Tranzeo TR-6019f initial semi-review

John Schmidt hath wroth:

Blush. I don't know if I can handle the praise. It doesn't happen very often.

You got documentation? All I get are a packing slip and a manifest.

Yeah, that's the problem with bridges. You can't just "move" users from one switched path to another as easily as you can move them in a router. Spanning tree protocol is suppose to be self-healing when a path is lost, I've found that a limited number of bridges and switches actually work. I don't think the Tranzeo bridges support spanning tree protocol. So, when you pull the plug, the users conections are reset rather than disconnected.

Suggestion: Do it often enough and the users will get used to being disconnected.

Bah-humbug. That takes all the fun out of IT. If they lose work, tell them it's their applications vendor that's really at fault in that it should be able to tolerate a connection reset without missing a beat.

It's in the data sheet. See:

is about 15 degrees both horizontal and vertical. You can be withing 7.5 degrees off and it will still work. However, there is no guarantee that the mechanical boresight is also the peak in the antenna pattern. Boresight alignment has its limitations with high gain antennas. It's also a bad idea to be right at the bitter edge of the 3dB beamwidth. The antenna pattern falls off rather rapidly beyond the -3dB points. Therefore, make an effort to swing both antennas back and forth, both vertical and horizontal, and try to find the middle of the peak. My guess is anything withing +/- 3degrees will work just fine.

The big danger is interference from other systems. The high gain and narrow beamwidth antennas will be a big help, but you should still watch out for sudden drops in speed or link disconnections.

Congrats and good luck.

"Ian" hath wroth:

Running the numbers at 11Mbits/sec, I get a fade margin of 36.3dB. This more an adequate for a 1.5 mile link.

trees are always an issue. Foliage attenuation varies radically depending on the size of the leaves, degree of moisture, and number of solid branches. How "sparce" are the trees?

Oh-oh. Interference, moving trees, or reflections. Did you test the units together at much shorter range to see if there are any problems? I usually do that and then cover one antenna with a wet towel to simulate a more distant connection.

It's difficult to predict foliage attenuation. Worse, it changes by season and weather. During the winter, it's usually fairly mild as the leaves have all fallen from the trees. During spring, when the leaves come back, it's often fatal. I have one short (3000ft) link that is super strong in the winter, and a marginal proposition during the spring and summer. I can't move the antennas so I have the trees trimmed. Fortunately, the tree has grown considerably in the last few years and I can now shoot the beam under the foliage canopy.

It will only change speed if it sees errors. However, those errors can be caused by relfections, multipath, moving trees, Fresnel zone diffraction, and interference. I can't tell from here.

That's what interference from another 802.11 system looks like. The link disconnects and then tries to reconnect. Meanwhile, the signal strength looks like a light show because it's trying to find a suitable speed.

Yep. Also check for local interference with a spectrum analyzer or Kismet passive sniffer (because it can also see clients and blank SSID's).

That's usually a big help with interference. It doesn't do much for reflections, foliage attentuation, or Fresnel zone problems.

Argh. Methinks the WRT54G v5 is a loser. Be sure that the access point and WRT54G are on widely seperated channels.

Any potential 2.4GHz RF sources in between or in line with the antennas? Factories using 2.4Ghz microwave drying ovens are always a problem.

Tough question. It is easily possible to have a very strong signal and a lousy connection due to multipath and reflections. However, that can usually be temporarily eliminated by simply moving the antenna around somewhat. Eventually, you'll find some location that is free of reflections (for the moment). However, this mess sounds like continuous lousy connections which reeks of interference, not reflections or trees.

"Ian" hath wroth:

Tranzeo Monitoring System

haven't had time to try it.

