I’m working in a five stars hotel, shaped like a capital i (which means, all the rooms are side by side, one after the other, in the same plan, floor by floor).
We have tried putting 2 rbmetals and 1 mANT almost in front of the building in order to cover all the rooms, but still, as the building is long and tall, we would have to add more and more wireless to cover every room, with the eventual trouble of 2.4GHz interferences.
One of our projects is to add ethernet cable in every room to be able to use VoIP phones, and I came up with the idea to add a hAP mini in every room (sealed on the wall), or at least one for 2 contiguous rooms, manage them with CAPsMAN, not less because we have thick concrete walls between rooms (~20cm+). I’m looking for comments about that, except the cost, which I know, would be huge (58 rooms, 4 suites).
What about the power transmission, it is possible to reduce it to the minimum, enough to cover the room, but with the least interferences with other rooms? What about the heat, is it safe to put them inside a sealed wooden box on the wall (no vent)? What kind of CAPsMAN device would be powerful enough to handle 58 CAPs (power needed, bandwidth needed?) + VoIP phone without flinching in case of a fully booked period, etc…
Also, if anyone know a good seller, capable of delivering me at least 60 of them, and eventually capable of doing a discount for this kind of volume please?
I really would like answers about the set-up I describe above, but I’m also open to suggestions.
It’s certainly more expansive… And I agree with using 2.4/5 together. But my questions remains (except for a ceiling router), what about sealing them in a wood box, Would the temperature be a problem? What about the transmission power, will it be possible to keep it lower as possible in order to avoid detecting too many AP on a device? What about the characteristics needed by a capsman/gateway to handle all that?
It is possible to lower TX power of wireless … using a trick of setting higher antenna gain. As ROS tries to keep EIRP (transmitted power plus antenna gain) within legal limitations, setting higher antenna gain without actually using high-gain antenna means lower transmit power.
Yuo’ll have to experiment a bit to find the optimum setting.
I’ll do some research about that as this is new to me. About the legal limitations, currently, our wireless routers are set on “no_country_set” because obviously my country (Republic of Congo) doesn’t exist in ROS and anyway we don’t have any regulation. So, what about the trick? Will it still work with no country set?
Bit of experience:
I have seen similar setup in The Sebel (Melbourne Docklands) - each room having own AP with SSID named according to room number. I can definitely recommend such decision as it delivered absolutely best wireless performance I have ever seen in hospitality business: https://www.speedtest.net/result/a/3382690514 (this was not peak but stable result, absolutely no congestion)
Although, it will be challenging to set up frequencies properly - if you leave it on “auto” it will end up with lot of interference as Mikrotik scans for free frequency only during interface start and nearby devices might be not ready yet. I believe best would be to split your devices into 3 groups and assign each group channel 1, 6 or 11. Then you can easily put them in order 1,2,3,1,2,3… so there will be always 2 rooms between same channel. On next floor, you just offset groups by one so you don’t have rooms with same channel exactly on top of each other.
With this approach, you won’t have to worry about TX power that much.
Re. “sealed wooden box” - nope, nope! Even non-wooden sealed box might be an issue. Wood is great for thermal insulation so it will just make your problem worse. Keep in mind that these devices need to dissipate plenty of power in heat.
Re. “hap mini” - personally, I would also advice against it, but I don’t really know what else to suggest. The price is simply unbeatable. Just keep in mind that it might be cheaper to keep your AP and phones on POE if you look at it from long-term point of view. The maintenance of adapters and power points does not seem high but over years, it can grow a lot.
Re. previously recommended “wsAP” - I am not sure how does it work with extra cables (for example if you want to daisy-chain your phones). Seems almost like you have to decide - either have cables or locked cover. Not both at the same time.
Re. 951G - the country info lists maximum power, not antenna gain. Sum of TX power + TX antenna gain must not exceed this value. Obviously, included antenna has specific gain which cannot be changed. By specifying antenna gain in your system, you do not change real antenna gain, you just prevent radio to transmit more, than legal limit.
If you set up false antenna gain which is higher than real gain, you will practically decrease maximum TX power without specifying TX power manually.
