Hantek 6022BE Data Timebase

jbh_ooi

I am using a Hantek 6022BE PC USB oscilloscope to record time evolution of signals. When I saved the data to .txt or .xls, the data shown only shows the y values in units of V. I need to be able to correlate these y values to a clock time because I am measuring quantities like period, frequency, pulse width etc.

In the .txt, I only have the CLOCK and SIZE. How can I convert these to time intervals. For example, CLOCK = 50.0 ns, SIZE = 1016.

Since the time axis on the software panel has 10 DIV, I thought that the time interval between each sample will be equally-spaced between 1016 points for a total time of 50.0 ns * 10. However, I noticed that the trace that was recorded is actually longer than 10 DIV. How do I know what the conversion ratio is? How about when SIZE = 130048?

I would appreciate any guidance on this, thanks a lot.

jbh_ooi

UPDATE: I have found the solution to the problem, and will share it here for anyone who needs it.


I have found this blog page useful, and have drawn inspiration from the discussion which led to the solution.
https://www.eevblog.com/forum/te ... e-20mhz-usb-dso/75/

Important information required to set the time interval between the sample points which can be found from the image by ying, as well as the exported data:
- Sampling rate (from the image)
- CLOCK (time/div)
- SIZE (buffer size)

From the blog post, I noticed a trend in the sampling rates set by the oscilloscope. When time/div = 2 μs, the sampling rate is set to 48 MHz. Even if the time-base is faster (< 2 μs), the sampling rate is limited to 48 MHz. ying has helpfully compiled the sequences of time/div into different rows according to the sampling rates and buffer size.

To get the time interval between sample points, consider these variables:
- CLOCK (the current time/div)
- ONSET (the largest time/div in the 48 MHz sequence/row, ie. the last value in the cell)
- RATE (the current sample rate)
- SIZE (the current buffer size)
- DIV (the total number of divisions in your data)

Firstly:
SIZE = number of samples/div * DIV

Since the 48 MHz sampling rate is the scope limit, we wil consider the onset where this limitation starts. So:
number of samples/div = RATE * ONSET

Substituting and rearranging gives:
DIV = SIZE / (RATE * ONSET)
Total time = CLOCK * DIV

Finally:
Time interval = Total time / SIZE = CLOCK / (RATE * ONSET)  --- [48 MHz]
For sample rates below the maximum rate, I suspect that the oscilloscope should be working optimally. In that case, ying's suggestion should work. Therefore:
Time interval = 1 / RATE  --- [all other sample rates]

This has given me the correct results so far (I haven't tested all rows because I don't necessarily have data in those ranges). For example,
- CLOCK = 50.0 ns
- RATE = 48 MHz
- SIZE = 1016
- Time interval = 50.0 ns / (48 MHz * 2 μs) = 0.521 ns
Hence, the time axis will have values like 0, 0.521 ns, 1.042 ns, ..., 528.815 ns. 1016 time points in total.

Hope this helps everyone.

ying

CLOCK: time/div
SIZE: the sample points, when the time/div is 10ms, the buffer size is 512K(512 sample points).
Unit: the unit of voltage.
The sample point has no relationship with X axis. The first sample point is saved at zero moment. The following sample point is saved  according to equal time interval.
time interval = 1/Sample Rate, when the time/div is 10ms, the sample rate is 1M. So the time interval is 1μs.

Then you can get one set of results andplot a graph. (X: sample point, Y: voltage)

jbh_ooi

ying replied at 2022-4-8 11:52
CLOCK: time/div
SIZE: the sample points, when the time/div is 10ms, the buffer size is 512K(512 samp ...

Hi, thanks for the input. I tried what you suggested but the results didn't come out correctly for me. I have a sample data having the following data:
CLOCK=50.0nS
SIZE=1016

According to your answer, the time interval for this set of data will be 1/48MHz = 20.833 ns. That means I can write out my x-axis values as (0, 20.833 ns. 2*20.833 ns, ... , 1015*20.833 ns) correct?

This didn't give me correct results, and it also did not look like the trace that I saw from the software screen. The largest x-axis value from the software should be at least 500 ns, but using your method, it is about 2 μs which is 4 times higher. Using your method also implies that the time interval between each sample point will be the same regardless of the time/div if they have the same sampling rate. Then how can I use the information about time/div in my time-axis?

Did I do something wrong? Please advise, thanks.