I would like to hear more. I just set up a TR-6019f access point at one end and a TR-CPQ-19f customer premise at the other end of a 1.5 mile link. I chose these units because at the one end we have to shoot through sparse trees. Set up and install was easy. I don't need to send much data back and forth it is just to share a high speed internet connection in a rural area. I had trouble right away. The units were reporting -57 signal strength and link quality of 45 which is excellent. So I went ahead and mounted the units. The problem was that the internet speed at the CPQ end was down to dial up speeds. Watching the CPE performance screen I could see the link speed was constantly changing all they way from 1 to 11. I tried changing channels, no difference. Locking the units a 2 Mbit/sec improved the speeds to barely acceptable levels.

I got prompt response from Tranzeo support both via email and phone. About all they could say was upgrade the firmware(there was just a new release) and I must have a fresnell zone problem. Yes I have a fresnell zone problem, sparse trees at one end but had hoped these radios would power through the line of site issue. The signal strength and signal quality would seem to indicate that they would. Well I upgraded the firmware on both units. With the new firmware the units still reported the signal strength at -57 and they removed the signal quality reporting feature in this release of the firmware, so no numbers available. But still the same problem, speeds constantly changing and very slow connection when set to auto, and had to lock at 2Mbit.

The new firmware brought on another issue. The CPQ has 5 signal strenght lights to aid in pointing. Before the firmware upgrade the signal lights were on steady, after the upgrade the every 30 seconds or so the the radio light and the signal lights go out, flash a couple of times and come back on. Tranzeo has no suggestions for fixing this. I guess I will have to take down the CPQ unit and bring it closer to the access point and see if the problem is fresnell zone issue. If they work properly at close range I will try switching to horizontal polarity.

Here are the details of my set up in case anyone has advice to offer.

TR-6019f access point connected to a linksys WRT54G V5 router. Mounted 20 feet AGL TR-CPQ-19f customer premise equipment connected to a computer. Mounted 10 feet AGL. Access point elevation approx 80 feet higher than customer premise. Clear line of sight except for sparse trees 75 feet from the CPQ Signal strength -57 Signal quality was reporting 45 Latest firmware on both units Both mounted on wood structures

A quick question for anyone, if signal strength and quality are excellent could the trees cause this problem or should I suspect radio problems? For sure the issue with the signal and radio lights going out seems tied to the firmware.

Ian

Just to add to my previous post. When trying to troubleshoot the link speed I set the computer at the access point end to ping the CPQ 1000 times. The results were one ping at 9 ms the rest at 2 ms and 0 packet loss.

Ian

"Ian" hath wroth:

Ok. So all you have is edge diffraction from the Fresnel Zone problems. At 1.5 miles, the Fresnel Zone is 29ft. Do you have 29 ft of clearance at midpoint including not hitting the ground? I don't think so if you're 10 and 20 ft off the ground as you indicated.

clearance at midpoint would be nice, but I would settle for perhaps 22ft. I should point out that objects in the Fresnel Zone do not just cause attenuation by blocking part of the signal. They create edge diffraction, which causes the signal level to vary in peaks and nulls at the endpoints. Move a few inches one way or the other, and the signal could easily be very strong, or very weak.

Tranzeo also transmits at +23dBm while most Linksys radios deliver about +15dBm. Linksys also has the coax cable and connector losses, which Tranzeo does not have with an integrated antenna. By all reason and calculations, it should work.

That's too low to have much of an effect.

Not Aironet but probably WaveRider. They use 900MHz because it goes through the trees. It won't cause problems unless the ISP is using

2.4GHz as a backhaul from his central access point. Look for the characteristic barbeque grill dish antennas.

A few years ago, we were having a local interference problem. None of the 802.11 sniffers would show anything. I dragged a spectrum analyzer and big dish around in the back of my pickup for most of day until we found the culprit. It was a power company remote telemetry link (SCADA) from a mountain top power switching station. For some reason, copper data lines were not available so they used 2.4GHz Proxim/Glenayre Lynx radios. These boxes transmit continuously even though there is little or no data to send. They also occupy half the

2.4Ghz band in each direction (full duplex). It was difficult to find because they had used solid dish antennas with every few side lobes. I had to be practically under the antenna (on a 40ft telephone pole) before the spectrum analyzer would see anything. PG&E was eventually convinced to run some copper and replace the data link.

