Ubiquiti AC Pro and AC Lite Access Points Reviewed

[thiggins]

We finally take a look at Ubiquiti's cheap AC access points.

Read on SmallNetBuilder

 

[DanH]

Great review.

The comparison to Orbi devices seems odd though since each Unifi AP is cheaper and they have a wired backhaul that kind of puts it in another use category than an Orbi/Satellite.

At the very least wouldn't it be more in line with real world usage to test 3 APs with wired backhaul versus say the Orbi with one satellite? I mean it seems the advantage for unifi APs would be that they don't have to rely on wireless backhaul so you can put them where coverage necessities would dictate, negating construction materials and even the drop off to much extent.

Thought that would get a mention, if they weren't going to be compared to other APs.


Not a knock on the review, just random thoughts I had when reading it. Enjoyed the review.

 

[matthewr21]

I also thought it was odd that this review says you can't manage UniFi switches or phones with the controller. That's actually completely wrong. Personally, I have a switch adopted to my UniFi controller, and the page for the UVP mentions the controller software...

While the internal photos were informative, the comparison between UAP-AC (old design from when 802.11ac was really new) and UAP-AC-PRO was also strange. The UAP-AC-PRO has fewer components because it is a newer, better, and more efficient integrated design rather than it just being a cheaper design (though this is a nice side effect).

 

[pete y testing]

To be honest im not surprised in the wifi results as ( and i think this is what ppl are not seeing ) we are used to seeing router testing and big honking routers and not AP's

AP's are not designed to be the biggest baddest wifi on the block , they are designed to work in multiples and so each one doesnt need to be maxed out , even the dws and mesh systems cant really be compared

it is a fact AP's are a class on their own and can and should only ever be compared with other ap's

however i would still recommend unifi stuff if ppl have structured ethernet and choose to use 2 or more ap's as thats exactly what they are designed for

dws and mesh systems are designed for those primarily dont have structured wifi and thats their place

"BHR" big honking routers are different again as they always transmit from a single location and push the wifi power envelope and beam forming algorithms to achieve what they do

this is why i suggest that ap's should be a separate category and tested and compared as such

pete

 

[joecoolman]

I'll chime in with my $.02. The range and throughput numbers clearly speak for themselves, but I've experienced a reliability with unifi gear that I wasn't getting from my other routers. Hopefully it's not $150 of placebo, but I've been quite happy with the little guy. The best thing I can say is I don't notice my wifi; which is fantastic.

 

[pete y testing]

While the internal photos were informative, the comparison between UAP-AC (old design from when 802.11ac was really new) and UAP-AC-PRO was also strange. The UAP-AC-PRO has fewer components because it is a newer, better, and more efficient integrated design rather than it just being a cheaper design (though this is a nice side effect).

the old draft ac unifi ap ( the square one ) was pretty much a dog , never had any joy with it and it cost a bucket load at the time , so your prob right that its not a good comparison but again the ap's are designed to run in multiples and spread the wifi throughout the house , meaning the clients are much closer to the transmissions

 

[jramskov]

I'll chime in with my $.02. The range and throughput numbers clearly speak for themselves, but I've experienced a reliability with unifi gear that I wasn't getting from my other routers. Hopefully it's not $150 of placebo, but I've been quite happy with the little guy. The best thing I can say is I don't notice my wifi; which is fantastic.

That's what I've heard from others too, it might very well not be the fastest, but it's very reliable and fast enough. It'd likely be my choice if I had the option, but half our house isn't cabled and it'll be quite the effort to do that which means I'm looking at Orbi instead.

 

[mdamda]

Does anyone have any info on how the Pro compares to something like say, the Asus AC68U in terms of WiFi Range / throughput at range?

Based on comparison of the SNB charts, the AC68u is superior to even the Pro.

Any pending SNB review of the UBNT AC HD?

 

[thiggins]

I also thought it was odd that this review says you can't manage UniFi switches or phones with the controller.

Latest UniFi release removed VoIP support. I was usure about switch support and could find no references to their support in the UBNT Community.

What switch features does Unify support and how are they shown in the console?

 

[thiggins]

DanH: All performance comparisons were on individual units. This was not intended as a mesh performance review. Comparison was done to mesh products because the nodes are basically 2x2 APs and are the only comparable products I had tested with the same process.

 

[DanH]

DanH: All performance comparisons were on individual units. This was not intended as a mesh performance review. Comparison was done to mesh products because the nodes are basically 2x2 APs and are the only comparable products I had tested with the same process.

Sounds good. I do have the Orbis because I believe in my apartment they were hands down the best option for me at the time. Now however I am moving into a new home with Ethernet drops everywhere and am really not sure if that would be the case versus wired APs.

