I just replaced an XR2 card at a site with an r52hn card, its using chain 0 (j4).

I found it odd that it doesnt output at 24dbm in B mode as it indicates in the current TX power tab.
How do I get the 24 dbm power output when using only one chain?
It reports 21dbm for all B modes 1,2,5.5 and 11.

Same applies to G modes, they specify 22dbm for the lower rates, I expected 25dbm.

Thanks!

I WAS TOLD IT DOES SO AT NO LOAD. When u get the clients connected it would give out more power

But I have 32 customers on it, 31 are B mode, 1 is N mode.
All are working well, just at the unexpected 21dbm power output.

The card will only give maximum power when both chains are in use. And that can only be if you use 802.11n (or ´mimo´) protocol.
You use ´b´ so only one chain thus 3dB less.

The card will only give maximum power when both chains are in use. And that can only be if you use 802.11n (or ´mimo´) protocol.
You use ´b´ so only one chain thus 3dB less.

So why is the card advertised as 24dbm in B mode? 25 dbm in G mode?

The card will only give maximum power when both chains are in use. And that can only be if you use 802.11n (or ´mimo´) protocol.
You use ´b´ so only one chain thus 3dB less.

So why is the card advertised as 24dbm in B mode? 25 dbm in G mode?

That's actually a good question!
I have been reading closely the .pdf (http://routerboard.com/pricelist/downlo ... ile_id=202)and see indeed that for ´b´ they give 24dBi where ´g´would give 25dBi in 6Mbit. etc.
I don't exactly know what the reason is for these differences is. Maybe some technician can give us closure on this?

ON the other hand, you use your radio to contact ´b´ clients AND one ´n´ client. Now we have three scenario's:
1. ´n´station is on b/g/n mode but adapts to ´b´ mode because AP is set to ´b´ mode.
2. AP in b/g/n mode adapts to ´b´ mode since he has some stations in its network working on ´b´ mode.
3. Both AP and station are in ´n´ mode but both adapt to ´b´ because we have other stations working in ´b´.

Now how each of these mode would now refer to the power outputs of the card is also not completely clear to me. Maybe some of MT can shine a better light on it?


On the side: I don't understand why you use ´b´ stations and one ´n´ in your setup. Your whole network will not be better than the ´b´ protocol can give. If possible try to use ´n´ only on all units. (Or just ´g´) That makes your network more stable with farther reach and higher speeds. But ok, maybe you have your reasons...

Well to answer your question, about B clients, they are being replaced with N clients, but the G clients will remain. But my question still remains to much confusion.

How does a G client, ever see 25dbm output, when it only has one antenna???

My only theory is terminating the J5 port with a dummy load, and enabling chain 1, and hope the power output comes out the chain0 port.

Every other theory I have doesnt make sense.

You can use 2,3 chains in a/b/g modes also - for diversity, all N-cards allow that, and then you will get your full power.

You can use 2,3 chains in a/b/g modes also - for diversity, all N-cards allow that, and then you will get your full power.

But on one connector? That's not what I've been reading in several comments MT made. I've red several times that using only one chain degreases the total output by 3dBm (halve).

Now I'm getting confused too.

Ok, I have one AP which has a R52nM card.
This cards has obviously 2 chains, and 2 antenna connectors. Working mode is 802.11a

If I use one chain/connector only, the power level (output) readings are:

Rate Tx Power Real Tx Power Total Tx Power
6mb 18dBm 18dBm 18dBm


Now, if I enable both chains/connectors I get this:

Rate Tx Power Real Tx Power Total Tx Power
6mb 18dBm 18dBm 21dBm


The total power output is doubled. Each antenna gets 18dBm though. So, no matter if one or if both chains are enabled, the power per connector is never more than 18dBm.
Only if you use both chains you get the full combined power. In ´n´ mode this means the full power to one ´link´ if this would be PtP. But in a/b/g you get half. This makes sense to me.
Basically you have 2 amplifier circuits in the radio. Each connected to one connector. So each connector gets 18dBm and both summed together gives 21dBm.

Real life check:
Both of the chains are connected to a sector antenna serving a different sector.
Now, if I disable one chain the clients that are reached (PtMP environment) via this antenna drop off or get their signa through the other sector but with some 5 to 8dBm lower signals. (Rx AT the client)
But the clients connected to the antenna that is still working, their receive signal (from AP!) is not changed. So this chain is not getting more power now.

So I think it is right to conclude that each chain comes with its power which is always the same. Only in combining the two chains in one ´n´ protocol environment you have benefit of the combined power. Not in any of the legacy modes.

