However, when it ONLY runs a SR9... runs solid as a rock.

Here's the setup:
- site 1 has internet connection coming in through ether1 on rb153, ROS 2.9.43
- site 1 has SR9 900mhz connected (atheros 5213) to 18db yagi, mounted horizontally 180ft up, brand new connectors & pigtails, freq: 2432, set @ 2ghz-5mhz, default power, txmode_manual
- site 1 has RB52 2.4ghz connected to an omni-type antenna, freq: 2462, set @ 802.11-B, (atheros 5214)

- site 2 is a remote location, w/backbone p2p link to site 1 via SR9 cards, ROS 2.9.48
- site 2 has SR9 900mhz connected (atheros 5213) to 18db yagi, mounted horizontally 150ft up, brand new connectors & pigtails, freq: 2432, set @ 2ghz-5mhz, default power, txmode_manual
- site 2 has TWO RB52 2.4ghz, one connected to a grid-style antenna (backhaul to another site), second one to a omni to service local clients; both these guys have diff. frequencies & are set in 802.11-B/G (atheros 5214)

both sites are line of sight, minimal obstructions. Normal signal between the SR9's is -78db/+20snr - gets the job done.
both sites are using 48v POE on ether1
both sites running data connections over ether2 and ether3 locally


---
Here's the problem:
when site 1 has the RB52 disabled and SR9 + ethers enabled it runs SOLID
when site 1 is running SR9 + RB52 + ethers it will lock itself up and requires a power cycle

when site 2 has the RB52's (or CM9s) disabled, SR9 + ethers enabled it runs great, SOLID
when site 2 has the RB52's (or CM9s) enabled + SR9 + ethers, it will lock itself up and requires a power cycle

these sites were previously working without issues, one day they just up and started acting like this. The only common theme here is that the BOTH seem to crap out on us when we're running bothe a SR9 + a CM9 or RB52.

We're seeing this @ 2 of our other sites right now as well, sometimes the backbone link will last a few minutes, sometimes a few hours, sometimes a few seconds.... if we ONLY run the SR9 cards though it will run all day long.


HELP this has kept us in the dark ALL WEEK (and messed up my thanksgiving), we're losing customers over it!!!
Any help is VERY MUCH appreciated! thank you!

-D

fcwireless -
Well first off you were fortunate that you were able to run the SR9 AND another 2.4Ghz card in the same 'box'. We tried that when the first SR9s came out - that's when we found out that the SR9 used a frequecny converter to downconvert 2.4Ghz to 900Mhz. No good for us - too much 2.4Ghz and other 800-900Mhz systems out there. We moved them to seperate enclosures, I built my own 900Mhz bandpass filters. Between the two we were able to use 900Mhz very successfully in mico-pop configurations.

Getting to your problem - I would imagine that I am repeating the obvious here but....
1) Have you changed channels on the 900Mhz - 2432Mhz converts to 912Mhz and has a 'harmonic' at 2457 when using the SR9.
2) I don't believe that your ethernet connections are locking up - more likely the entire RB is tied up trying to make connections. The SSIDs are different - right - between the SR9s and the 2.4Ghz radios?
3) I see you have ROS 2.9.48, did you also update the underlying RB firmware? /system routerboard print - see if the current firmware and upgrade firmware are the same version - if not then do /system routerboard upgrade answer yes. It takes about 5 seconds, no reboot is necessary.
4) There have been significant weather changes in the past few weeks...fall weather, make sure you have solid DC ground connections, this seems to help with noise issues - whether seen or unseen, as well as this also seems to help on 'out of channel rejection' properties.

Thom

we've updated most of our firmwares and so far, the site 1 board is running stable with it's RB52 and SR9 cards BOTH enabled, so far so good... we've been a few hours now with no issues.

However, this is odd... we have a few boards (RB153) which are saying they are at RB firmware 2.7 but it says that the upgrade version is 2.6. What gives?

fcwireless -
Well - have your systems stayed stable?

As to your firmware question - this does happen from time to time that MT produces a board with newer firmware than is currently being distributed with the software. In a version or two the new firmware should catch up...

Thom