If you set up false antenna gain which is lower than real gain, you will increase maximum TX power. That however, might cause stronger than allowed signal and lead to legal issues. ()
Technically, you can say that the number is also “maximum antenna gain in dBi” because if you set 30dBi antenna gain while you use it in 2402-2472 range, you will achieve 0dBm TX power, therefore disabling the output.
Not exactly true. TX power expressed as 0dBm means TX power of 1mW. Which is almost but not quite zero. Negative TX power, when expressed in dBm, is completely valid.
Re antenna gain: by definition a (hypothetical[*]) isotropic antenna has gain of 0dBi. A real dipole omni directional antenna has gain generally around 2dBi (but much depends on exact length of it, should be certain ratio of wavelength to maximize gain). Typical directional antennae (such as yagi, log-periodic or panel antenna) have gain around 10 dBi (up to 20 dBi). A parabolic (i.e. satellite) antennae have gain 20-40 dBi (depending on diameter of the dish relative to wavelength). A typical in-device (so called patch) antenna has gain around 0dBi .. or even slightly negative if its badly constructed. Although a very good patch antenna can have quite good gain.
So: antennae built in routerboards have typically gain of 0-2dBi.
[*]isotropic antenna is purely hypothetical as it is impossible to construct antenna that would have completely symmetrical radiation pattern. That would mean that radiating element would be a electrical monopole which doesn’t exist. The closest to it is a small dipole which has symmetrical radiation pattern in directions perpendicular to the dipole axis but has a distinct radiation pattern in vertical cross section … in direction along the dipole axis it doesn’t radiate any EM energy. Which is why it has gain of around 2dBi (when considering direction with highest radiated power) … the energy that is not radiated along the dipole axis is actually radiated in direction perpendicular to dipole axis.
I was more thinking about the same SSID for every room to avoid a user having potentially 10 SSIDs available and not knowing what to choose (Room, above, below and 2 adjacents + externals). Unfortunately, we have this kind of users. Our hotel is abbreviated “OPH”, and we can detect external SSIDs like “APE Direction” or “Site CMA”, and yet, we have users sometimes calling the reception saying they can’t connect. I think it would be better for user experience to have the least number of choice, and not trying to connect to his neighbor AP “voluntarily”.
Did you manage them using CAPsMAN, if so, which device did you use to manage? Do you use a hotspot there, open wifi, or individual WEP/WPA?
Great suggestion! If I stick on the idea to give the SAME SSID to every room (or basically, all around the hotel with hotspot for instance), keeping the lowest TX power will insure me a device will stay connected on the room’s AP (vs a very weak neighbor signal), isn’t it?
Ok. I was wondering if heat was an issue there because I have a personal hAP lite TC, it doesn’t seem to dissipate that much heat… Anyway, in that case, I’ll have to consider @mistry7’s suggestions (wall mounted/ceiling).
Why not telling us what AP you are using . I agree with the maintenance issue. VoIP phones usually have 2 ports, so I can plug directly the phone to the network, then the AP to the phone. But I don’t know how much I can rely on the phone as a switch, we’ll have to buy phones with PoE IN and OUT or simply PoE OUT with the right voltage. I also wonder if PoE will reach the furthest room. And it also implies having to buy a kickass switch/router to be able to provide PoE to all these devices…
It seems you can plug all your cables then lock the cover with special screws, according to the product page. And according to that page, you can connect both analog and LAN on the back of the device. It has PoE IN and PoE OUT. Which would be great to chain an IP phone (same for the cAP). With a quick search, the wsAP seems to be a little bit cheaper than the cAP, and as we are still on analog phones, it would ease us the transition to IP phones.
This is a five star hotel, so please use the hAP ac lite or the wAP AC or another AC product and run 5Ghz. This will give you plenty of channels to avoid interference. Turn the power down and run the same SSID unless you have a particular need to do otherwise.
Example 5Ghz layout
5Ghz is better for density
For additional research on this subject, you can read about Wireless Networking from this site. Choose the eBook PDF for a small download. Some light reading here.
Yup, you are absolutely right, sorry for that misinformation. Not sure what I was thinking when I wrote it. Maybe I tried to simplify it too much.
You are right. I always forget that some people get confused so easily.