Anybody that has different arguments please tell them because this is very interesting stuff....

I see and understand that now. My concern is that on most traditional ptmp setups people have a dual polarity antenna sector and face it one way. Any B or G clients at any reasonable distance are never gonna see one of the polarities in enough signal to combine the received power to 25 dbm.

The only way I see it working is if the B or G client can see both chains with equal signal strength.

So I think it is right to conclude that each chain comes with its power which is always the same. Only in combining the two chains in one ´n´ protocol environment you have benefit of the combined power. Not in any of the legacy modes.

If this is the case, the datasheet is wrong for B and G power outputs.

Anybody that has different arguments please tell them because this is very interesting stuff....

Yes, in case of legacy a/b/g mode you will get 18dBm per chain, but if both chains are enabled and connected to (for example) dish antennas pointing the same direction, on opposite side you will get stronger signal, cause instead of just one 18dBm source you will get 2x 18dBm sources that are trying to send you the same information.

It is not MIMO, but you can use all features of:
http://en.wikipedia.org/wiki/Antenna_diversity
And end effect sometimes is much more than +3dBm

This also allow to shape your PTMP link coverage of single wireless interface - for example you can make setup where 20 clients from 5km range connect to cahin0 sector antenna, and other client connects from 30km away to chain1 point-to-point dish. and from software point of view they are connected to the same wireless interface.

Yes, in case of legacy a/b/g mode you will get 18dBm per chain, but if both chains are enabled and connected to (for example) dish antennas pointing the same direction, on opposite side you will get stronger signal, cause instead of just one 18dBm source you will get 2x 18dBm sources that are trying to send you the same information.

It is not MIMO, but you can use all features of:
http://en.wikipedia.org/wiki/Antenna_diversity
And end effect sometimes is much more than +3dBm

Very interesting! Nice alternative for the case mimo ain't working for you! (And probably getting better results on a 20Mhz wide band link where otherwise 40Mhz was needed on a ´n´ link just to get a very stable link to run with mediate throughput.)

This also allow to shape your PTMP link coverage of single wireless interface - for example you can make setup where 20 clients from 5km range connect to cahin0 sector antenna, and other client connects from 30km away to chain1 point-to-point dish. and from software point of view they are connected to the same wireless interface.

I am doing sort of the same. I use two sectors to server a wide area with two good antenna's while using only 1 frequency (first saving!) and one radio. (Second saving in power, Third saving in device=costs + lower temperature production in box)

But now it becomes interesting to know, from where the radiocard splits the two chains.
Physically for an antenna it makes no difference how many stations are communicating with it.

Now each AP has a limit of the maximum number of simultaneously associated clients it can serve which are also active in a sense the indeed do send and transmit data streams (so not associated but idle).

Lets assume that a single chain AP could serve 20 stations each demanding 2Mb in total so this 40Mb is the traffic stream that has to be handled by the AP while the AP also has to serve each client one by one (TDMA).
Lets assume that this 40Mb in itself is not the limit of the routerboard to process but actually to server 20 clients with it is. (Radio only has so many cycles per second to server 20 stations with this 40Mb.)

Now, if we would use 2 chains, and split these 20 stations equally over both chains, would that now mean we can have each client even use 4Mb without loss of performance of AP? Or we can now have again 20 stations on each chain (so 40 in total) all demanding 2Mb without loss of performance of AP?

The process of TDMA (´talking´ to each individual client in order) is that done on each chain separately or before the data is split?
If not (TDMA is a single process of radio CPU) than actually we won't be able to increase the amount of stations attached to an AP by using 2 chains.
If TDMA takes place on each chain individually (which I don't think is the case) than indeed we would be able to ´talk´ with more stations at the same time where then the rest of the AP circuit takes care of the data stream process. (Routing, Qeue, filter etc.)

I am afraid that using both chains in itself will not give us the possibility to have more stations associate with one AP.

if we are talking about perfect conditions, then no, you will not get any benefit when using both chains on non-802.11n situation. In real world where there are many variables and varying distance/signal quality over scope of one AP, you could benefit greatly, like having small omni for laptop clients and some sector to reach some further away clients as macgaiver described.

I understand the different ways people can use the card with different antennas, but I'm still confused on how one b/g client can ever see the advertised 24/25 dbm power output.

I understand the different ways people can use the card with different antennas, but I'm still confused on how one b/g client can ever see the advertised 24/25 dbm power output.

You don't get it physically on TX side, but at RX side signal level/quality from 2x21/22dBm >= 1x24/25dBm.

