MikroTik

Just announced to distributors and available, the new R52Hn looks like a winner. Mikrotik took the best of the R52H, the R52n, and the R5H to create the complete radio card. It supports both 2.4GHz and 5GHz, 802.11a/b/g/n, with up to 320mW transmit power, up to -97 dB receive sensitivity, 2x2 MIMO with dual MMCX connectors, and improved ESD protection, all in the same price range as an R52H.

I can't wait to get my hands on these and test them in my network. If it performs as well as it looks in the specs, I'll be using this card everywhere.

The complete data sheet:

http://www.roc-noc.com/pdf/R52Hn/R52Hn.pdf

Tom

MMCX are THT?

yes, it also has ESD protection on the ports, which is not mentioned on your datasheed. We will fix the PDF.

back side of the card, for the ones interested in the connectors and heatsink

Nice, I want!!!

Exciting.

I hope it doesn't go up in smoke when changing channels, like the R5H.
I also hope it won't suffer from 2.4 GHz interference when running at 5.8 GHz because of the dual band operation.

it's on the web now: http://www.routerboard.com/pricelist.php?showProduct=94

Nice.. can't wait it will be available at balticnetworks.

Весчь!!!

MMCX, Brilliant!

Thanks Mikrotik

wow... can't wait...

and i hope it has good quality better than R5H, 3 of 5 of my R5H just blown and still on RMA.. already wait for 2 months and still no coming back..

Today I set up 2 RB800 routers each with a new R52Hn radio card and dual omni antennas. I only had about 30 minutes available to set up and test them. I used the vpls bridge. I was running the bandwidth test program on the RB800s which is not ideal. I'll try to get more time tomorrow and run the test program on other RouterBoards.

Freq was 5180. 802.11a mode worked fine as expected and I had about 20.5Mbps throughput with tcp both directions. CPU was at 11%

Next I turned on 802.11n mode with 2 chains and saw the throughput jump to about 50Mbps tcp in both directions. CPU was at 100%.

Wow, that is impressive. Guess I'll have to deploy some 802.11n in the field.

It also shows that you need a fast CPU for optimal 802.11n. I plan on using the RB411AH for all my outdoor PtP 802.11n links until an RB411G comes out.

Tom

I was running the bandwidth test program on the RB800s which is not ideal.
[...]
It also shows that you need a fast CPU for optimal 802.11n...
Tom

...Maybe it just really shows that one should not test with bandwidth test program running on the device itself

I have a question: Based on what you said in another thread, I take a guess that R52Hn will not fit in RB411U with indoor enclosure?

Will 3 of these cards fit in a rb433ah? Wonder if Mikrotik will ever modify boards to accept major brands of high powered radios since they are thicker? Shore gets old having to put two boards with 3 radio slots each, just to get to use 3 high powered cards not to mention breaking my piggy bank and 532A's with daughterboards are running low at my place.

50+ RB's and counting

but in the end... a hellava lot better than all the other alternatives.

MMCX mikrotik radio!

Heya I've got a question: Is the reported 25dbm achieved in single chain mode or this is sum of both chains(meaning that single chain setup will have 3dbm less)?

that's combined power level of two chains

Okay, thanks for quick answer!

Anyone else test these yet to see how well they perform? Do they outperform the R52n? I've had no luck getting anything greater than 30 Mbps UDP on 5 GHz or 2 GHz. I've tried several different people's configs who say they get close to 100, but not me!

for full 11n performance use both chains, you should get more then (in perfect conditions up to 200Mbit should be reachable)

Using 1 chain I only get around 7 Mbps. I'm testing indoors at close range. Do I still need to keep the antennas a certain distance apart and use different polarizations?

Using 1 chain I only get around 7 Mbps. I'm testing indoors at close range. Do I still need to keep the antennas a certain distance apart and use different polarizations?

What RouterBoards?

What frequency?

Dual antennas?

CCQ?

I tested a pair last weekend in RB800s on our office conference table. I was seeing about 100 Mbps in single chain and 200 Mbps with two chains. 5180, 5GHz-N-only, two 5 dBi omnis in each - both straight up (vertical), about 10 feet apart. My CCQ was in the 90-100% range.