Dang! I am sorry, this was terrible misunderstanding. In mentioned hotel, I saw this from guest point of view. I was not admin/tech guy in there. It was quite a while so I do not remember exact details. Those devices were definitely not Mikrotik/Unify, just some not-well-known brand. Also I know it was not a hotspot - just pure unlimited network. Unfortunately I forgot if there was encryption or not.
It should but it depends on the client device itself. There are some brands who claim they have revolutionary band-steering or roaming on AP, but technically, all they usually do is break standards. It is only client’s decision, where it connects. Some client devices actually get confused if you enable some weird roaming function on your AP. (usually works the way that when signal is lower than treshold, they send a “disconnect” frame to client)
It does not seem but it is. Nobody can beat basic physical laws. hAP lite, hAP mini and mAP lite are actually lowest-powered devices from mikrotik with specified consumption only 3.5W. That is maximum value so the real consumption might be actually lower. I don’t have hAP lite/mini next to me but mAP lite is consuming 1.5W when idle or 2W when I run iPerf via wifi or btest with some nonsense firewall rules which consume around 60% of cpu. Within minute I can feel that it’s temperature is above ambient temperature as it feels gently warm. If properly insulated, the temperature would grow and grow and grow… I don’t think it would burn but internal components would definitely be not happy about it. (another example - aquarium heaters for small tanks are often around 5W which is enough to heat un-insulated water by 10 degrees from ambient temperature)
Simply said - the power is low, so even small airflow will easily dissipate heat. But if airflow is blocked and box is made from material with low thermal conductivity, temperatures will grow.
In few hospitality venues I unfortunately inherited UniFi setups which are causing serious troubles but managers refuse to spend more money on networking (they say they spent enough). In office building I started network by my own design and (again due to lack of funding) I ended up with hAP ac^2 which has sufficient routing power, excellent switch chip, reasonably fast wifi and PoE-in. Thanks to it’s universality I can use it as anything in every possible scenario - router, firewall, switch, ap … In the end, I don’t need to keep many different spare devices which really simplifies my purchasing and stock control. Only downside is lack of PoE-out so I can’t daisy-chain them or plug PoE phones to it. Not sure though if it would work well for you as you don’t really need so many ports and you would have to hide/protect it same as hAP mini.
Cant talk about every brand but I have few Yealink phones connected like that and there are no issues reported.
Good luck with that. I didn’t find anything like that (but maybe i was just looking wrong ). If you find some, even later, let me know here please!
PoE should be able to reach approximately 100m, but with greater distance, you add more loss on wire which means it needs to draw more power from supply. I have no idea about wiring in your building so I will not guess at all.
Absolutely, but you would need switch anyway.
Really quick estimation: Lets say you have 60 AP and 60 phones. If you go non-poe way, it means you need to dedicate 2 power outlets in each room for this. Unless you have plenty of outlets already, you will have to install new and that will cost some money. If it is just 10 USD per each room (I bet it would be more for labour+material), you end up with at least 600 USD for extra outlets for whole building. Now, look at price of POE switches - for example CRS328 with PoE is just 375 USD while CRS326 without PoE would be 199 USD. Extra price for PoE is 176 USD per 24 POE ports = 7 USD per port…
(In near future, there is supposed to be CRS354-48P as it was presented around April 2018. It will certainly be even more expensive but price per poe port might be actually lower. Unfortunately pricing or release date is still unknown)
Only normal screw is included. The “special” screw must be purchased separately and it always come with tool as well so to get 60 screws you will end up with 60 tools as well… Personally, I would not waste money on that because the screw and thread sizes are standartized so you can buy same size with unusual head from your nearest hardware shop.
Not sure about possibility of locking it with cables attached. That is not explicitely stated. Other similarly designed devices are different: Look for example at PowerBox Pro. There are clearly visible sealed holes on bottom of the cover. Similarly wAP (and other models from same family) has those holes on bottom and it’s description specifically says: “there is a special opening on the back of the case”. I couldn’t find either holes on bottom nor opening on the back from any photos/videos of wsAP.
It would be good if some owner of wsAP describe, how it really works. If we don’t get some reply here, I might ask my distributor.
