UBNT-Brandon
Occasional Visitor
Meant to also mention we reached out to offer AC-HD, switches, and other new models to try out as well.
Ubiquiti AC Pro and AC Lite Access Points Reviewed
- Thread starter thiggins
- Start date
sfx2000
Part of the Furniture
I see it - UniFI's do ok, nothing exceptional, but similar to other AP's...
I also saw your thread/post over on the UBNT forums, and that's not a fair post either... in that post you had serious questions on how he tested. I think the methodology is very fair in how Tim does his tests - as that's pretty much apples vs. apples in the chamber. Pretty objective vs. someone over on some other tech site doing subjective payola with little editorial oversight...
Tim's been pretty fair if there was a true error in the testing, which I respect as an engineer...
Not sure what's the dynamic between UBNT and SNB's editors, but please solve it.
sfx2000
Part of the Furniture
At the same time - not very different from my Airport Extreme AC's that I run in AP/Bridged Mode (which is basically just AP) - with a Netgate SG-2440 (pfSense) doing the routing.
I was considering the UniFI AP's to replace them as Airport Extreme AC's are getting hard to find...
Perhaps not now unless this editorial conflict is resolved...
mlg321
Regular Contributor
Relax. Consider ourselves lucky that someone from UBNT is posting here. I would take that as a clear sign they are trying to "solve it". There's more than one way to skin a cat, I'm certain that Tim will get to the bottom of whether his testing was flawed or not.
Ps @thiggins I am seeing "this message is awaiting moderator approval, and is invisible to normal visitors." Posted 03/15/17 7:05pm.
Last edited:
I'll be honest . . . Something does seem a bit off. I have done basic testing with a UAP-AC-Pro and a Lenovo laptop with an Intel 2x2 card. The 2.4ghz performance over UDP was definitely over 100mb/s. As for not connecting past the lower MCS rates I wonder if that's due to the DSSS rates being disabled.
UBNT-Brandon
Occasional Visitor
Just as a quick update - Tim and I have connected so we'll be discussing over Skype the new test setup when I'm back from WISPAmerica on Friday. Will get it dialed in.
pete y testing
Very Senior Member
perhaps thats the way it should have been handled in the first place
UBNT-Brandon
Occasional Visitor
Yes, reached out right away when I saw the review - I agree.
pete y testing
Very Senior Member
i mean by before your post on the unifi forum suggesting that tim was totally wrong before examining the processes and procedures used
UBNT-Brandon
Occasional Visitor
So it is/was necessary to point out the error.
UBNT-Brandon
Occasional Visitor
So I'm working with Tim to get his new test setup running such that an error like this does not re-occur.
In the meantime - it's necessary to make the error clear. Once corrected, Tim and I will be able to show the source of the error, and show folks here how to prevent such an error as well.
Like I said, it's pretty easy/common to see this sort of leakage when testing higher attenuation values. It's not an easy thing to get right.
I hope that makes sense - does it?
Thanks,
UBNT-Brandon
In the meantime - it's necessary to make the error clear. Once corrected, Tim and I will be able to show the source of the error, and show folks here how to prevent such an error as well.
Like I said, it's pretty easy/common to see this sort of leakage when testing higher attenuation values. It's not an easy thing to get right.
I hope that makes sense - does it?
Thanks,
UBNT-Brandon
UBNT-Brandon
Occasional Visitor
Oh, also, to be clear - I had examined the processes and procedures used. Like I said above (and in the Ubiquiti community posts) - such an occurrence is common, and readily identifiable if you do this a lot.
So it's pretty quick to see should this be what you do all day. And also easy to know the mechanisms to lead to it.
For example you can see the 'flattening out' at high attenuations (after throughput has been monotonically decreasing). This only occurs if the signal path is no longer through the attenuation.
In this case one could add an arbitrary amount of attenuation and the throughput would remain flat - because the signal is 'leaking' around the attenuators.
It's kind of like getting an eye for fixing problems as a mechanic. If you've seen the same thing many times before, it becomes easier to diagnose accurately and quickly.
Cheers,
Brandon
So it's pretty quick to see should this be what you do all day. And also easy to know the mechanisms to lead to it.
For example you can see the 'flattening out' at high attenuations (after throughput has been monotonically decreasing). This only occurs if the signal path is no longer through the attenuation.
In this case one could add an arbitrary amount of attenuation and the throughput would remain flat - because the signal is 'leaking' around the attenuators.
It's kind of like getting an eye for fixing problems as a mechanic. If you've seen the same thing many times before, it becomes easier to diagnose accurately and quickly.
Cheers,
Brandon
pete y testing
Very Senior Member
only if such tests arnt squed to advantage any one product over another and still doesnt explain why the orbi didnt or wasnt effected , as its easy to setup and report results to the benefit of a certain product , the base line needs to be exactly the same for each device and how its tested otherwise the results are meaningless
i assume you have tested none ubiquiti gear in your posted test chamber to at least get some sort of base reference which is not designed on your specs
tim tread very carefully as to how any why changes need to be made and how they effect base line and subsequent results , your testing method has held the test of time and without judgement apart from now !
Last edited:
This is a big reason I like the UBNT community boards and by extension the products. The UBNT staff are very responsive and helpful. I thought Brandon's post on the UBNT board was extremely insightful and at a minimum very interesting.
I've been a fan of Tim and his reviews since his first review web site around the turn of the century. I'm also a satisfied Ubiquiti customer for both my business and personal use. And in particular Brandon and the rest of the Unifi crew are always responsive to issues. I'm certain that the two will work together to add clarity.
DanH
Regular Contributor
@UBNT-Brandon if you need someone to test out some HDs in an environment that is totally not setup in any fashion conducive to said tests, and then never return your HDs back to you. I will be that guy for you. I require 3 of them and a US-16-150W (which I also will not return). I will also not publish any results from said tests (that I totally promise I will run). Just saying I am a good guy like that. Heck Ill even test them if I can get them at the old beta store price, cuz I am nice.
sfx2000
Part of the Furniture
I think everything will work out - there are two sides to every story, but everyone needs to respect the findings...
If there's an issue with the test bench - that's ok, as we'll all see better results across all devices under test.
Similar threads
