well. I got my boards today, and they are a bit different than I thought.. but still can be good.
They only have 16 meg flash on them. I thought since they had a compact flash slot that they would boot off that.. like a wrap board.. but they did not, and they are like the routerboards in the onboard flash.
you can get 32megs flash but you have to special order them, so I have to send mine back in order to get the other 16meg flash soldered on.
That’s an Intel IXP “X-scale” CPU board. No ROS support on that.
I have a couple of those with some other software running on them.
Performance is quite impressive but without a decent OS (and I mean RouterOS) it’s not so useful. So they’re sitting around unused.
If I get the flsah up to 32 megs.. why wont the routerOS run on it?
what do you mean? it does
about the power issues - RouterOS uses the values that are written in the cards EEPROM. if the results do not satisfy you, you can contact the manufacturer of the card.
about the power issue
I am not saying the issue is a mikrotik one at all. I am saying it is an issue.
I was talking with our head engeinner about that very thing.. calling ubitquiti. I am just trying to get all the info I can.
In my opinion Mikrotik is the best thing since sliced bread
Every time I use other vendor, I see how much I really love MK. This whole power issues is stemming from 900mhz, The UB cards state they use 700mW.. I assume the power needed for that is very high.
I am testing the motorola 900mhz canopy and it works ok, but everything else about it sucks!. I want to be able to use MK for 900Mhz and for that I think we will need very high power, that is, if UB every gets the SR9 ready for production.
Mikrotik routerOS= GREAT
Motorola BAM/Prism = CRAP!
Try to test sr-5 with 5MHz channel bandwidth, there shold be more
I know that these 400mW is “marketing” but i don’t care, that’s about 6dB more than other cards, sometimes this can help a lot (form unstable -89 to -83)
And there is still option that your power meter isn’t calibrated on all bands.
I’ve testes prism 2,5 mPCI and PCMCIA with HP spectrum analyzer and never saw 200mW, only demarctech 300mW model have about 180mW real world.
So do not take this results real, because that’s only pointer, when you can see what happeneing with power.
I don’t belive that expensive Rhode&Shwarz, HP or Anritsu power meters and spectrum analyzers cost that much because they are most famious brands…
PS. Have you dissasemble your sr-5’s ?
There’s one hardware method to power up card with direct 5V so the amp should give about 3-4dB more clean power
PPS. Which connector ? UFL/MMCX antenna-a or antenna-b ??
PPPS. Have you tested Tonze 684HA ?
We tested IXP 425 533MHz with other (non-MT) software.
Actually the hardware is very impressive, with ethernet - wireless bridged throughput as fast as the best x86 platform, i.e. no CPU bottleneck.
however that software wasn’t complete enough for us to use “for real”, so it’s left sitting in the lab.
Rigth now I think the only options are fast x86 platforms.
We have AMD “NX” fanless at 1GHz, which we demonstrated 200Mbps ethernet-wireless bridge throughput split between 5 radio interfaces, all running on a single box. A press release on our website if anyone likes that sort of thing
My take on it is this: If there is going to be no “port” of ROS to IXP platform, then optimised x86 with low-heat mobile processors is the way to go.