Triggering and Synchronization

To begin with, we generate a trigger output with the Signal Generator channel 1. As this tutorial is an extension of the Basic Waveform Playback Tutorial, configure the SHFQC as follows:

After configuring the output using the table above, we use the SHFQC to generate a trigger output. There are two ways of generating trigger output signals with the Signal Generators AWG: as markers that are part of a waveform and played with sample precision, or by controlling trigger bits through the sequencer.

The method using markers is recommended when precise timing is required, and/or complicated serial bit patterns need to be played on the Marker outputs. Marker bits are part of every waveform, and are set to zero by default. Each waveform is represented by an array of 16-bit words: 14 bits of each word represent the analog waveform data, and the remaining 2 bits represent two digital marker channels. Hence, upon playback, a digital signal with sample-precise alignment with the analog output is generated.

Generating a TTL output signal using a sequencer instruction is simpler, but the timing resolution is lower than when using markers. The sequencer instructions play at the sequencer clock cycle of 4 ns, whereas the markers are part of the waveform and therefore have a resolution of 0.5 ns. The method using sequencer instructions is useful to generate a single trigger signal at the start of an AWG program, for instance.

Let us first demonstrate the use of markers. In the following code example we first generate a Gaussian pulse. This is identical as in the Basic Waveform Playback Tutorial, where the generated wave already included marker bits - they were simply set to zero by default. We use the marker function to assign the desired non-zero marker bits to the wave. The marker function takes two arguments: the first is the length of the wave in samples; the second is the marker configuration in binary encoding, where the value 0 stands for both marker bits low, the values 1, 2, and 3 stand for the first, the second, and both marker bits high, respectively. We use this to construct the wave called w_marker.

The waveform addition with the '+' operator adds up analog waveform data but also combines marker data. The wave w_gauss contains zero marker data, whereas the wave w_marker contains zero analog data. Consequently the wave called w_gauss_marker contains the merged analog and marker data. We use the integer constant marker_pos to determine the point where the first marker bit flips from 0 to 1 somewhere in the middle of the Gaussian pulse.

There is a certain freedom to assign different marker bits to the Mark outputs. The following table summarizes the settings to apply in order to output marker bit 1 on the Mark output of the Signal Generator channel 1.

Figure 2 shows the AWG signal captured by the scope as a yellow curve. The green curve shows the second scope channel displaying the marker signal. Try changing the marker_pos constant and re-running the sequence program to observe the effect on the temporal alignment of the Gaussian pulse. After the waveform has finished playing, the marker bit returns to a value of zero automatically, as no more waveform is being played.

Let us now demonstrate the use of sequencer instructions to generate a trigger signal. Copy and paste the following code example into the Sequence Editor.

Each AWG core has four trigger output states available to it. The setTrigger function takes a single argument encoding the four trigger output states in binary manner – the integer number 1 corresponds to a configuration of 0/0/0/1 for the trigger outputs 4/3/2/1. The binary integer notation of the form 0b0000 is useful for this purpose – e.g. setTrigger(0b0011) will set trigger outputs 1 and 2 to 1, and trigger outputs 3 and 4 to 0. We included a waitWave instruction after the playWave instruction. It ensures that the subsequent setTrigger instruction is executed only after the Gaussian wave has finished playing, and not during waveform playback.

We reconfigure the Mark 1 connector in the DIO tab such that it outputs AWG Trigger 1, instead of Output 1 Marker 1. The rest of the settings can stay unchanged.

Figure 3 shows the AWG signal captured by the scope. This looks very similar to Figure 2 in fact. With this method, we’re less flexible in choosing the trigger time, as the rising trigger edge will always be at the beginning of the waveform. But we don’t have to bother about assigning the marker bits to the waveform.

For this part of the tutorial, connect the cables as illustrated below.

In this section we show how to trigger the AWG with an external TTL signal. We start by using the Signal Generator channel 1 of the SHFQC to generate a periodic TTL signal. As shown in Figure 4, the Mark output of channel 1 is connected to the Trig input of channel 2. We monitor the marker and signal outputs of channel 2 on a scope.

Compile and run the above program on the AWG core of channel 1. Then configure the Mark 1 and Mark 2 to use Output 1 Marker 1:

Next we configure channel 2 to respond to the trigger generated by channel 1. Internally, the AWG core of each channel has 2 digital trigger input channels. These are not directly associated with physical device inputs but can be freely configured to probe a variety of internal or external signals. Here, we link the AWG Digital Trigger 1 of Channel 2 to the physical Trig 2 connector, and we configure it to trigger on the rising edge.

Finally, we modify the previous AWG program by adding a while loop so that the sequence can be repeated infinitely and by including a waitDigTrigger instruction just before the playWave instruction. The result is that upon every repetition inside the infinite while loop, the AWG will wait for a rising edge on Trig 2.

Compile and run the above program on the AWG core of channel 2. Figure 5 shows the pulse series as seen on the scope: the pulses are now spaced by the oscillator period of 8 μs, unlike previously when the period was determined by the length of the waveform w_gauss. Try changing the trigger signal frequency or unplugging the trigger cable to observe the immediate effect on the signal.

For this part of the tutorial, connect the cables as illustrated below.

In this section we will show how to use the Internal Trigger Unit to synchronize the outputs of multiple channels of the SHFQC.

To configure the internal trigger, set the following settings in the tables below.

The number of repetitions (ranging from 1 to more than 1e9) determines how many triggers will be sent out. The holdoff time (minimum 20 ns, resolution of 4 ns, maximum 4000 s) determines the time in seconds between the individual trigger events, typically chosen to be longer than the longest part of the sequence. For example, in a Ramsey experiment, the hold-off time might be slightly longer than the length of two pi/2-pulses plus the length of the longest evolution time and the length of the readout pulse. After entering the settings above, the Internal Trigger Unit is configured but is not yet sending out triggers. We will enable the triggers after uploading our sequences.

Compile and run the sequencer code below (which is the same as in the section above) to SG channels 1 and 2 of the SHFQC.

The sequencers are now waiting until they receive a trigger. Enable the internal trigger by clicking Run/Stop button in the Internal Trigger section of the DIO Tab. Figure 7 shows that the outputs of channels 1 and 2 are synchronized in time.