Senao 200Mw cards

[Anonymous]

I would like to use a couple of these to make a longer distance, very inexpensive link.

Have any of you accomplished over 20 miles with these?

Can it be done by stepping down the speed rate from 11 to 5.5 Mbit?

Or lower?

The power and sensitivity combination make it perfect for low power and low cost distance, if they can be made to associate and pass packets reliably...

These would be the cards... MMCX antenna connectors, 200mw power, appear very much to be Senao...

http://www.ecost.com/ecost/shop/detail.asp?dpno=155861

Thanks

[georgew]

The senao 200mw cards should work fine at that distance.

The only downside is thruput. There is a delay factor that needs to be tuned past a distance of 6 miles to maintain maximum thruput. This setting is only available on the orinoco clones. Without the ability to adjust the timing, the link thruput will diminish with distance. Other than that, the cards should preform well for you. Put a 24db antenna at each end
and it should work fine.

[Anonymous]

I understand the timing thing...

What I don't understand, is how that timing factor is changed with the lower modulation rates.

Or, if those cards have already been tweaked...

[lonnie]

They are untweaked from the report we have had from others.

[georgew]

Because of the nature of the timing problem, switching to a lower speed is not going to help, the same problem exists at all speeds.

The timing problem causes a loss of available airtime, switching to a lower speed reduces the airtime even more, so it is counter-productive, unless it improves thruput by reducing retransmissions and errors.

You may have to go to a slower sped anyway, but no need to do it unless it actually helps.

[bobbyc]

Yeah, senao's impletation of 'high range' means long range indoors... not miles...
Bob C

[Anonymous]

Teletronics (www.teletronics.com) sells Senao cards for thire 200 MW connectorized cards. They claim 50km distances are possible.

I called tech support, emailed, and otherwised bugged them about getting the real truth on their cards, and they gave me rather vague answers.... except that they were very certain that the data rate had to be dropped.

Why it helps with the data rate being dropped, is that the time from the moment the packet is sent, to the time the ACK is seen by the sender, is so short that even at light speed, it only has time to go 21 miles

At lower data rates, the ACK time is much longer. Data rate doesn't matter, because the time starts at the END of transmission, not waiting for a whole packet to travel. At lower data rates, the ACK time allowed is longer, for some reason.

How much and all that, I do not know.

[lonnie]

I would say somebody does not know what they are talking about, but that would not be nice. So, I'll just keep quiet and ask you to give it a try and have you then go and ask them about your results.

[georgew]

I have no doubt yo can get really long links.

But the tech was just plain wrong about the slower speeds addressing this issue. The slower speeds improves sensitivity, and throws away half of your airtime at the same time.

Part of the issue with the tech support is the scope of their mission. The only people at Senao that could answer the question correctly are not going to be in the support department, and they probably don't speak english. The tech support people are there to help load drivers and read the user manual to people.

If you read their spec sheets on range, you will see what they mean by long range, and the implication is you must reduce speed to go farther, but that has nothing to do with the timing issue.

Up 23 dBm transmit output power
Long operating range, up to three times
range of standard products
11 Mbps –300m/450m ( 23 dBm output power)
5.5 Mbps –400m/600m ( 23 dBm output power)
2 Mbps – 500m/750m ( 23 dBm output power)
1 Mbps –800m/1200m ( 23 dBm output power)

The distance tuning issue is in the chipset's capabilities, and is outside of the capabilities of the card maker to change. The Prism chips simply don't have the needed tunability, when/if it is added, it will be a new chipset.

[Anonymous]

I would say somebody does not know what they are talking about, but that would not be nice. So, I'll just keep quiet and ask you to give it a try and have you then go and ask them about your results.

No, speak up, Lonnie.

I've now read conflicting information on this, and strangely enough, both were considered "experts".

However, after doing a LOT of searching, and finally stumbling onto what I was looking for, I believe that's incorrect now.

So, again, in theory, lowering the speed won't help.

Still, I'd like wrangle out of Teletronics why they claim 50km links are possible. I tried, believe me, I tried, and the people I talked to simply would not get into the discussion - I don't think they had any idea what they were talking about.

I also stumbled across something else, that hinted, but would not say, that some firmware may not strictly comply with the standard, and so will enable operation that otherwise isn't possible.

[Anonymous]

I have no doubt yo can get really long links.

Part of the issue with the tech support is the scope of their mission. The only people at Senao that could answer the question correctly are not going to be in the support department, and they probably don't speak english. The tech support people are there to help load drivers and read the user manual to people.

The distance tuning issue is in the chipset's capabilities, and is outside of the capabilities of the card maker to change. The Prism chips simply don't have the needed tunability, when/if it is added, it will be a new chipset.

Let me quote something interesting for you...

From the teletronics website

http://www.teletronics.com/tii/products/wireless_lan/LANc11_hp2.html

(in free space)
11 Mbps = 1380 Ft. (420 M)
5 Mbps = 1950 Ft. (600 M)
2 Mbps = 2550 Ft. (720 M)
1 Mbps = 3900 Ft. (1200 M)
50 Km with optional amplifier and antenna

So who's fibbing to whom?

There seems to be a whole lot of FUD about all of this..

So, is TT lying, or does someone out there tweak firmware?

(I hate questions like that...)

[georgew]

I have no doubt yo can get really long links.




(in free space)
11 Mbps = 1380 Ft. (420 M)
5 Mbps = 1950 Ft. (600 M)
2 Mbps = 2550 Ft. (720 M)
1 Mbps = 3900 Ft. (1200 M)
50 Km with optional amplifier and antenna

So who's fibbing to whom?


I don't think you understand.

Long links are possible, with any card, period.

