Brocade ICX Series (cheap & powerful 10gbE/40gbE switching)

[custom90gt]

It isn't the power supplies. The ICX7650 has different fans and it runs the PSU harder hence sounding harsher and louder than the ICX6610

I measured SPL for both the switches inside my enclosed 42U rack, doors closed and standing around 5 feet away from the rack

Using Decibel X on the iPhone. This is reference + background

View attachment 28902

Switching the ICX 6610 on

View attachment 28903

After a minute or two the ICX6610 ramps the fans down

View attachment 28904

Now switching the ICX7650 on - note the different harmonics

View attachment 28905

The 7650 takes somewhat longer to ramp the fans down than the 6610 - and also has different and more unpleasant sounding harmonics

View attachment 28906 and View attachment 28907

My rack is in the basement but at night when the house is quiet, you can hear the 7650 - but the 6610 I think has a lower frequency and blends into the background noise very nicely

Since those frequency peaks are annoying - I've ordered some MLV for the rack - will try to put acoustic foam and MLV so that I can dampen those frequencies.

The 7650 pulls a quite a bit more power - it consumes 120W min at idle, the 6610 does 80 ~ 85W at idle

View attachment 28909 View attachment 28908

Appreciate the very thorough posting. I may have to pass on the 7650 since my daughter's room is near the utility room. It's a shame because it neatly meets all of the things I would want in a switch.

 

[hmw]

Appreciate the very thorough posting. I may have to pass on the 7650 since my daughter's room is near the utility room. It's a shame because it neatly meets all of the things I would want in a switch.

I’ll get my soundproofing materials over the weekend and will update the post or create a new thread to discuss

 

[itronin]

sorry work has been really busy - hoping to shut it down, pull the psus on Friday, and maybe get some comparable sound readings of the 7650 and 6610. re-iterating. audibly I cannot tell the difference between the 7650 and 6610 - but everyone's ear's are different.

and yeah, IIRC they are the same psu's. I have two B's in my 6610 and 2 fans. my7650 has two psu's and 2 fans + the front 4x10gbe module installed but I'm not using.

my 7650 has two A's... I have old man ears I guess! @hmw just posted a really fantastic comparison in this thread so I don't think I need to try and replicate collecting audio samples.

 

[casperghst42]

Hi,

I got an 7150 C12P running SPS08095g, and am hitting my head against the wall. It is a very nice switch, which is mounted in a wall rack enclosure with adequate ventilation (2x12cm fans).

I am trying to run 2x 10GBaseT transceivers, one cable is ~10-12meters(cat7) and one is 2meters (cat6a), my problem is that the transivers get so hot that I loose connectivity, running it with one transiver works. One DAC Cable (2m) and one 10GBaseT transitiver works most of the time, but still with drop outs.

Have anyone found a way to cool the two SPF+ ports in that switch, enough to make 10GBaseT transceivers work without problems?

 

[neb50]

Hi,

I got an 7150 C12P running SPS08095g, and am hitting my head against the wall. It is a very nice switch, which is mounted in a wall rack enclosure with adequate ventilation (2x12cm fans).

I am trying to run 2x 10GBaseT transceivers, one cable is ~10-12meters(cat7) and one is 2meters (cat6a), my problem is that the transivers get so hot that I loose connectivity, running it with one transiver works. One DAC Cable (2m) and one 10GBaseT transitiver works most of the time, but still with drop outs.

Have anyone found a way to cool the two SPF+ ports in that switch, enough to make 10GBaseT transceivers work without problems?

Which 10GbaseT transceivers are you using and what speeds do they need to support?

https://www.amazon.com/dp/B078SNK1MY?psc=1&ref=ppx_yo2ov_dt_b_product_details - This one properly supports 1, 2.5, 5, 10G in a Mikrotik but gets hot.

https://www.amazon.com/dp/B09ZYGYZ7X?ref=ppx_yo2ov_dt_b_product_details&th=1 - This one supports 10G and may support the rest, but I haven't tested yet. This one runs much cooler and currently am only needed 10G connection speed.

I am only running a meter or two cable with both, so the heat is not due to long cable runs.
I also have not tried either in my ICX7250 switches so can't comment on how they work in those, but would assume 10G would work with both and 2.5G would be questionable with either.

 

[casperghst42]

Which 10GbaseT transceivers are you using and what speeds do they need to support?

I also have not tried either in my ICX7250 switches so can't comment on how they work in those, but would assume 10G would work with both and 2.5G would be questionable with either.

I got two H! Fiber 10GBaseT (cheap from Amazon) which I thought was running too hot - especially the one which runs to my desktop would reset to 100mb/s or 1GbE, the one which runs to my server (2 meter cable) would be ok.

