For your application, the Nyquist theorem doesn't apply. In most cases where you are generating and acquiring response digital data, you will almost always want to use the same clock for both generation and acquisition.
I would recommend reading up on Round Trip Delay elimination in the HSDIO help to get a good understanding of the system you are trying to build. This doc can be found in the NI Digital Waveform Generator/Analyzer help under Programming->Features->Eliminating Round Trip Delay. This will talk about a source synchronous method of generating and acquiring data. This would requir you routing the clock through the ASCI and back into the NI-6541, as well as a synchronous trigger.
If you cannot change up your ASIC to do the above, I would then recommend using a pattern match start trigger for acquiring your response data. Here is the programming flow: 1) Configure the generation session using an onboard clock 2) Configure the acquisitin session using an onboard clock and a pattern match trigger that matches when IRQ_Real is HIGH 3) Start the Acquisition session 4) Start the Generation session 5) Wait until Done 6) Read acquistion data from device 7) close both sessions.
Looking at your HWS waveform, I know that the IRQ_Real signal will be low until the ASIC starts generating data to it, which is why I chose to use the pattern match on that signal when it is HIGH. Now to do the software compare, you will need to get the IRQ_Real data out of your HWS file, and remove all of the leading zeros such that the first value is a HIGH. Now your expected response data will be aligned to the pattern matched data. Lastly send both the acquired IRQ_Real data and the modified HWS IRQ_Real data to the software compare VI discussed in the previous post.
One last thing that may or may not be an issue (depending on how fast your clock is) is data delay. If the acquired data is totally incorrect, there is a chance it is because the IRQ_real data coming from the ASIC is transitioning at the same time as the onboard clock, thus the 6541 is sampling when the IRQ data is in an invalid state. To get around this, you can use the Data Delay feature to delay when the acquired data is sampled within a clock period. You will have to just coerce this value manually until the data starts looking correct. I would start with 50% delay, then 25% and 75%, etc, until the data looks good. There is an example for using data delay, In the examle finder, go to Modular Instruments->NI-HSDIO->Dynamic Acquisition->Dynamic Acquisition With Data Delay.vi
good luck,
Jared
Message Edited by JaredW on
01-25-2008 03:57 PM
