Determine achievable speeds on AMDS

[mohamed-dek1]

Context

Related Issues / PRs: (To do: list all related issues)

• Design and Implement a Daisy Chain Protocol to be used with the AMDS and FBC #80
• Design a spec for the Daisy Chain Protocol #91
• Develop and test the Daisy Chain Adapter Board #81
• Implement daisy-chain firmware and demonstrate first data exchange on AMDS #92

We need to be able to daisy chain 3 AMDS (or FBC REVB) boards and have all the data arrive and be parsed by the AMDC at a reasonable data rate. The spec is being made in #91, it will be tested in #92, and now we need to test achievable speeds to determine if this spec will work. This issue will act as a go/no go point for this spec.

Approach

Close this issue by:

• Testing the data rate we can get with the current Daisy Chain spec
• Determine whether we need to revisit the spec or continue with testing on the FBC REVB

[elsevers]

@mohamed-dek1: This timing diagram suggests that three fully populated AMDS boards daisychained together would require approx 23 us to transmit data (1us + 4us + 3 x 6us), which would limit control rates to below 43 kHz.

Can you please confirm this calculation? If it is correct, I think that is probably okay, but likely emphasizes the need to be able to detect that not all cards are populated expedite transmission.

[mohamed-dek1]

Per 3/11:

Thank you @elsevers, based on the propagation delay listed in this comment (640 ns), each AMDS will receive the sync signal in less than 1us and begin their data acquisition in parallel. Depending on if the time it takes for data to propagate between the AMDS boards is shorter than from AMDS to AMDS, we would be limited to about 43 kHz. However, if it takes longer for the data to get between AMDS boards (unlikely), it might be bigger.

[mohamed-dek1]

Per 3/24:

With the most recent successful data exchange across 3 AMDS boards taking 66.272us, we are currently limited to about 15kHz. There is an issue working on improving the current implementation so that we can meet our timing requirements.

• Demonstrate 3x AMDS’s in series using the DCP meeting timing requirements #95 (comment)

[elsevers]

Timing benchmarks are documented in a PR comment thread starting here:

• Add AMDS firmware changes to implement the DCP #102 (comment)