Author Topic: Keysight/Agilent 81160A: Highest time-resolution?  (Read 286 times)

0 Members and 1 Guest are viewing this topic.

Offline frizensamiTopic starter

  • Newbie
  • Posts: 2
  • Country: sg
Keysight/Agilent 81160A: Highest time-resolution?
« on: February 06, 2024, 05:16:05 am »
Hey all,

I've been trying to generate a series of 10ns-width pulses that have precise delays from each other from the 81160A for a while now. In general, the signal is:

1. One 10 ns width pulse
2. Wait [(2304 ns * M) + N] (M is just an integer, N is a delay that should have at least 1 nanosecond resolution, ideally higher)
3. Repeat 1 and 2

The issue i'm facing is not getting a high enough time resolution for these delays from 81160A. The two options I've tried are:

1. Store and output an arbitrary signal

While I only have around 200 pulses total, the entire length of the signal is 3.73 ms. With an equally sampled signal at 1 ns, this requires 3.73 million points in memory, but my 81160A only can store 131,072 points. 3.73 ms / 131,072 points gives me a time resolution of 28.4 nanoseconds, which is too coarse for my application.

2. Store and output a pulse pattern signal

I have tried representing each part of the signal by a bit, and the device has enough memory for this representation (2M bits), but the smallest timescale is 3.03 nanoseconds which is again just above what I need.

For context: these are signals to test a LIDAR sensor, which is why the time resolution is important. I can also sacrifice the accuracy of the pulse width if necessary (e.g., by generating a much longer / more jitter-y pulse) as long as the start of the pulse is timed correctly.

I'm guessing this is more a user error than a device limitation since the 81160A can do 2.5 GSa/s, and I'm pretty new to this. Does anyone have any idea how I can get the signal I need?

Thank you very much in advance.
 

Offline 2N3055

  • Super Contributor
  • ***
  • Posts: 6451
  • Country: hr
Re: Keysight/Agilent 81160A: Highest time-resolution?
« Reply #1 on: February 06, 2024, 11:55:56 am »
Hello.

If I can understand you properly, you need:

A burst of 200
10ns impulses
with period of [(2304 ns * M) + N] (M is just an integer, N is a delay that should have at least 1 nanosecond resolution, ideally higher)

A burst should repeat every 3.73 ms.

Is that it?

What I explained here should be easily acomplished by using exactly that flow:

Set pulse waveform, set pulse width to 10ns, and period to whatever you need. This period is your hi res interval.
Enable burst mode, 200 pulses, and set repetition rate to 3.73 ms.

Or did I misunderstand something?
 

Offline frizensamiTopic starter

  • Newbie
  • Posts: 2
  • Country: sg
Re: Keysight/Agilent 81160A: Highest time-resolution?
« Reply #2 on: February 07, 2024, 02:37:59 am »
Ah I think I wasn't clear enough in my original post. What I intended to say was that the values of M and N change between each pulse. That is, the signal can be thought of as a sequence of say, 200 (M, N) values.

Another simpler way of thinking about it perhaps is:
1. Generate 10 ns pulse
2. Sleep for an accurate period of N nanoseconds (ideally sub-nanosecond resolution)
3. Repeat steps 1 and 2 with differing values of N, around 200 times

I hope that helps!

The trouble I'm facing is that while my signal looks pretty straightforward, because of the low maximum duty cycle (10 ns / 2304 ns at the least, approx 0.4%), when equally sampled for AWG memory, most of the points are "wasted" as 0 values. 
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf