Pulse Calibration Workflow

Experiments of a 14 kDa IDP on 850 MHz.

13C and 15N 90-Degree Hard Pulse Calibration

lock
atmm: select “1H” and match/tune; then select “13C”; then “15N.” Finally, click Close and Store Values.
loopadj
topshim gui: set Dimension → 1D; Optimise for → 1H; After → Z-X-Y-XZ-YZ-Z; Only → yes. Click Start.
Once completed, set After → off and run:
```
topshim ordmax=7
```
This optimizes all parameters containing Z up to the 7th order (the maximum order of Z is 8). One strategy is to first roughly optimize Z up to 5th order, then the combinations of Y and Z up to 5th order, and finally Z up to 7th order.

pulsecal_fig1

re 1 to switch to the 1D ¹H PP; then P1 to check that the 90° high-power pulse length is correctly set. Next, run:
```
zg
efp
```
to verify that the signal from D₂O isn’t distorted (it must appear as a perfect symmetric bell shape).

pulsecal_fig2

Calibrate 13C 90-Degree Hard Pulse Length

Copy the ¹³C and ¹⁵N pulse-calibration experiments to a new directory. These contain a special pulse sequence: when the 90° high-power pulse length is optimal, the signal is zero; if shorter than optimal, the signal is positive; if longer, the signal is negative. Enter the ¹³C experiment and run:
```
pulsecal
```
to calibrate the ¹H pulse length (P1) first. pulsecal will also read and update P3 (the 90° high-power pulse for ¹³C) from the “prosol” table.

pulsecal_fig3

Run ased and inspect P3. It should be the default value from the “prosol” table (14.2 μs here). Set it to a lower value (e.g. 5.0 μs, or divide by 2) to get a spectrum you can phase—otherwise, a near-90° pulse will nullify the signal. Then run:
```
zg
qsin
ft
.ph
```
to acquire and phase the spectrum. Correct phasing is crucial, as this spectrum feeds into the next step (popt); without it, signals will have dispersion character (positive and negative peaks), making the 0-signal point ambiguous.

pulsecal_fig4

Zoom into the signal region to be optimized. Right-click → Save Display Region To… → Parameters F1/F2 → OK.

pulsecal_fig5

Run popt to open the parameter-optimization window. Set:
- GROUP → 1
- PARAMETER → P3
- OPTIMUM → ZERO
- STARTVAL → 12
- ENDVAL → 16
- INC → 0.4 Then click Save and Start Optimize, overwrite → y.

pulsecal_fig6

Run P1 and note its value (e.g. 13.07 μs). Do the same for PLW1 (e.g. –11.46 dB). These will be needed for getprosol.
Once acquisition finishes, run:
```
re 1 1
```
to invert the signal sign. Skipping this means spectrum “999” won’t load correctly.

pulsecal_fig7

Run:
```
re 1 999
```
to display a series of spectra (11 spectra with pulse lengths from 12.0 to 16.0 μs in 0.4-μs increments). If the optimizer doesn’t find an exact zero within these values, it interpolates. Here it reports an optimal P3 value of 13.382361 μs.

pulsecal_fig8

Run P3 and note its value (e.g. 13.4 μs). Do the same for PLW1 (e.g. –21.68 dB). These will also be needed for getprosol.

Calibrate 15N 90-Degree Hard Pulse Length

Run re 2 1 and then pulsecal. Compare the output 90° pulse value (13.10 μs) with the one previously obtained—they must be very similar.
Run ased and set P21 (the parameter controlling the ¹⁵N 90° hard pulse length) to a lower value, e.g. 10.0 μs, than the default from the “prosol” table (34.0 μs). Then run:
```
zg
qsin
ft
.ph
```
to acquire and phase the spectrum.

pulsecal_fig9

Zoom into the signal region to be optimized. Right-click → Save Display Region To… → Parameters F1/F2 → OK.

pulsecal_fig10

Run popt to open the parameter-optimization window and set:
- GROUP → 1
- PARAMETER → P21
- OPTIMUM → ZERO
- STARTVAL → 32
- ENDVAL → 36
- INC → 0.4 Then click Save and Start Optimize, overwrite → y.

pulsecal_fig11

Once acquisition finishes, run re 2 1 to invert the signal sign, then re 1 999 to preview the optimization. The result may look too noisy to infer an accurate P21 value—this is because only 4 scans were used.

pulsecal_fig12

Repeat the process with more scans. Run:
```
NS 16
popt
```
and narrow the pulse-length range: STARTVAL → 33, ENDVAL → 37. Then click Save and Start Optimize, overwrite → y.

pulsecal_fig13

When optimization finishes, run re 2 1 followed by re 2 999 (the former is required to load spectrum “999”). The 11 spectra with progressively lower signal will be better resolved, and the optimal P21 (e.g. 35.218787 μs) will be displayed—interpolated from the intensities.

pulsecal_fig14

Run P21 and note its value (e.g. 35.2 μs). Do the same for PLW3 (e.g. –25.55 dB). These will be needed for getprosol.