I then got a FS transiver for 80m which though would be better suited for the run to my dekstop ... but that is even more unstable than the one from H! Fiber. I should probably have gone for their new 100m certified ones which only uses 1.2w.

I now run 1GbE (normal port) to the desktop and 10GbE to my server which is stable, I would like to have more speed, but can live with it.

I just wanted to know if I am the only one facing problems with 2 10GBaseT transivers at the same time, as the switch is passive cooled.

 

[kpfleming]

I suspect the designers of the 7150-C12P expected the SFP+ slots to be used primarily for stacking, and using DACs, AOCs, or fiber transceivers. The unit is a little warm on its own, without any modules in the slots, so I'm not at all surprised that putting 10GbaseT modules in those slots causes problems.

If the server is only 2 meters away from the switch, an AOC would be a good choice, assuming the server has an SFP+ slot available.

 

[casperghst42]

I suspect the designers of the 7150-C12P expected the SFP+ slots to be used primarily for stacking, and using DACs, AOCs, or fiber transceivers. The unit is a little warm on its own, without any modules in the slots, so I'm not at all surprised that putting 10GbaseT modules in those slots causes problems.

If the server is only 2 meters away from the switch, an AOC would be a good choice, assuming the server has an SFP+ slot available.

At this point, so do I. I used to use a DAC cable, but I wanted to use 10GBaseT after I got a X10SDV-6C+-TLN4F - turns out that this was a bad idea. I'll go back to use a x520-t2 and a DAC cable.

 

[gb00s]

I am confused. I just bought some 6610's 24 & 48's non POE switches. I cared a lot to get switches with power supplies of B or C revisions. In my opinion they are not a lot quieter than the A revision and I could have saved a lot of time of searching. These C revisions psu's where still horrably loud.



However, I was working at one 24 port switch and accidentally pulled a power cable. I couldn't hear the switch anymore in the environment and thought, f.... all my config (not saved yet) is gone. But the switch kept working with the other power supply I couldn't hear. So I checked the psu and it's a different one but also 'Rev C'. So this switch had 2 different Rev C versions with a wide gap of noise.



The second psu is the super quiet one. I tested both psu's independently in regards to the temps and the super quiet version keeps the temps almost at the same level compared to the loud Rev C version. I checked @fohdeesha 's power supply section of his guide I remembered and found this photo which looks quite similar to the first image of the loud Rev C version.



Does somebody have any knowledge of the super quiet psu? For me the super quiet psu seems to be a Delta power supply while the loud Rev C version is an Emerson psu. He a comparison from the outside.



Any hint is much appreciated. Thank you in advance.

 

[custom90gt]

I am confused. I just bought some 6610's 24 & 48's non POE switches. I cared a lot to get switches with power supplies of B or C revisions. In my opinion they are not a lot quieter than the A revision and I could have saved a lot of time of searching. These C revisions psu's where still horrably loud.

View attachment 28920

However, I was working at one 24 port switch and accidentally pulled a power cable. I couldn't hear the switch anymore in the environment and thought, f.... all my config (not saved yet) is gone. But the switch kept working with the other power supply I couldn't hear. So I checked the psu and it's a different one but also 'Rev C'. So this switch had 2 different Rev C versions with a wide gap of noise.

View attachment 28921

The second psu is the super quiet one. I tested both psu's independently in regards to the temps and the super quiet version keeps the temps almost at the same level compared to the loud Rev C version. I checked @fohdeesha 's power supply section of his guide I remembered and found this photo which looks quite similar to the first image of the loud Rev C version.

View attachment 28922

Does somebody have any knowledge of the super quiet psu? For me the super quiet psu seems to be a Delta power supply while the loud Rev C version is an Emerson psu. He a comparison from the outside.

View attachment 28924

Any hint is much appreciated. Thank you in advance.

Super interesting. I have a revision A power supply that is much more loud/annoying than my revision C Delta power supply. I do not have any revision C Emersons to test though.

 

[gb00s]

Ok, but @fohdeesha his photo looks like the Emerson's sticker. Delta stickers, at least mine here, looks much different.

Uploaded a short vid here >> VID_20230512_235938.mp4 with the fan noise. As all my other psu in the other switches sound the same, I guess the psu/fan is not broken.

Edit: sound is the same even with the switch fully booted.

 

[hmw]

Ok, but @fohdeesha his photo looks like the Emerson's sticker. Delta stickers, at least mine here, looks much different.

Uploaded a short vid here >> VID_20230512_235938.mp4 with the fan noise. As all my other psu in the other switches sound the same, I guess the psu/fan is not broken.