Tom

Using 1 chain I only get around 7 Mbps. I'm testing indoors at close range. Do I still need to keep the antennas a certain distance apart and use different polarizations?

What RouterBoards?

What frequency?

Dual antennas?

CCQ?

I tested a pair last weekend in RB800s on our office conference table. I was seeing about 100 Mbps in single chain and 200 Mbps with two chains. 5180, 5GHz-N-only, two 5 dBi omnis in each - both straight up (vertical), about 10 feet apart. My CCQ was in the 90-100% range.

Tom

I only wish mine worked half that good... I'm using 1 RB600 and 1 RB133. I've tried 2 GHz and 5 GHz and the performance is roughly the same. 5 GHz is a little better, I assume because of a cleaner channel, but not by much. Max I ever got UDP was 34 Mbps with both chains. The CPU usage only got up to around 20% I think. I'm using 2 backfire antennas on one routerboard and 2 omni's on the other. I've also tried connecting one of the R52n's to an AP on the linksys WRT610N and throughput was way worse than just connecting a computer to it with a wireless N adapter... I don't know if I have bad cards or what. I've also tried 2 5 GHz turbo cards and they perform better.

I only wish mine worked half that good... I'm using 1 RB600 and 1 RB133. I've tried 2 GHz and 5 GHz and the performance is roughly the same. 5 GHz is a little better, I assume because of a cleaner channel, but not by much. Max I ever got UDP was 34 Mbps with both chains. The CPU usage only got up to around 20% I think. I'm using 2 backfire antennas on one routerboard and 2 omni's on the other. I've also tried connecting one of the R52n's to an AP on the linksys WRT610N and throughput was way worse than just connecting a computer to it with a wireless N adapter... I don't know if I have bad cards or what. I've also tried 2 5 GHz turbo cards and they perform better.

You are using a RB133?! That's your bottleneck.

RB133 has a 175MHz processor that will be at 100% all time and it cannot use the card.

RB133 has a 175MHz processor that will be at 100% all time and it cannot use the card.

I checked and the CPU was definitely below 50%! Anyway, I'm pretty sure I tested with 2 RB600's when we had them and it wasn't much better. I'll probably just try a few more things and then end up trying the new R52Hn I guess. Thanks for the help.

According to this: http://www.routerboard.com/pdf/routerbo ... _tests.pdf

RB133 will do 50-110Mbps with 1500 packets. Therefore as long as default packet size of 1500 was used it would not be the bottleneck.

Note that passing ethernet traffic trough the board and passing the wireless traffic is different thing, also note that N links are using frame aggreation for gaining more speed which uses CPU power.

RB133 wont pass more than 15mbps on wireless.

RB133 wont pass more than 15mbps on wireless.

And that is if you are lucky using 802.11a/b/g. Forget about 802.11n with 175MHz CPU. It will fall to its knees and crawl.

Tom

How does it perform with 2 chains on the AP and one on the CPE?

I am seriously considering putting up an 802.11n AP, but I still haven't found a reasonably priced dual antenna setup for the CPE.

Did anyone else notice, doing the math these things require max 2.1 A. That's way higher than any of the other Mikrotik cards.

Did anyone else notice, doing the math these things require max 2.1 A. That's way higher than any of the other Mikrotik cards.

How do you figure?

Did anyone else notice, doing the math these things require max 2.1 A. That's way higher than any of the other Mikrotik cards.

How do you figure?

P=IE so I=P/E 7watts/3.3volts = 2.21amps

I am guessing that the 7watts number has to be wrong as no other high power cards draw near that much power.

It wouldn't be the first time there was an error on a data sheet.

Tom

Nope, it is fine. There are two powerful PA's on it, and if you transmit with both chains (!) at max power, then it could be close to that number. Of course in real life, it probably won't be as high.

Did anyone else notice, doing the math these things require max 2.1 A. That's way higher than any of the other Mikrotik cards.

How do you figure?