(From distance two smaller light bolts shine better together than one slightly bigger alone.

if we are talking about perfect conditions, then no, you will not get any benefit when using both chains on non-802.11n situation. In real world where there are many variables and varying distance/signal quality over scope of one AP, you could benefit greatly, like having small omni for laptop clients and some sector to reach some further away clients as macgaiver described.

Hence I'm using it that way. I have two 90º sectors now serving a 180º ´pie´ with a gain and reach (vertical pane is more open) I would reasonable not be able to get done with one 180º antenna.

My point is a bit theoretical. If the NV2 working AP serves two antenna's over its two chains, is the tdma process also split in two or not.
If it was you really would have two separate antenna chains with their own characteristics while rest (which rest?) of the data process done by radio is done in one circuit and in doing such gain some performance.
Since both chains basically have their own amplifying circuits I was wondering where actually the tdma proces takes place.
If that process is done in each of the chains we would truly have a dual fully synchronised antenna system.

If this last would be the case, and after a while 3 or 4 chain cards would evolve (it is covered in the 802.11n protocol) you could finally make your 3 or 4 sector antenna AP setup with full synch. something some are dying for in another topic!

But I am afraid the separation doesn't go that far. Only the amplifying circuits are dual, but all the extra processing is just one process?

I understand the different ways people can use the card with different antennas, but I'm still confused on how one b/g client can ever see the advertised 24/25 dbm power output.

You don't get it physically on TX side, but at RX side signal level/quality from 2x21/22dBm >= 1x24/25dBm.

(From distance two smaller light bolts shine better together than one slightly bigger alone.

Your reasoning might be right. But I don't think that the .pdf talks about what other side would receive. It talks on what the power is of the card and they do mention the high levels for the low rates in b and g. It doesn't look right to me.

Maybe some of MT can explain this?

I can test it myself and when I disable one chain the other chain is not increasing in power output. It stays the same which also makes sense since the 3dB increase comes from doubling the chain and thus the power which is 3dB.

The only way to get a double power output in b or g mode is using both chains in b or g (a) and have two antenna's connected to it serving the same stations at the same time. That station now might see a near 3dB increase in signal by having both chains working in the AP in b or g (and a!) mode.

I understand the different ways people can use the card with different antennas, but I'm still confused on how one b/g client can ever see the advertised 24/25 dbm power output.

The only way to get a double power output in b or g mode is using both chains in b or g (a) and have two antenna's connected to it serving the same stations at the same time. That station now might see a near 3dB increase in signal by having both chains working in the AP in b or g (and a!) mode.

1000% agree, but who uses these cards to stick 2 rubber ducky omnis and make a hotspot? I want these things to do 10 km PtMP Dual polarity sector setups where the legacy CPE will NEVER see the other polarity chain and never see the advertised power. Id like to hear from mikrotik too about it. So sure the advertised power might be right, but only in certain very limited usage scenarios with near perfect conditions.

1000% agree, but who uses these cards to stick 2 rubber ducky omnis and make a hotspot

Well, to be honest, if it wasn't that we have mimo nowadays, using both chains for one and the same frequency is not such a strange idea after all. Its called diversity and ´macgaiver` already mentioned that. If you follow his link you'll learn that cell phone operators use it that way sometimes.

I want these things to do 10 km PtMP Dual polarity sector setups where the legacy CPE will NEVER see the other polarity chain and never see the advertised power.

What is it you want now? Setting up a dual polarity sector where a CPE will only talk to one of the polarity antenna's/chain? That's well possible. Off course you only get the power of that chain. Not the power of both. But it even gets better. If you would now use a legacy mode working CPE but also with a dual chain card connected to duo pol. antenna you can again use both. I don't know how ROS would deal with it. But either you get double the receipt energy or ROS simply picks the strongest signal. In anyway it will improve your link and make it less vulnerable for one polarity interferences.

So sure the advertised power might be right, but only in certain very limited usage scenarios with near perfect conditions.

Isn't this the case in many things you buy? A Ferrari will only give you 300km's per hour if the situations are ideal. Like good road, skilled driver, plenty of length and no other traffic.

Most brochures/sales info show you what is possible under the best circumstances. MT is nothing better than any other product manufacturer.
You heard about 300Mb over Wifi mimo products? Yes, but how many people really see this happening?
You buy 10Mb adsl line? How many people do really get this?

etc. etc.

If you want more power for a link you can buy higher power output card. Like the R52H. And other makers even have higher powered cards. Just don't think raw power will do every job. Sometimes it ruins more than it creates...