Synchronous Capture - Analog Discovery 2 / Digital Discovery Trigger

[alexplcguy]

Hello, I have a process where we are using several of the AD2 devices to capture a few uS of information on a trigger (T1) event generated by a Digital Discovery board. The Digital Discovery is looking for an external Logic trigger input, and then executes a pattern which pulses external equipment as well as these AD2 devices. The rate at which the external Logic trigger is occurring is fast (2mS or so), however, we are using a custom Java interface to pull the latest buffer data from the AD2s and this occurs on the order of 100mS. The problem is that when the first Device's buffer is pulled it does not necessarily correspond to the same event captured in the next Device's buffer when it is pulled. It is very useful for us to make sure that the buffer we pull across all devices, is from the same trigger event. This will help us paint an accurate picture of what is going on. The equipment cannot run slower and we cannot speed the buffer pulls to under the 2mS, due to data processing constraints. Is there a way through configuring the AD2 (HoldOff / Buffer Stop, etc) or the DigitalDiscovery Pattern output that we may be able to achieve synchronous captures on the order of 100mS while the remainder of the triggers controlling the equipment continues to run at 2mS?

Thank you!

 

[attila]

Hi @alexplcguy

A. Use longer Run time in Pattern Generator which generates trigger for AD2s.
B. Use trigger source: "Logic" instead "Logic Detector". The first one only fires when a Logic Analyzer capture is performed and not on every trigger event. Set higher Update rate, like 200ms, under top-right COG 

The Scope has HoldOff option but this, depending on trigger sequence, may not make the captures of multiple devices synchronized.

[alexplcguy]

@attila

Thank you for the response. Unfortunately I am unable to test these options as the devices are in use. I have ordered more and I'm waiting for the delivery. Until they arrive, for clarification:

A: Can you change the Run Time for a Pattern Generator for a single signal? Currently the same pattern generator is used to run other equipment. Although everything else needs to run fast (2mS), the Trigger of the AD2 should be slower (100~200mS), but sync'd with the triggers of the Pattern Generator for other devices. 

B:I am just completely confused about what these things actually are (Logic Analyzer vs Trigger Event, Update Rate?). I see Frequency on the "Clocking" tab but not Update Rate. I will read up on these things and test when the units arrive. 

I also don't suspect HoldOff will work

[attila]

Hi @alexplcguy

A. There is one Pattern Generator in the device so the state, trigger-wait-run are the same for all signals.
With Digital Discovery (and larger devices ADP2230, ADP3X50) you can specify Repetition for each signal, so it generates it only the given number of times, but this is not available with AD1,2,3

B.

 

[alexplcguy]

Thank you, Yes, our system is set to "Logic Detector" so I suspect the "Update Rate" does not affect us, as this is related to "Logic". We do not have an issue with missing pulses or anything like that. I think the issue is more a matter of the Pull Rate of the computer. I understand that one option is to "Repeat" the faster signals after the Logic Trigger, however, for safety reasons, we need the incoming trigger to ONLY trigger equipment ONCE per incoming event. If incoming triggers stop, so to should the outputs (immediately). Another issue proposed here is that the Trigger Rate is variable up to 480Hz so, we cannot have any repetition because the rate would be off.

To demonstrate the issue, I've whipped up a sketch. Trigger input up up to 480Hz goes to equipment. Equipment generates signals. Signals get captured with AD2 devices. AD2 buffers are pulled by a computer at 5Hz (with custom Java Software). However, you see by the picture that (at High Trigger Rates) the computer will pull the buffer associated with Trigger Event #1 across some devices but for other devices it pulls the buffer for Trigger Event #2. It does not help us to paint a clear picture across all equipment which are working together as a system.

Please let me know if you would like any additional configuration information. I am starting to suspect I will need a hardware solution.

[attila]

Hi @alexplcguy

If the trigger rate is 480Hz and your capture rate is only 5Hz then you will have missed triggers and the captures won't be synchronized.
The trigger rate should be less then 5Hz and to synchronize the captures makes sure all capture devices are armed before the (first) trigger.

[alexplcguy]

ok, yes, unfortunately Lowering rate is not an option.. I was just curious if, on the Digital Discovery board, there is a way to have ONE signal operate on a different logic or implement a minimum hold-off on the pattern output for a single signal or something like that. Maybe this can be a future firmware enhancement. I will design a hardware solution. Thank you for your input. 

[attila]

At what sample rate are you capturing ? If this is ~MHz you could use record mode, continuous data streaming on ADs.
The AD3 and newer devices have reference clock option which makes the sampling interval stable over time. AD1,2 only works with internal oscillator so there will be shift over time between captures with multiple devices.

You could lower only the Trigger signal frequency sent to the ADs or connect the Trigger only to one 'master' AD and configure this to output 'Scope' trigger event on Trigger2 which you connect to the other ADs. This way the captures will be in sync.

The ADP2230, ADP3X50 provide device buffering (memory segmentation) which can perform captures with 1us latency.

[alexplcguy]

On 6/18/2024 at 12:15 PM, attila said:

.... or connect the Trigger only to one 'master' AD and configure this to output 'Scope' trigger event on Trigger2 which you connect to the other ADs. This way the captures will be in sync.

This is an interesting concept. Yes, I may try this first. I've also designed a board that will only pass the trigger after a designated configurable "Hold" time. I will try your method here which will be much easier and less time consuming and not require physical space as I'm already very constrained there too. 

As for continuous recording, we are only capturing a 20uS burst of information at variable rates up to 480Hz. So, recording doesn't tell us much and the software is not designed to interface in this manner. We use the tiny frame of information to monitor and control various equipment in real-time.