Anyone else using these cards? Xagyl FLR9G30 and XC900M are 900MHz cards that are compatible with the XR9.

I've got a few FLR9G30 out there and they are more resilient to interference. I was wondering if anyone has compared the old FLR9G30 and the new XC900M?

i have several XC900M and they are GREAT! The noise floor with the XR9 bounced around from -50 to -80 something. With the XC900M the floor is steady in mid to high -90's. Beside replacing the AP's XR9 am using the XC900M in all new CPE and will replace some in existing CPE.

We have a lot of interference here with electric meters "calling home" every few minutes. The XC900M solved most of our 900 mhz problems.

Thanks.

Always looking at ways to increase capacity and quality. Glad to hear it other people are having good results as well.

It is hard to find info about these cards online.

What channel width, nstreme, nv2, 802.11?

I have more bandwidth when using nstreme but better stability and jitter with standard 802.11, using 5MHz wide channel. I was hoping nv2 would help a bit, but I can't update ROS on this AP right now.

i use 10mb width with the standard 4 chanels. The XC900M has a jumper that allows you to use 5 channels. I use Nstream with best fit and 512 as size.

On my AP's I use RB433AH Ros 4.17 and currently use RB411 for cpe. I am looking at the Rb411L to save money. All are Ros 4.17. Again I see horror stories about the 5.X and don't want to introduce problems. I am not sure about upgrading remotely. I tried once switching to NV2 and lost the tower AP. I had to physically drive to site and reprogram AP.

I am afraid to try NV2 I have see a lot of post about problems.

Thanks for the karma

I've got a 900MHz customer on a 411L running v5.12. The only problem I had was the channel width scripting from v3 and v4 doesn't work on v5. Ran into a few snags getting that going. Other than that, seems to hang with my 3.30 clients pretty well.

So far v5.9-5.12 seem to be working fairly well. The problem I've got is the first AP I ever built was on my tower. I've got two 5GHz cards, a 2.4GHz and 900MHz all on an RB600. If I change from 3.30 some of the weaker 2.4GHz clients have serious issues. If I had built it all on separate boards, as I intend to from now on, it would be a lot more flexible.

3.30 was a good ROS. I upgraded to 4.17 and everything worked the same or better. Still not sure on AP using 5.x unless it is a new tower not a upgrade. If 5.x will work with all existing customers with out having to change them i may upgrade. Does the any setting really work? will the ap see 802 and nstream and NV2 at the same time?

I don't know how far you can go with the "any" setting. I can tell you this, I have a script that changes the channel width, nstreme on and off until it connects. With the "any" setting, enable or disable nstreme or nv2 and the CPE just connects like nothing has changed. No delay worth mentioning.

I upgraded receintly from 4.11 (nstreme) to 5.9 (nv2) (all XR9):

nv2 seems to work better with all CPE having the same data rate
frame control disconnect problems (supposedly fixed in 5.14)
AP reboots every day (coredump)
nv2 does worse than nstreme with links with CCQ below 75
nv2 does worse than nstreme at low power levels

Unfortunatelly the last points are sort of catch 22 for nv2 - if I lower the power to increase CCQ (self interference) nv2 jitter increases, stability decreases. Self interference is an issue for me especially evident when data flows both ways.

Instead of lowering the power I had to switch from 20 to 10MHz channel to increase CCQ to make it reliable (minus the daily AP reboots and hourly control frame disconnects on one radio)

Thanks for the tip on XC900M - I'll try to use it on my APs even though I already have notch and passband filters installed.

Just wanted to add to what I wrote above. NV2 did not work out for me and I reverted back to nstreme (on v5.9 this time).
During my testing nv2 throughput was about 25% less and very uneven, oscillating widely especially on nodes with <75CCQ. Latency was also all over the place, going as high as 1700ms(!?) and never as low as nstreme. One node was disconnecting every 10 minutes on 5.14 and every hour on 5.9.

wow.

Do you have polling enabled on nstreme?

wow.

Do you have polling enabled on nstreme?

My settings:

Enable Polling
Disable CSMA
best fit 3200
long preamble (it increases CCQ)

Note I originally stated that nv2 worked better - I based it on one way UDP tests. Once I tested using TCP, receive and transmit from three CPEs at the same time the throughput was much smoother and faster on nstreme.

I was wondering if anyone has compared the old FLR9G30 and the new XC900M?

I'd be curious to know the same thing. I tried out an FLR9G30 back when it first came out. It was better in some respects and worse in others. I gave up on it and stuck with the XR9's that I knew and trusted. I hear that the XC900M cards are a major improvement, but I'd be interested to hear some real world tests between it and the XR9, and between it and the previous FLR version.

I have one of the XC9 cards on order and UPS says I should have it tomorrow. I'm going to put it on a tower that has been having a lot of noise issues and see how it does. I'll report back when I get a chance to try it out.

Cool. Definitely interested in your results.

I've had better results with the FLR than the XR9, but not complete corrections of the problems.

The AP is much more reliable. I wasn't really connecting anyone with the XR9 AP as the noise floor would completely drown out the CPE transmitting back to the tower. I don't have that problem anymore.

There are still a few customers that have issues, one with the Xagyl card on the CPE as well, but it could still just be fresnel problems or any of a dozen other issues. I don't have to adjust for noise as often, but it isn't a magic bullet for certain.

I spoke with one of their techs. They are very proud of the improvements and admit it is an evolutionary change rather than revolutionary. They've got some more improvements on the way. He didn't get into too much detail, just enough that I know that they are not finished yet, sounds like they are on the right track. The new cards should be compatible with the old.

So far my results with the card have been somewhat mixed. I've also been chasing an interference issue on this tower, which hasn't left me with much time to do a before and after comparison. I do know the noise floor is being reported much better with the new card. I'll do some more testing one of these days and try to get a better idea of how it works.

Just a follow up on this really old post...

Still running FLR9G30 in my AP, XC900M and a few XR9 as clients.

Running 5.26, Nstreme runs best on 10MHz channel, I've seen as high as 8mbps. Nv2 is much more stable, *everyone* stays online, but I've only seen as high as 5mbps.

I'm still looking for info on the RBMetal9HPn and how it compares to the XC900M.

A later update...

The FLR9G30 is still in the air. Most of my customers have the same card now.

NV2 stability wasn't as it appeared. The radios would remain associated, however the customer would lose Internet access.

5.26 was better, but NV2 was still unstable on this particular AP. 6.20 works well, though I've still had little luck with NV2 on 900MHz. It looks great in testing, but I always get calls after I enable NV2.