– Copied from a customer query –
We want to design a carrier board on which two SPEBlox-Long modules can be mounted inside a case. The idea is that we can have a redundant 10Base-T1L connection from within a single housing.
Are there any suggestions you can give us, is it for example handy to include a PPTC as an input protection for the SPEBlox-Long boards?
You mention you are trying to integrated two SPEBlox Long boards into your case to achieve a redundant 10BASE-T1L connection.
When you say redundant, do you mean that both boards connect to the same ethernet networks? Or do you mean both SPEBlox are on independent networks?
Sorry for the delayed response. We want to have two parallel SPEBLOX-Long boards in parallel which are also connected on the other side of the 10BASE-T1L with two parallel SPEBlox-Longs.
The 10Base-T connections from the two SPEBlox-Longs are connected to a managed switch configured with LACP. See below for a overview.
Understood thanks for clarifying. So both 10BASE-T1L connections are distinct, and you will use the switch to manage the redundancy. One thing to bear in mind, if you have two parallel links between two switches on a network, you may have an issue with network loops. That being said, with LACP you can solve that network issue, or use RSTP on the managed switch (if available).
Regarding suggestions for integration, there are no specific considerations you need to be aware of for what you want to achieve. Using a PPTC is not mandatory, and it is unlikely that SPEBlox Long will fail in a way to cause a short circuit on the power lines. That being said, it wouldn’t hurt to add one.
I’m not sure this will work as a redundant connection. LACP has a static assignment of traffic to one of the two links (usually based on source and destination port and IP). If one of the links goes down, it will send all traffic via the remaining link. So far, so good. However, the requirement is that a non-functional link goes down. I don’t know the internals of speblox, but with a high probability, the Ethernet part will remain connected regardless of the connection ability through the twisted pair. But maybe I’m wrong.
Hmm. If LACP uses LACPDUs to determine whether a link has gone down, then won’t a non-functional SPE connection will still show up to the ethernet port on the other side as a lost link.
That being the case, the managed switch would still be able to detect when a link has gone down on the SPE side?
Yes. With LACPDU it should figure out automatically that the link is down. So it is definitely needed to test the behavior and reaction time of the particular setting of LACP.
Yes, a bit of an unknown. Will report back findings from the customer once we have it.
I’m the customer and the Etherwan SG300 series switches we currently use can detect that the link goes down. In the example below the 10Base-T1L connection of port 2 was unplugged. Reestablishing the 10Base-T1L brings the blocking state back to forwarding.
Nice, thanks for posting!
