NI-DAQmx write raw takes up to 5ms (too slow)

[GoPSU]

I never solved it, but I changed the architecture of my system to make everything continuous and based on hardware clocks generated by a timing card that drives the chassis backplane clock using slot 2. For example, I have a 10ms half buffer that is hardware synchronized so that the output is always perfect to within one sample. The only issue is that the system cannot ever respond faster than 10ms. It always responds within 20ms (two half buffers). It doesnt need to respond that fast anyhow for my needs. If you need faster response, use smaller half buffers.

[arse]

It sounds like you are trying to do something similar to me.  I have a simulation cycle that ticks over a 1ms and it works but there are occasional delays that I can't get rid off.

Have you had any experience with Channel.MemoryMappingEnable?

Thanks.

[GoPSU]

No I have no experience with that.

If you explain your system in more detail I may be able to help come up with a solution. 

What is your desired delay between input and output?

Do you need to do more than to set/update a control voltage?

What I was doing was mutichannel sonar array stimulation so I needed a fixed high sample rate high fidelity set of signals. The minimum system delay determined how close I could get to the target. We still use that system to this day and there are four copies of it, so it worked well and is stable.