Thanks a lot for your documentation. I’m indeed in a developing country, so it’s very useful. BTW, our bandwidth is 10Mbps (fiber) for ~4’500€… Developing, but very very expensive…
Hey!!! This is actually a great idea for us! As we’ll have an individual router for each room, we could allow users in rooms to connect to an open network without hotspot, and enable it only on common areas
The complete story was that we have a fake wooden wall, which can be opened just by lifting panels. But in the end, the issue is the same, there is no airflow, so we won’t take any chances on that. The best would be to use a ceiling wireless router. Outside, ventilated, not reachable. And if the customer wants to connect using a cable… Then he will connect to the phone
Well… I wanted to let the customer connect using RJ45 or Wifi in the room, but I wonder if it’s a critical requirement. Routers and phones usually have at least 2 ports. The routers are more likely to provide PoE out, good to power the phone. The phones usually don’t, so it’s best that way. If for any reason (financial for instance) I HAVE to buy mini, I will embed them in the AC case (ventilated, refrigerated at most. The case is bigger than the actual AC, so there is space to put it without a risk about moisture)
As said, I think it’s better to chain the phone on the router. PoE IN phone ← PoE IN/OUT router ← PoE OUT Kickass switch
Out technicians told me they will be able to add the cables under 100m each. So it’s fine.
I’ll try to evenly distribute power/network between 2 switches. Really quick estimation about power use: cAP ac consumes 13W, a phone found online consumes ~3W. So 448W for 24 rooms. CRS328 can handle 450W for whole 24 ports. It should be fine.
I think you got almost everything sorted Just one note I forgot to mention earlier: There are many PoE standards (or proprietary solutions) and the difference is not just voltage (very common misconception). If you look for PoE phone and you want to daisy-chain it behind AP, it needs to be compatible with Mikrotik’s passive PoE, not just 802.3af/at because there are no models, which can pass through 802.3af/at. (except RB960PGS and RB960PGS-PB).
Again I am little unsure about wsAP AC lite official page says “PoE in 802.3af/at, PoE out Passive PoE” but in brochure states “PoE-in 802.3af/at, PoE-out passthrough (Ethernet 3), 18-57 V” which suggest that PoE-out is exactly same as PoE-in (otherwise it is not real passthrough).
Similarly, cAP ac official page says “PoE in 802.3af/at, PoE out Passive PoE up to 57V” and brochure confirms it with “PoE-in 802.3af/at, PoE-out (Passive, Ethernet port 2), 17-57 V”
Due to that weird “passthrough” word in wsAP AC lite, it might be able to feed 802.3af/at into your phone. It would be much easier to find such phone as 802.3af/at is very common.
But don’t take my word as fact on this. It is just my suspicion and it needs to be confirmed by some owner.
Almost . The last thing is the manager. We currently have 2 RB2011, the main gw on which our wifi routers and desktops are connected via different switches, and on which is connected the 2nd RB2011, on which there is the IT office (few computers), the IP cameras and all the IT stuff (VMs, NAS, etc…). We also have a RB3011 waiting to be setup. Ultimately, all the phones (64 phones for rooms, without counting office phones) and room routers will be connected to 3 switches connected to 3 ports on the same router. Would a 2011 or 3011 be capable of handling all that, phones + wifi without flinching? Or will it be better to connect the 3 switches to a router, itself connected to the gateway? I may be not clear enough, I’m taking about CPU power.
I’ll pay attention to that also. I forgot to mention, I will also have to pay attention to our UPS… The electricity in the server room wasn’t designed so well, so we’ll have to check all that also. Anyway, when I’m done (and before), you all are welcome to visit our facility
@axe50397: I must apologize for my previous statement about heat… I realized that my GrooveA 52 has temperature sensor on PCB so I can relatively easy test these conditions.. After all it seems that my fear of overheating was too big and you should have no issues:
10:40 monitoring enabled while idle system sitting on my desk (ambient temperature in office was around 22°C). wlan was turned off and there was almost zero traffic (just winbox and snmp). Real consumed power consumption was 2.2W and temperature on PCB was around 45°C. Surface was gently warm.
