Synchronizing a mix of TSN and non-TSN devices

[JKSH]

Hello,

We're planning to deploy a system of 10 CompactRIOs on a single network.

If I've understood TSN correctly, then if I connect all the cRIOs to a single TSN-compliant switch, all of their system clocks would automatically sync to each other without further intervention from me. Is that correct?

Now, if one of the cRIOs has to connect to the network via a non-TSN switch, I understand that it can no longer get the tight nanosecond-level synchronization offered by IEEE 802.1AS. However, I still hope to keep the "isolated" cRIO synchronized to within ~100ms of the others -- is this possible? What would I need to do to make it work?

Finally, how do I ensure that the TSN-compliant cRIOs and a server PC have their times synchronized (again, to ~100ms)? The server PC is unlikely to have a TSN-compliant network card.

[JKSH]

Anybody...?

I even submitted a Service Request at https://sine.ni.com/srm/app/newrequest about a week ago and got no reply. Unfortunately, due to the issue at https://forums.ni.com/t5/Feedback-on-NI-Community/Error-when-accessing-service-request-page/m-p/4046... I can't see my list of submitted requests and I can't tell if NI has received my query or not.

[Jerry_X]

You might want to repost this in this area for more TSN experts to chime in: https://forums.ni.com/t5/LabVIEW-Time-Sensitive/gp-p/5237.

[mcduff]

Two years later ...

Were you able to solve this issue? I want to connect some TSN FieldDAQs and cDAQ chassis together through a TSN switch. If you solved it, a few questions

1. Switches have long lead times now, can I daisy chain together for testing until switch arrives?
2. System will use Windows as a controller, not a cRIO. Where will it get the TSN time from? I will have a TSN clock when the switch arrives, but for testing purposes would like to attach to Windows computer and start programming.
3. Do you know the difference between Host Time and I/O Device time, that is, how to convert between them?

Thanks

[JKSH]

---

@mcduff wrote:

Two years later ...

Were you able to solve this issue?

Not properly. And the TSN forum linked by @Jerry_X is now gone.

We went with the traditional route of getting a PTP (IEEE 1588-2002) time server, to be safe.

---

@mcduff wrote:

If you solved it, a few questions

1. Switches have long lead times now, can I daisy chain together for testing until switch arrives?
2. System will use Windows as a controller, not a cRIO. Where will it get the TSN time from? I will have a TSN clock when the switch arrives, but for testing purposes would like to attach to Windows computer and start programming.
3. Do you know the difference between Host Time and I/O Device time, that is, how to convert between them?

---

This might be a case of the blind leading the blind, but I'll try answering with what I know.

1. Yes, one of the attractive core ideas of TSN is that you can daisy-chain the chasses together and they will all auto-sync. The TSN-capable FieldDAQs, cDAQs, and cRIOs all have internal TSN-capable switches; you don't strictly need an external switch. (My original question was about mixing TSN-capable and non-capable devices, which complicates things)
2. All of the connected TSN-capable devices check each others' clocks and elect a master among themselves. If the master goes offline, they will choose a new master.
3. I don't have anything to add on top of https://www.ni.com/documentation/en/temperature-input-device-for-fielddaq/latest/manual/time-based-t...