P=IE so I=P/E 7watts/3.3volts = 2.21amps

I didn't notice the 7 watts line.

So, you'll really have to be careful about what boards you put these in by making sure it can output enough power to the cards.

How does it perform with 2 chains on the AP and one on the CPE?

I am seriously considering putting up an 802.11n AP, but I still haven't found a reasonably priced dual antenna setup for the CPE.

Me too. I am only waiting for a MT/3rd party solution that make a nice and need solution with two antenna's embedded in one box that can carry the typical router boards ready for the ´n´cards.
When will we see such CPE solution? This would be the perfect way to keep these Wimax guys out of my back garden....

UBNT is expanding its leap in this field.....
Hurry up MT!

How does it perform with 2 chains on the AP and one on the CPE?

I am seriously considering putting up an 802.11n AP, but I still haven't found a reasonably priced dual antenna setup for the CPE.

Me too. I am only waiting for a MT/3rd party solution that make a nice and need solution with two antenna's embedded in one box that can carry the typical router boards ready for the ´n´cards.
When will we see such CPE solution? This would be the perfect way to keep these Wimax guys out of my back garden....

UBNT is expanding its leap in this field.....
Hurry up MT!

What price point do you guys consider reasonable for a CPE?

Scott

does R52hn support channel width of 5,10,20,40 Mhz?

Hello,

just RB800+RB604 + 3x R52Hn in test (exchange from UB5). First look? Cards not recognized. After soft reboot - ok, but at booting time there are about 16-17 beeps. Power outage = cards not recognized. Again soft reboot - OK. Anyone have this issue?

But cards are MUCH BETTER than UB5. I will exchange R5H to R52Hn to compare this pair

mjj,

Yes I have the same problem, my rb600 wont recognize them. I have a tech on a tower right now trying to get it working.

Anyone have a suggestion?

What version of RouterOS are you running the cards under?

It was 3.2 now its 3.30. Do i Need 4.x??

It was 3.2 now its 3.30. Do i Need 4.x??

Yes

Looking at change logs looks like i need it.

can I do this upgrade remotely?

NVM, just following instructions.

how come you have only 2 antenna connector when you need 2 antenna for 2.4Ghz and 2 antenna for 5Ghz ?

how come you have only 2 antenna connector when you need 2 antenna for 2.4Ghz and 2 antenna for 5Ghz ?

you just drop some acid and beuty of this card just vanish ) =))

this card use conector as you want ... one or two antenas, or 2 antenas ... one for tx and other for rx ... what ever you can imagine ... just read more about 802.11n and you will be in knowledge

yeah but for MIMO in .11n you need 2 antenna of each frequency ... => total of 4

the pdf says it 2x2 MIMO so ...

AcidDrop,

You need one antenna for each MIMO. So you need two antennas for 2xMIMO.

The cards supports multiple frequencies but _not_ at the same time.

You can use it for 2.4GHz with 2.4GHz antennas _or_ 5GHz with 5GHz antennas.

Tom

I was running the bandwidth test program on the RB800s which is not ideal.
[...]
It also shows that you need a fast CPU for optimal 802.11n...
Tom

...Maybe it just really shows that one should not test with bandwidth test program running on the device itself

I have a question: Based on what you said in another thread, I take a guess that R52Hn will not fit in RB411U with indoor enclosure?

Yes that is correct. You cannot fit a radio card with MMCX connectors in the RB411U or RB411UAH and use the Mikrotik indoor case. The miniPCI card slot is just too close to the case top. The same holds true with the top card slot in RB433 and RB493 boards. They are too close to the case top so you can't use MMCX cards in the top slot of those boards and close the indoor case.

Tom

AcidDrop,

You need one antenna for each MIMO. So you need two antennas for 2xMIMO.

The cards supports multiple frequencies but _not_ at the same time.

You can use it for 2.4GHz with 2.4GHz antennas _or_ 5GHz with 5GHz antennas.