We are not talking about what is possible. We are talking about what is well tuned and what is not.

Using the prism cards you CAN get really long links. On those links the performance characteristics of the links will display a predictable loss of thruput past 6 miles. The links are possible, and strong and good... and they will be slower than links at the 6 mile point.

No one is lying, you are assuming I am saying something I am not. I never ever said there was a brick wall at 6 miles.

The actual calibrated distance I believe is 10 kilometers, a nice round number. But regardless of the tuning, you can go farther. In fact people have been going really far without tuning for a long time. My first link was farther, and it worked well.

The technical issue is that side by side, an Orinoco through a 23db amp will have better thruput than a 23db prism based card. The only downside to this is if the lost thruput is important. If you are feeding Internet data off of a T1, the radio link is probably faster than the T1 still. On the other hand, if this is a lan to lan link, the extra speed may be important.

[tony]

During our testing with our 57-mile link, we tried intersil cards on each end (not the TT ones however) and it linked fine, with great signal however the throughput could not exceed 16KBytes/sec with extreme packet loss. The cards would only connect at 1Mbps.

Lucent cards that use the default timing acted in an identical manner, with exactly the same throughput numbers. With the tuning in place, the Lucent cards (still in service) link up at 11Mbps, with ~400+KBytes/sec throughput and no packet. We would love to get the timing information for the Intersil cards, however according to the Intersil firmware development team, there are no hooks or other means such as firmware patching that will allow us to adjust the timing.

[lonnie]

I would say somebody does not know what they are talking about, but that would not be nice. So, I'll just keep quiet and ask you to give it a try and have you then go and ask them about your results.

No, speak up, Lonnie.

I've now read conflicting information on this, and strangely enough, both were considered "experts".



I guess you have to decide which "expert" you have faith in, especially when they differ in advice. I won't speak on this topic any further.

[Anonymous]

I would say somebody does not know what they are talking about, but that would not be nice. So, I'll just keep quiet and ask you to give it a try and have you then go and ask them about your results.

No, speak up, Lonnie.

I've now read conflicting information on this, and strangely enough, both were considered "experts".



I guess you have to decide which "expert" you have faith in, especially when they differ in advice. I won't speak on this topic any further.

Well, let me put it this way... you guys have demonstrated that not ONLY do you know what you're talking about, but that you have witnesses that prove it.

So, when things conflict, I take your word above pretty much anyone else's...

Is that a problem? :)

[lonnie]

Thank You. Now get back to work.

[bobbyc]

I just threw a 200mw senao prism 2.5 card in my home starOS router desktop, associated to a hermes AP at 8.5 miles with distance setting at 17 miles.

RX from my house: 525KB/s over 15seconds.
TX from my house: 475KB/s over 15seconds.
Bob C

[bobbyc]

Is this incorrect?
http://www.personaltelco.net/index.cgi/MaximumDistance

[georgew]

Is this incorrect?
http://www.personaltelco.net/index.cgi/MaximumDistance

I know Jim, and when he is not a troll on the attack, he is an experienced microwave engineer who should know what he is talking about.

It would seem that the prism chipset radios have a hard limit on distance, and the limit is extended for each speed increment. At 11 megabits, they run full speed up to 10 km, then the speed drops to zero at 35km. So there is a brick wall after all.

I'm suprised that the slower speeds don't have the same problem in the same place, that tells me that the slower speeds must have slightly less available thruput than one might assume comparing the bit rates.

[Anonymous]

When I have some parts to try it out with...

I'm going to try locking a couple of these cards at particular speeds and seeing what they do at various distances.

I have a situation where I can temporarily set up a link and have the opportunity to test it at around 5, 10, 16, 22, and 28 miles with good LOS and all that jazz :)

[bobbyc]

Well I'll have some interesting results by the end of the day.
Right now I have a hermes AP card at tower A, and a hermes client card at tower B. Signal is -77/-77 Noise -98/-98. Distance is 13 miles. From the AP's perspective, the new tower C is ~15 degrees to the right of tower B... and is about 15 miles. I am going to throw 200mw cards in tower B and tower C today, and do a thruput test between tower B and tower A. I am then going to drive to tower A and rotate the grid antenna to split the diffence between tower B and tower C.
Confused? If all goes well and the beamwidth is wide enough on the 24dB grid of tower A, Then I'll get away with not using amps.
I'm hoping that since the AP is going to remain hermes, I'll be able to set the distance to 15 miles, and tower B and C can get decent downloads. Uploads will hopefully be at least 2mbps, but we'll see how prism timing effects that.
Bob C

[bobbyc]

Oops; I just did the topo map and tower C is 20 miles from tower A. Prism chipset is out of the question. Going to have to use a amp/hermes at tower C.
Bob C

[dkii]

I've got a couple of links that are > 15 miles. If I use anything but a hermes card, or don't set the distance setting correct, the RX speed is fine a little lower, but not bad, since the setting is correct on my AP and it it using a hermes. I've tried it with a prism, and a TI card, and a smartbridge(which I believe has a prism chipset) and all 3 of them would not transmit at over 20KB/s at that distance. Smartbridges claims 21 miles.

[bobbyc]

Well yesterday I got to the 20 mile away tower and hooked up the starOS router, and threw a 200mw card in... and got a signal of -96 to tower A... but I hadn't gotten to tower A and rotated the antenna to split the difference between tower B and C.
So I played it safe and climbed tower C and rotated the antenna towards tower B, and am feeding it via a sector AP from tower B for now. Initially I got -76 signal because tower C is out of tower B's beamwidth; but I then drove to tower B and rotated the sector towards C; and I didn't harm any of my clients' signals off of tower B. Now tower C has -69 signal and is good enough for now.
Bob C