Afonso Teles

Hi Jeff,
Yes, there is a significantly newer version of DPOJET available.
DPOJET is updated together with the firmware of the machine. Here is the latest version of the firmware (and DPOJET): https://www.tek.com/en/support/software/firmware/firmware-dpo7000-dpo-dsa70000-dpo-dsa70000b-c-mso70000-c-doc-win-xp

On measuring relative jitter between two clock source's, that is possible by selecting a measurement like TIE, RJ or DJ, then going into "Configure", "Clock Recovery" and setting the Method to "Explicit clock - Edge" and selecting the desired clock channel.

There was a known problem with measurements on the TBS2000X.
This has been fixed with the latest firmware release. I ask that you install it and try again: https://www.tek.com/en/support/software/firmware/tbs2104x-firmware-release-v132136

"curve?" data from spectrum view should always be in 8 byte floating point format, so data:width doesn't need to be set explicitly.

Vrms is a measure of the voltage integral adjusted to allow for easier measure of power.
dBm is a measure of power relative to 1 mW.
The relationship between them is giving by the usual P = V^2/R relationship. The dBm display on spectrum view assumes a 50 Ohm load, this may or may not be true.

The fact the first harmonic is at ~13.5 dBm comes from the inherent properties of applying the Fourier transform to a square wave. I will not go through the math here, but there are plenty of sources discussing the Fourier transform of a square wave.

I do not know how your LabVIEW program works, but I assume you simply passed it the data from the "curve?" command to a plotting function.
It then picked the highest point as 0 dB (not dBm). This explains what you are seeing.
This seems like a quirk of how the LabVIEW plotting function works.

Hi,

The lowest measurable frequency is the inverse of the acquisition time (which is ten times the time/div). This means that to get measurements for lower frequencies, you should increase the time/div.
To do this you can go into the Horizontal/Acquisition menu and set the Horizontal mode to Manual and then increase the record length (increases processing times) or, if that isn't enough, the sample rate. Make sure not to drop the sample rate to the point that it might cause aliasing of your signal. In your specific case, the sample rate should stay at 25 MS/s and above.

Measurements down to ~10 Hz should be possible, with 25 MS/s sample rate and 250 M record length.

Hi,

Correct me if I'm wrong, but I presume you are using a DPO/MSO/MDO5034 and not the TDS5034.

A fresh install of Windows might be worth considering if using LabView is important to you.
I would suggest trying TekScope Utility. It should allow you to easily save waveform data manually or on trigger or timing events.
If it does not do what you need it to do, then you may want to try sending SCPI commands manually through OpenChoice and see if the data you receive matches with what you'd expect. Then it's a simple matter of stringing those together into Python or some other language of your choice.

If you post your LabView program we might be able to provide more direct help on it.
I would also point to this post that features several known working LabView remote control examples.

Hi Jeremy,

Just to be clear, the data is already in log form if using dB units. If using Vrms then the data will be linear.

The data follows the units that spectrum view is set to.

"DAT:WIDTH 8" Is actually not necessary as when we set the data:source to CHx_SV_NORMAL, it automatically sets the data width to 8. This is because the spectrum view data is encoded as double precision floating point, which takes up 8 bytes per point, as mentioned on page 2-420 of the programmer's manual.

So it's there just to make it clear to any reader that the data is 8 bytes per point.

2. That is correct, the number of data points available is equals to the record length that the instrument is configured to.

3. I have attached a Python script I wrote that exemplifies how to read the spectrum waveform data from the scope. This can be a bit tricky as it is a double float and needs to be converted into a usable format from its packed binary representation.
To your initial question, no you should not apply a log function to this data as it is already in log form. From page 2-429 of the manual:

> CHx_SV_NORMal, CHx_SV_AVErage, CHx_SV_MAXHold, CHx_SVMINHold: When DATa:SOUrce is set to one of these traces, the data points represent the amplitude of the Spectrum View frequency domain traces in the current units, and are floating point values

1. Yes, it is possible to transfer both the time domain waveform and the spectrum view data in one trigger event. This might require stopping acquisition to allow time to transfer both waveforms.
The argument for DATA:SOURce should be "CHX_SV_NORMAL", as specified on page 2-439 of the programmer's manual for the MSO4.

How are you processing the data? It should be floating point data (as mentioned on page 2-429), so they take a bit more work to read than time domain waveform data.

2. That is up to how many points you wish to read. If you wish to read all the points then you can send "DATA INIT" to set it to the default settings.

3. I believe this issue goes back to I discussed in 1., that the data is floating point and so needs to be read differently from time domain waveform data.

Hi Jeremy,

They are significantly different. The following comparison explains it well: <a href="https://www.tek.com/en/documents/fact-sheet/3-series-mdo-spectrum-analyzer-vs-oscilloscope-fft-comparison">https://www.tek.com/en/documents/fact-sheet/3-series-mdo-spectrum-analyzer-vs-oscilloscope-fft-comparison</a>

Main points to take away:

- ~15 dB better dynamic range than scopeFFT

- Significantly higher bandwidth (Up to 3 GHZ)

These may, or may not, be important to you, depending on what measurements you are taking.

One hour is quite a long acquisition for a scope. This is the kind of measurement that is usually best done by a digital multimeter.
There is also a question of the sample rate you need this acquisition to be done at, which you did not specify.

I would suggest using TekScope Utility (https://forum.tek.com/viewtopic.php?t=140451) and using its Data Logger feature to record several waveforms that add up to one hour of acquisitions.