Tom

I understand better, thanks.
I'm a bit disapointed though

can someone give me some technical data about signal noice ratio? I found no document or chart regarding it.
I figured out on 3 different r52hn cards operated either with antennas or shuts that noise floor ist very high. I were operating them on 433ah and 433 plain.
i get values between -65 to -80dbm, which is far beyond of such as old fasioned r5h, dmca82 or xr5 (all of them around -95 to -105 dbm).
pse let me know if someone else experianced such behaviour too.
It's pretty unbelievable that the new card will behave worse than the old ones.

thanks in advance, Robert

you talk about PtP or PtMP?

point to point (PtP). But Measurements were done under non link condition to get rx values

did update to 4.10 with board fw update and somehow the ground floor level dropped to -115dbm!
such value will be fantastic.

i give the same procedure to another board for being sure and will update the posting

just finished verification. must be a rb433ah 4.9 os thing. after already full update (os and fw) all is running fine (-115dbm 5g und -112 dbm at 2.4 g band).
Congrats, fine work!

Hello,
after several weeks of testing R52Hn noticed the following things:
1) after changing from R52: improve the sensitivity,
2) after changing from R5H: similar sensitivity (unchanged)
3) what's interesting: customers with the signal at level of -75 or worse, do not hold online (using a WinBox = disconnects).

Has anyone noticed the effect of instability in links with weak signals?

change HW retries to bigger value

HW = 10. I'll check this out (HW = 8 before).

You can compare, same site just change R5H to R52Hn on AP side. FYI noise floor before -97dBm and noise floor after -110dBm. As you can see signals drop sometime around 5-7dBm.
AP ROS4.9, most of CPE ROS4.5

before - R5H (5GHz-a mode) after same time for signal settling - R5Hn (5GHz-a/n mode) signail drop is pretty high, don't you think ?

regards

Tomas

I have replaced many back bone links with the R52Hn. So far we had 8 failures. It is all on the tx side. Is there anybody else that also had failoures?

I've noticed one fundamental thing right now: if you don't need 'N' R5H is better solution than R5Hn. We have right now many working links at R5H on both side. Antennas have VSWR below 1:1,14 & RX/TX are not less 58,5 Mbps, distances: 0,8 - 4 km.

I've noticed one fundamental thing right now: if you don't need 'N' R5H is better solution than R5Hn. We have right now many working links at R5H on both side. Antennas have VSWR below 1:1,14 & RX/TX are not less 58,5 Mbps, distances: 0,8 - 4 km.

How you can achieve tx/rx not less than 58,5 Mbps ?? R5H is only 802.11a, so maximum is 54Mbps right? But I need to confirm, that where we change from R5H to R52Hn, there we have more struggle with hwretries, troughput, latency or CCQ. I guess, problems are more with "drivers" in ROS 4.9 than with hardware. So we kindly waiting for new ROS versions. We will continue on testing

How you can achieve tx/rx not less than 58,5 Mbps ?? R5H is only 802.11a, so maximum is 54Mbps right? But I need to confirm, that where we change from R5H to R52Hn, there we have more struggle with hwretries, troughput, latency or CCQ. I guess, problems are more with "drivers" in ROS 4.9 than with hardware. So we kindly waiting for new ROS versions. We will continue on testing

That's simple with R5H, because it supports 5GHz-turbo mode. I almost everytime can see 'P throughput' at around 70mbps and it actually reaches real-life values somewhere near that.

As for the R52Hn card itself, so far I cannot tell anything good about it, since I had no success in any of its supported modes (a, n). The performance is worse than R5H (however, I will try to increase hwretries as mentioned in this forum). Another odd thing I've noticed is that when doing a scan, I can see my AP with signals around -53dBm, which of course would be very good, but unfortunately I'm sure that such signal level is impossible in location that card is set up. And what's more, when it connects to AP, it shows ~-75. I would definitely credit this as some sort of a fault. More like a software one, I believe. Also I had no luck in adopting these cards with MIMO mode either, using 2 sperate antennas (with different polarization and spaced around 40cm between. What better antenna setup could there possibly be?). Which forced me to move back to my old 11a-turbo setup for now.

Anyway, I'm not giving up, in not very far future I'm going to give another shot with these r52hn cards using all the experience from 802.11n topic in this forum