I was hoping the comparison was $400 Orbi plus satellite (because that is the only way you can buy them and really the only way they work) vs 4 of the $100 APs wired up with cat backhaul and then compare throughput and coverage. That would be how each is used in real life.

That obviously isn't feasible for you to do, and I get it. Once again I appreciate the review and work you do it is valuable and first rate.

 

[clearzero]

I read the Unifi board almost daily and this review has a good thread going about it. Looks like UBNT has issues with your 2.4GHZ testing and has reached out to you to help with the testing process.

Is that true? You started the review saying you were on the outs with them based on the Amplifi review. They seem eager to set this straight.

https://community.ubnt.com/t5/UniFi...tBuilder-UniFi-AC-review/m-p/1865861#U1865861

 

[pete y testing]

Looks like UBNT has issues with your 2.4GHZ testing and has reached out to you to help with the testing process.

pretty arrogant of that staff member to post the way he did before first waiting for a reply from tim , and suggesting that their method cant be somewhat flawed suggest that something else is at play here , perhaps a Volkswagen diesel cover up type thing ????

 

[UBNT-Brandon]

UBNT-Brandon here. Yes reached out and I'm working with Tim.

So the case here is control of the RF path. When doing high attenuation testing often the propagation path ends up not being through the attenuators, and rather through some leakage.

Signs to know this is happening is 'flattening out' in terms of throughput at high attenuation values - where adding more attenuation doesn't further decrease throughout.

This happens in all sorts of RF testing and is tricky to avoid actually.

I'll be working with Tim to see current techniques and precautions and how to prevent the leakage.

So the interesting bit here is since the RF is going around the attenuators at high attenuation levels, you can just keep adding attenuation and you'll see no difference.

In RF testing you'll see cases where a tech will come back and say 'I added every attenuator in the lab and we still had signal!'. This is actually really common to hear... and is another sign that the RF path is around the attenuation, not through it.

Anyways, perhaps will do an article on this as well and blog... it's an interesting subject which impacts WiFi to satellites to walkietalkies.

In fact one of the times where this is seen is on the DANDE student satellite when doing link margin testing there. Engineers working on it saw the same, and had the same confusion.

So the interesting bit is with the thermal noise floor equation (which can be approximated by -174dbm + 10*log10(bandwidth_Hz)) you can calculate the physics imposed limit on communication channels - and quickly tell if leakage is corrupting your measurements.

You can also do this for various MCS rates, which is why Omar was asking for the MCS rates observed, which will tell exact leakage.

Anyways, as you all know, UniFi - and its impact on the world - is my passion.

So I love to do anything/everything I can do to make sure anyone using the system is getting the most out of it, and the world knows what it can do and how it can help to better connect schools, hospitals, businesses more efficiently in terms of dollar cost and time cost world wide.

Best,
UBNT-Brandon

 

[sfx2000]

@UBNT-Brandon - please explain why it's the UniFI's that have trouble and not the other devices under test?

I've been in the business for over 20 years, and reviewing @thiggins test plans and test environment - and the effort he's put into it, I'm reasonably certain he's done a good job there.

Just as a bit of background on me - I'm a Principal Engineer with experience at the Telco, Device, and Chipset level - mostly on protocols, architecture, and RF across 3G/4G/LTE/Wimax/WiFi systems.

 

[sfx2000]

So the interesting bit is with the thermal noise floor equation (which can be approximated by -174dbm + 10*log10(bandwidth_Hz)) you can calculate the physics imposed limit on communication channels - and quickly tell if leakage is corrupting your measurements.

Ok - but that would impact all devices under test - so...

Tim's chamber has been fairly consistent over time - not saying there might be an error, but I'm not seeing one. But like you perhaps, I'm seeing the report - however, without a good explanation from the vendor, I'll stick with SNB's observations.

 

[UBNT-Brandon]

So each is set up at each time. And little leaks can have huge differences.

As Tim mentioned he tests one at a time. So there's potential for different levels of leaks during tear down and re-setup.

Cheers,
Brandon

 

[UBNT-Brandon]

So sfx2000 notice the Eero and Luna plots verse others also see this. I can probably find the plot to point it out.

Cheers,
Brandon

 

[UBNT-Brandon]

Oh and not saying he's done a bad job... just pointing out what's going on here and trying to help.

Also I can show you what I'm talking about. Once you see you'll immediately get what I'm talking about here.

Cheers,
Brandon

 

[sfx2000]

Read on SmallNetBuilder

@thiggins - Clean up/Clarify page 3 - this is where folks are getting confused - especially the charts that don't include the UniFI's...

this will likely clear things up... Looking at the charts, I'm actually not seeing a problem here.