View Full Version : XR2 results compared to R52 and Tranzeo Atheros
wispwest
04-13-2009, 07:52 PM
Just received the XR2 today excited to see it out-perform my other cards. We have a tower using a Mikrotik Routerboard and a 2.4Ghz mini-pci card. Over the past 4 months I've been testing different scenerios to see which card would perform the best using all the same conditions (channel #, etc). I started with a MIkrotik R52 card (80mw) into a 500mw Amp. My signal at home was -53db with my bullet!! The downfall, was my data rates fell to 1-2Mbps. So I figured out the AMP injected noise, although a better signal, the speed rates sucked. Without the amp, my signal went to -73db and my speeds were much better (18-24Mbps). 2nd test, I used a wireless card ripped out of an old Tranzeo AP, I believe its 200mw, my signal at home went to around -64db, and my speeds were a steady 36mbps. The best I have seen so far. Ping times were around 4-6ms to our DNS servers.
XR2, which is "high powered" "carrier-class" etc, signal is now 62db. Ping times went up to 8-12ms. Speeds are slower, around 18-24mbps. Its worse than the TRanzeo Atheros card. I'm really dissappointed, anyone have any idea's? I'd like to believe maybe there's a setting that needs tweeked or something, although I've kept the same channel # and everything as the 2 other test cards....is it because I didn't hook the ground wire maybe? But that sounds silly.....
wispwest
04-13-2009, 09:20 PM
Yes, I forgot to post this in the other "XR" series thread. The card seems to be doing "ok" now, but the throughput is for some reason slower than the other cards I tried.
J_Hughes
04-19-2009, 01:37 AM
If you don't mind supplying a little more info, I have these questions:
- What CPE are you using
- What is your speedtest methodology
- What is the distance from Tower /w what foliage
In my experience with XR2 (600mw) vs R52h (350mw) the XR2 performs 25% - 40% better (throughput) with around -3db to -5db better signal on average. Never had an opportunity to try a Tranzeo radio
wispwest
04-19-2009, 08:34 AM
I'm using a Mikrotik Routerboard, everything is staying the same (Antenna, distance, etc..) the only thing I'm changing is the Mini-PCI card in the Mikrotik.
J_Hughes
04-19-2009, 11:16 AM
I'm using a Mikrotik Routerboard, everything is staying the same (Antenna, distance, etc..) the only thing I'm changing is the Mini-PCI card in the Mikrotik.
well, I ask these things because perhaps there are some optimizations that your XR2 needs that the tranzeo does not need.
- What CPE are you using (if this is simply a PtP this question could be changed to what is your other side. board/radio)
If you have a tranzeo CPE vs a ubiquiti CPE, perhaps there could be changes made here that would be needed for a XR2 but not the Tranzeo. Such as; Multicast Rate, Rate Algorithm, ACK timeout, etc...
- What is your speedtest methodology
If your "speeds" are simply an observation of the signal rates vs actually measuring throughput, it could be simply a charicteristic of the XR2 vs tranzeo. Consequently, a the XR2 if measured with a speedtest it could prove to actually have more throughput, but report a lower Signal rate.
Thanks,
Josh
wispwest
04-19-2009, 12:09 PM
Its an AP that serves around 20 people at fixed locations with external antenna's. I, at my house have a Bullet 2 with a 19db Flat Panel signal is -62db. I thought since its a bullet it would work even better to the XR2 "ubiquiti to ubiquiti". We have a speedtest at our core server room. http://speedtest.wispwest.net One thing I noticed it the RX Rate stay's at 54Mbps more often which slows things down to a crawl. I can lock the TX Rate at 36Mbps and I get a "real" upload to our server of around 13,000 kbps. The download, when in 54Mbps, goes from 1500k-2000k. If only I could lock the RX Rate at 36Mbps I'd be set, but Air OS didn't decide to put that as an option so now I'm set at whatever the AP puts me at. As we all know, when you go higher up in speed rate the more sensitive the connection is, often giving you LESS speed instead of more.[/url]
J_Hughes
04-19-2009, 02:28 PM
Hi wispwest,
Actually, you can control the RX rate for your CPE's with a setting in RouterOS.
You first have to enable Advanced Mode in the radio's wireless interface section (bottom right button)
Then go to "Data Rates" Tab, and select manual data rates.
Then Uncheck 48mbps and 54mbps Rates
Let me know your results
wispwest
04-19-2009, 10:18 PM
Already tried that, its too good to be true. When I uncheck the 48 and 54Mbps rates it for some stupid reason screws everything up. All the clients on B Mode 11Mbps don't register back at all, at a total of say 30 clients only 5 remain registered, so I had to revert that idea real quick.
skyhook
04-20-2009, 12:19 AM
Hi wispwest
what is the distance from AP and Stations (min/max)?
All the clients on B Mode 11Mbps don't register back at all, at a total of say 30 clients only 5 remain registered, so I had to revert that idea real quick.
How long have you waited?
In the RouterOS logs (enabling wireless debug in logging), can you see attempts to connect from CPEs in B mode?
What is the CPEs model?
J_Hughes
04-20-2009, 04:54 AM
10-4 with Skyhook on that,
I have had problems with manual data rates not working only when the lower modes are whacked. It will never try the upper modes if they are gone.
When you are having your bad throughput problems while at 54Mbps mode, what is the CCQ on RX/TX?
Another thought is to try EWMA rate algorithm (maybe then it won't try for an unstable RX mode)
skyhook
04-20-2009, 05:03 AM
10-4 with Skyhook on that
what does it mean?
J_Hughes
04-20-2009, 05:42 AM
10-4 is slang, been around for as long as I've been alive. Maybe only in USA. Usually used on radio. Pilots, Truckers, HAM Radio...
Here's wikipedia's take: http://en.wikipedia.org/wiki/Ten-code
Meaning:
* Affirmative
* Ok
* Understood
J_Hughes
04-20-2009, 05:50 AM
concerning the data rates..... I just checked it out, by default the B modes are not checked. So, you would have to check all the B modes you want supported and one Basic Rate for B mode (usually the lowest rate is set for the basic rate)
http://img6.imageshack.us/img6/852/datarates.jpg
wispwest
04-20-2009, 06:13 PM
^^^^^ that's exactly how mine looked. Until I look at registered stations, the only stations remaining were the ones that had B/G capability. Try it yourself, if you have multiple stations, lock some in on B only, and others on B/G mixed. Then do as we speak by un-checking only 54Mbps and 48Mbps. You'll lose all your B clients for whatever reason. I love Mikrotik but this is something that really pissed me off. Oh and all stations range from 0.3 Miles to 10 Miles.
If you haven't already tried, 90% of the time I've had better throughput on speed rates LOWER than 48/54 Mbps. Unless your really close to the AP and have an outstanding signal, your better off lowering your rates to get FASTER.
rmichael
04-20-2009, 09:42 PM
XR2, which is "high powered" "carrier-class" etc, signal is now 62db. Ping times went up to 8-12ms. Speeds are slower, around 18-24mbps. Its worse than the TRanzeo Atheros card. I'm really dissappointed, anyone have any idea's? I'd like to believe maybe there's a setting that needs tweeked or something, although I've kept the same channel # and everything as the 2 other test cards....is it because I didn't hook the ground wire maybe? But that sounds silly.....
Try disabling adaptive noise immunity (ANI).
J_Hughes
04-20-2009, 10:03 PM
I use 5mhz channel size for all my clients, B mode does not support channel sizes other then 20mhz unfortunately. But if you did as said, I'm sure you've got your info straight