, vice versa...


Glenn Facey said...

Hi Liming,
I know nothing about MTUNE (Varian instrument?) so I am unable to comment.