12:00 I enabled wlan interface (just enabling wlan without traffic increased power consumption to 2.7W), connected it to local wifi and started btest to utilize both wlan module and cpu. Groove was still sitting on desk. Power consumption went up to 4.7W and temperature on PCB slowly stabilized around 52°C. Surface was warm and I noticed that a lot of heat is actually dissipated via antenna connector which was hot.
13:10 I wrapped Groove into several layers of bubble wrap to simulate thermal insulation. Temperature raised to 66 degrees. After few hours, I could feel heat through all layers of wrap. Also I noticed that consumed power was around 5W.
16:20 btest disabled, unit left intact in wrap, temperature is slowly decreasing.
16:40 Unit is removed from wrap. Plastic surface is hot and antenna connector even more - it almost burned my hand.
Official specs of groove says it was tested to +70°C (ambient temperature) so PCB should be capable to reach much more without being damaged. Also, with so high temperatures, it seems that heat dissipate well enough.
Maybe you can really save some money by placing your devices behind fake walls (unless your temperatures in there go too high)
Thanks for your reply, that’s not a problem. Anyway, as you said, heat needs to dissipate somewhere. Actually, I wrote about closed wooden box, but we do have fake walls with large panels removable by hand.
More precisely, one side of the fake wall is open (not visible by the client) but there is no airflow. Indeed, the budget would be way better, but not so sure… First, there is no lower cost AP with 2.4 and 5GHz. Which means, if for budget reasons we start with something like the hAP mini, we will have to progressively change them. The final price will increase a lot more, and we will have too many spare APs.
For all these reasons, I think the best would be to buy at least, 1 AP in every 2 rooms. i.e: Channel 1, Empty, Channel 14, Empty, Channel 7, etc… That way, nothing to change when we add routers in between. Half the budget to start. hAPs would have been great, but limited to 2.4GHz.
After all, I understand it would be good to anticipate and consider the 5Ghz bands, but if the distribution is well designed, is it a critical requirement…?
Going 2.4GHz only is very risky, since then you are designing it for current minimum bandwidth and not future, and assuming that nobody else will add more 2.4GHz AP’s in the area. If you have 2.4GHz in every room, you are going to have to lower the power substantially to help get rid of interference (or only put a full power AP every two rooms instead of every room), and lack of line of sight (if you place the device in a wall) is going to mean that the power is going to be divided equally between the room on one side of the wall and the room on the other. If there is metal in the wall it can reflect the signal causing interference between the AP and itself. Even your wooden box idea I would avoid – the best place for the AP is in the open or with a thin plastic barrier, not in a wall or wooden box, and even if you put it in the open and lower the power, if someone else puts in a few high power 2.4GHz AP’s near the hotel, there goes your reliable Internet and there is nothing you can do to fix it short of going to 5GHz.
With 2.4GHz there are only 3 channels you can have that do not interfere with each other. You might be OK if the hotel is in the center of a giant parking lot, far from other AP’s, but if it is in an urban area with other 2.4GHz AP’s nearby then the service will be pretty bad. This is especially the case in places like yours that lack regulation, and so other 2.4GHz AP’s can be turned up in power well beyond what they need to be and cause needless issues with other AP’s. I’ve stayed in some rather high end hotels (Hyatt Regency, etc.) where the wifi was often terrible due to interference, even though the signal received from the hotel AP was quite strong.
The wsAP AC makes the most sense. It can be mounted on the wall instead of in the wall, then you ensure that the AP in each room is going to provide the best power for that room (instead of sometimes getting better signal from the next room over), and get 5GHz which means that you will be safe even if there are nearby high power 2.4GHz interference sources. It is worth the extra cost, unless you want to put in a solution that you are going to have to replace entirely in 2 years time. The 2.4GHz radio is lower power than the 5GHz, which means that the power is already turned down for you, suitable for the hotel setup.
If you want to save money, the safest way is by putting in wsAP AC units every two rooms instead of every room, although that doesn’t provide you a solution for the other IP phones (mikrotik does say that you can power IP phones from the wsAP AC in https://i.mt.lv/cdn/rb_files/1536211562wsAP-wg.pdf, “The E3 port supports PoE output, so you can power an IP phone from this port.”). Adding extra wsAP AC units for every second room later on will be much cheaper than the cost of replacing every single AP in the entire hotel.