Edit: sound is the same even with the switch fully booted.

Can you check the temps & the fan speed when corresponding the PSU is inside - are they different ?

 

[Nehalem501]

Hello,

I've recently acquired an ICX 6450-48 (non PoE version) and its (single) stock fan is very loud.
When reading this thread, I understand the fan should slow down after a while after the unit has finished booting, but mine seems stuck at full speed all the time without anything connected to the switch (I kept it on for over an hour once to check if the fan would slow down after some time).

I've already followed the guide to update the firmware to the latest version (08.0.30uT313).

The temperature displayed when running show chassis is always between 25° C and 40° C and for the fan it says Fan ok, speed (auto): [[1]]<->2

I don't know if it is supposed to work like that, but the fan blows air from inside the switch to outside the switch.

Is there something wrong with my unit, or should I just change the fan? (which one would be the quietest? I would prefer to keep some form of cooling rather than running it completely fanless, summers in France can get pretty hot and most homes don't have AC here)

Thanks for your help.

 

[NablaSquaredG]

Is there something wrong with my unit, or should I just change the fan?

Yes, Fan MOSFET fried. Seems to be a common issues with those switches.

 

[Nehalem501]

Yes, Fan MOSFET fried. Seems to be a common issues with those switches.

I'll try to replace it. Thanks.

 

[hmw]

Super interesting. I have a revision A power supply that is much more loud/annoying than my revision C Delta power supply. I do not have any revision C Emersons to test though.

I have a Rev B PSU that has newer firmware and is newer than the Rev C PSU. And I have Rev A PSUs that are the newer than B or C. No idea what Brocade does. OTOH, tested both B and C PSUs and it makes no difference to the noise

 

[hmw]

Appreciate the very thorough posting. I may have to pass on the 7650 since my daughter's room is near the utility room. It's a shame because it neatly meets all of the things I would want in a switch.

Decided to do some minimal sound proofing for the server rack. Started by adding some 1lb Mass Loaded Vinyl to the removable side and back panels



And followed it up with some cheap 1.5" acoustic foam



The front of the server rack is glass and has perforations for airflow - this wasn't covered completely, just enough so that air now flows through the lower part of the front door but the sound from the switch is blocked by a small MLV strip



All in all, took a few hours to cut and stick everything in place. The result is an appreciable reduction in noise - something like 5dB ~ 6dB. However - as noted previously, the switch has a really unpleasant acoustic profile and even when muted can be annoying. As you can see below - the noise reduction is great compared to my previous readings HERE



The problem is the harmonics at 1.6KHz - this is the unpleasant whine that I noted earlier. While muted it is still annoying. Luckily the rack is in the basement and I cannot hear it anywhere upstairs. If you have a utility closet, that might be enough to block the noise completely

 

[happydaisy]

I am trying to get PIM working for SSDP from VLAN50 to VLAN 1, without much success so far.
So I thought I would go back to basics and try to get SSDP working within VLAN50 with IGMP-snooping enabled, still not having much luck, if everything is patched through the ICX7250, multicast traffic just seems to get dropped.


The network consists of the main router (Mikrotik RB4011), the main switch (ICX7250-24p), a mini mikrotik serving out DHCP on VLAN50 only (RB750GL), it isn't doing any routing, just providing NTP and DHCP services to VLAN50 and finally 2 Ruckus R610 access points, one connected to the switch, one to the RB4011 router

With IGMP-Snooping disabled on all three devices (Mikrotik RB750GL, RB4011 and ICX7250-24p), everything works as expected. SSDP multicast queries flow through from the server to the client device on the wifi access point, no matter what the access point is connected to or how the devices are wired up.

With IGMP snooping enabled on the RB750GL and RB4011, SSDP queries can still be seen on the client laptop connected to the R610 via the main router, however, only if the ICX7250-24p is bypassed.

I am pretty sure I am missing something fundamental on the ICX7250.
So far I have tried the following commands:

enable
configure terminal
ip multicast

<test>

ip multicast active

<test>

ip multicast flood-unregistered

followed by

vlan 50
multicast

<test>

multicast active

<test>


Firmware on the ICX7250 is 09.0.10e
Mikrotik's both on latest stable 7.9 releases.

I will also say I have not dealt much with multicast/broadcast traffic before, still learning the ins and outs.

 

[akw2312]

Is it possible to force the fan speed on ICX7250?

I have changed the FAN on ICX7250, but the switch can not receive enough airflow on mode 1.

 

[itronin]

"Brocade E1MG-SX-OM 33210-100 Compare Brocade REAL & FAKE 1GbE SFP-SX"

This bay listing amused me!