Hello LabJack support,
The Frequency In (Extended Feature Index 3, Rising Edges) seems to miss the very first positive edge anytime after the DIO#_EF_ENABLE is switched from 0 to 1, however, this is not the case for Falling Edges (Extended Feature Index 4), the falling edges works as I believe it was intended.
I have only tested this using DIO0. Kipling and a Custom C# code I wrote reproduces the same error.
EDIT: This is consistent behaviour for both One-shot and Continous mode.
.png%3Fitok=PHGBxf6M "EEV Blog Reviews LabJack T7-Pro & Software Tools")
.jpg%3Fitok=B1YcuDFK)
.png%3Fitok=Ybt-OlgU "LabJack + JEDI One Can Create a Powerful Distributed DAQ Ecosystem")
I see the same thing. A firmware investigation has been started.
I have been testing various conditions and hardware settings. So far, I have found...
The first rising edge is captured properly only when the line is high when the DIO_EF is configured and the signal idles high. I have tried several hardware configuration options with no good results. At the moment this seems like a hardware nuance.
While we look at this, I recommend trying L2L.
I switched to use Negative Edges for now so this is not causing any major problem for me at the moment. but please keep me updated if this ever gets fixed.
One interesting experiment is to check if you miss a negative edge while the Input idles high first. (exactly the opposite)
We have been unable to get the hardware to capture that first edge. A note has been added to the Frequency In section of the datasheet.
Hello Labjack team,
I am a little bit disappointed by the response I have received about this issue, would you care to share some more details on why you think this is a hardware issue.
This is a core functionality that I need to make sure it is working properly,this will be implemented in 7 different stations and to be honest the "Hardware" answer is not very convincing. would you care to elaborate? do you think the pull up resistors are causing an issue? do you have an internal chip that ignores the first pulse? Have you tested this on the negative edge configuration? does it reproduce the same results on negative edges?
How L2L would measure time between pulses from the same source, this is not very trivial by the way, I would need to connect the same signal to 2 different channels, and then have extra hardware to invert it on one of them.
Is there any other way to measure the time intervals in between consecutive pulses?
Regards,
Walid
The L2L allows you to specify the edge of interest for each line. An inverter should not be necessary.
Several combinations were tested. During testing, I noticed that the idle state of the line also affected the detection of the first pulse. Idle state being the state that the line is in most of the time; idle state of 1 indicates a duty cycle greater than 50%. All combinations of initial state, idle state and edge direction were tested. Below are the results. Correct results are bold.
I have not been able to produce a reasonable explanation for why some combinations work where others do not. In the bad cases, the first edge does not trigger a capture by the hardware.
Tests were conducted using a lua script.
+----------+---------------------+-------------------+--------+
| DIO_EF I | State before Config | Signal Idle state | Result |
+----------+---------------------+-------------------+--------+
| 3 | 1 | 0 | 3.11 |
| 3 | 1 | 1 | 0.03 |
| 3 | 0 | 0 | 2.63 |
| 3 | 0 | 1 | 2.74 |
| 4 | 1 | 0 | 0.03 |
| 4 | 1 | 1 | 0.02 |
| 4 | 0 | 0 | 0.03 |
| 4 | 0 | 1 | 2.84 |
+----------+---------------------+-------------------+--------+
Many configurations of the hardware were tested. None of which produced any improvements.
Thank you for your detailed response and for the test results, I am more confident now to use the negative edge which seems to suit the application.