WS-12-250-DC Temperature and Fan Speed

[bgoode]

Summer is coming to the Desert Southwest...temps are already close to 100F. Today a separate temp sensor in the box housing the WS-12-250-DC is showing ambient 37.3c/99F. The enclosure is actively ventilated (not cooled) with a continuously running redundant (2) exhaust fan that draws ambient temp air from the bottom of the box and exhausts out the top side of the box. Summer time ambient temps can reach 46C/115F on the desert floor...this box is at a higher elevation and will likely see 40C/105F.

So...what can you tell us about how the fan is controlled? What switch monitored temps should be we watching? What is the MTBF for the fan? I assume that the higher the RPM the shorter the life of the fan. I also assume, under the above conditions, if the fan stops...the switch does too. I've always been concerned with a fan design for our remote locations...this switch is the first for us. What has installed experience shown?

This is critical backhaul site and a PIA to get to...we have a spare switch on the shelf but a temp caused outage would result in at least a multiple hour outage.

The switch is at v1.3.8, Board Rev. D, Power Supply Firmware v28...Today's 99F ambient temp shows RPM consistent between 9450 - 9990:

[sirhc]

Well first off we have sent out alert to get people with SMART DC switches to upgrade to v1.4.0rc12 as there were several bugs that were important (review release notes)
I STRONGLY ADVICE YOU UPGRADE.

Since upgrading from such an older firmware (for the SMART DC anyway) I would be on site and prepared for whatever. Biggest concern is when upgrading from v1.3.8 with SMART DC model it "could" require a full power cycle.

The SMART DC switch will shut itself down if it gets too hot. First it will shed POE ports in an attempt to cool down but ultimately it will shut down if too hot.

Fans are rated at something like 50K hours MTBF but I am sure some fail a lot sooner?

If this is such a critical location I would look at ways to keep it as cool as possible, maybe a shade visor or plant a tree.

The fan is controlled mostly by the switch board temp, Kicks on at 40C, at full by 50C, kicks off at 35C

After you upgrade to v1.4.0rc12 please post up the FULL device Status Tab again.

[bgoode]

The box is fully shaded and insulated...without the interior temp of the box would be much higher....maybe +20F

[bgoode]

"The SMART DC switch will shut itself down if it gets too hot. First it will shed POE ports in an attempt to cool down but ultimately it will shut down if too hot."

How hot is "too hot" and which temp sensor(s) would determine that? What temp should we be watching?

In what order will the POE ports shed (the same as "Power" Priority)? Will the POE ports shut down come back up after sufficient cooling?

[sirhc]

In what order will the POE ports shed (the same as "Power" Priority)? Will the POE ports shut down come back up after sufficient cooling?

It uses the priorities specified on the Power Tab. Remember the higher priority ports get shut off last, lower priority get shut off first

[Dave]

- I tested this design at an ambient of 55 degrees Celsius with a 200 plus watt POE load and vin=24 volts...runs no problem & well within design specs.
- if it does get to hot (uses power supply heat sink & power controller temperatures) it will shed ports.
- it will then turn ports back on after cools to within acceptable operating temperature for ten seconds.

[bgoode]

"- I tested this design at an ambient of 55 degrees Celsius with a 200 plus watt POE load and vin=24 volts...runs no problem & well within design specs."
Was that with the fan operating properly (and at what RPM)?

Still would like to know...

How hot is "too hot" and which temp sensor(s) would determine that? What would the "too hot" readings be?

[sirhc]

PHY and CPU are rated up to 125C

Most other parts on switch board are rated to 75C or higher

Board temps of 65C is OK

I think the power supply controller is rated above 85C?

If you're asking for a precise temperature things will fail then we can not provide that as most electronic components come with a range not a precise definitive temperature the component will fail.

They do not say "exactly and precisely at this specific temperature this will fail"

[Dave]

- we will start shutting down ports when power controller gets above 110 degrees Celsius for 20 seconds, or 105 deg for over 30 seconds.

[bgoode]

Thanks Dave...that's the info I was looking for.