Apply the Optimum ¹H, ¹³C, and ¹⁵N Pulse Lengths to All Experiments

Now apply the optimum 90° pulse lengths of ¹H, ¹³C, and ¹⁵N to every experiment you wish to record. Run:

re 41
getprosol 1H 13.07 -11.46 13C 13.4 -21.68 15N 35.2 -25.55

Then, for each spectrum, repeat:

re 42
getprosol 1H 13.07 -11.46 13C 13.4 -21.68 15N 35.2 -25.55

re 43
getprosol 1H 13.07 -11.46 13C 13.4 -21.68 15N 35.2 -25.55

re 44
getprosol 1H 13.07 -11.46 13C 13.4 -21.68 15N 35.2 -25.55

re 45
getprosol 1H 13.07 -11.46 13C 13.4 -21.68 15N 35.2 -25.55

re 46
getprosol 1H 13.07 -11.46 13C 13.4 -21.68 15N 35.2 -25.55

If you had run only getprosol 1H 13.07 -11.46, the ¹³C and ¹⁵N values would be pulled from “prosol” and would be suboptimal for this sample.

pulsecal_fig15

Finally, verify that all spectra can be recorded without issues. For each experiment:

re 41
o1      # Check transmitter-frequency offset [F2,F1]
gs      # “Stop”

re 42
o1      # optional
gs      # “Stop”

re 43
o1      # optional
gs      # “Stop”

re 44
o1      # optional
gs      # “Stop”

re 45
o1      # optional
gs      # “Stop”

O1 should be identical in all experiments.

pulsecal_fig16

Run:
```
re 41
multiexpt 6
```
to estimate total experiment time. If satisfied, run:
```
multizg 6
```
to start the recordings.

Bottom-up 13C and 15N 90° Hard Pulse Calibration on 850 MHz by Me

lock
On high-field cryo-probes, Bruker recommends doing automatic tuning-matching from the nucleus with the highest frequency to the lowest, namely 1H → 13C → 15N. You can either select the nuclei manually or, better, issue:
```
atma high
```
to do it automatically.
Once finished, hit:
```
atmm
```
for manual refinement. Optimize the 15N. The Wobble sweep width of 13C is very wide (12.017 MHz by default); therefore, we must zoom in on the baseline of the signal. Set it to 4.0 and click Set, then refine the position of the minimum. Likewise, set the Wobble sweep width of 1H to 4.0 MHz (default: 12.754 MHz) and refine manually.

pulsecal_fig17

topshim
loopadj
topshim gui: set Dimension → 1D, Optimise for → 1H, After → Z-X-Y-XZ-YZ-Z, Only → yes. Click Start. The “3D” option is more thorough but much slower; it’s usually unnecessary unless you’ve recently replaced the probe. It helps for water-containing samples in wider tubes (e.g. Shigemi 5 mm), but in a 3 mm tube there’s little X-Y inhomogeneity to correct. To save time—since we already completed the 1D optimization—issue:
```
topshim ordmax=7
```
to optimize all Z combinations up to 7th order (default is 5th).

pulsecal_fig18

Let’s measure and check the 1D proton now. Run:
```
re 1
ased
```
The existing P1 value (13.130 μs) is too high and will give a very strong signal (probably left by a previous pulsecal or getprosol). Lower it to 1.0.

pulsecal_fig19

Run:
```
zg
efp
.ph
```
for manual phase correction. Zoom into the peak and verify that the shape is symmetric.

pulsecal_fig20

Calibrate 13C 90° Hard Pulse Length

Nothing new.

Calibrate 15N 90° Hard Pulse Length

Nothing new.

Apply the Optimum ¹H, ¹³C, and ¹⁵N Pulse Lengths to All Experiments

Run re 1 and zoom in on the peak region. Place your cursor in the middle to see the maximum O1 value (water’s resonance frequency in this sample), here 3991.1 Hz. However, the optimum may deviate by a few Hz; therefore, issue:
```
o1calib
```
which calculates it. Then apply the output O1 value to the spectra you want to measure and, using:
```
gs
```
assess the FID shape to decide the exact O1. This is important because experiments with water-signal suppression work best when O1 is exactly at the water resonance. Here we measure only gradient-selected experiments, where it’s less critical; regardless, it’s ideal to set an accurate O1.

pulsecal_fig21

When it finishes, run:
```
O1
```
to display the found value; in this case, 3994.21 Hz.

pulsecal_fig22

Execute getprosol with the ¹H, ¹³C, and ¹⁵N calibrated parameters for all experiments to be measured.
Run re 8 to switch to the 3D experiment and experimentally find the optimum O1 value. Hit gs, go to the Offset table, enter the O1 value from o1calib (3994.21 Hz), and press Enter.

pulsecal_fig23

Then enter the previous value, 3991.10 Hz, and press Enter. You will see a less wavy FID, indicating it is better than 3994.21 Hz. This occurs because the modular differences between observed and resonance frequencies are minimized, giving a smoother FID. In principle, one could manually alter O1 and observe its effect, but here it is redundant. Finally, click Stop.

pulsecal_fig24

Update the O1 value in all experiments you plan to run.

Authors

Thomas Evangelidis