It's difficult to simulate field conditions, but not impossible. I like to set it up at close range and see what breaks. However, be careful going too close. The 802.11 timing relys somewhat on the speed-o-light time delay between endpoints. You'll need to put some minimum distance between endpoints to do a proper test. I think about

3 ft is the minimum. Anything you can put between the antennas will suffice as an attenuator. However, try to use absorbant material rather than reflective. The reason I use a wet towel is that it's easy to throw over a panel or dish antenna. Ideally, I should be using black carbon foam, which is the same stuff used in RF anechoic chambers.

Can you put the antenna in the trees and run a very long CAT5 cable? I'm not sure how long a PoE CAT5 run can be with these units. The data part of the CAT5 can be much longer than 300ft. I've done almost

1000ft without problems (using 10baseT-HDX). You can always run the power over larger diameter copper cable such as Romex instead of CAT5.

Worth a try. Just remember that you're changing 3 things at once. The Fresnel Zone, the folliage attenuation, and possibly the interference pickup.

Prism2 (not sure) so it should work just fine.

External antenna pointed in the direction of the line of sight. You want to pickup as much as possible. You can temporarily ignore anything to the sides. Start at one end and point the antenna along the line of sight in both directions. Repeat at the other end. If that doesn't show anything interesting, find a high place nearby and slowly spin the antenna 360 degrees. Anything that causes that much trouble should show up with Kismet as a fairly strong signal. So, if it really is 802.11 intereference, you should have no trouble finding it. Don't be suprised if it's a client radio and not an access point or repeater.

I've never seen a WRT54GL. It's my understand that it's a WRT54G v4 with a $20 higher price tag because Linksys had to re-negotiate the contract with the vendor after the v5 problems.

Don't just consider the houses in between. Also anything that's along the line of sight including beyond the end points. The signal doesn't just magically stop when it hits the other end of the link.

Good luck.

However, trees are always an issue. Foliage attenuation varies

I live in Canada. Right now there is no vegetation on the trees, you can see the other end through the trees. There is one row of trees and once the vegetation comes back I might not be able to see through them. I ran a test with linksys stuff in the summer attached to 15dbi antennas and had a connection, but of course wanted as much margin as possible so thats why I'm using the Tranzeo 19 dbi units.

I have checked it on calm days and with no vegetation on the trees I know it is not a moving tree issue. As far as interference goes we are in a fairly sparse area but a mile away I do have a 802.11 setup. The tranzeo CPQ picks it up as a signal 0f -94. There is alot of oil field sacada equip around running on 400 mhz I think and the ISP here runs aironet equipment on 900 Mhz. I did bench test them but it was more to get familiar with the setup screens and to configure and test the encryption before installation. To be honest I do not know if the units were hunting for a connection speed. I like the wet towel idea.

At the end with the trees the guy said we can cut down some brances in the summer if we have to.

I think I'll try moving the CPQ closer, they are close to where I live the CPQ is easily accessable and it won't take long to do. Horizontal polarity wouldn't take very long as well. I'll have to check if the wireless card I have will work with Kismet - SMC2532W-B. I purchased it because it is 200mw and can connect to an external antenna. When checking for local interference is it best to attach an external antenna and do a sweep or just use the card as is?

Yeah I got them the V5 before I knew of the issues with it because I have a ton of V4's out there working flawlessly. Hopefully the WRT54GL will be better. Anyone have any comments on it? At any rate it is easy to take it out of the mix and try transfering files between PC's.

We are in a very sparsely populated area aprox 1 house every mile or so but there is one house in a direct line in between at an elevation that should be below fresnell problems but maybe I'm wrong. I don't think they are running any equipment but you never know.

If it is interference hopefully horizontal polarity will be the ticket.

Thanks for you help Jeff, much appreciated. I'm not in a panic, the connection seems to be hanging in there fast enough for the internet, but I'll post back as soon as I get a chance to investigate further.

Ian

It looks like at mid point I have 19 ft. so could be a factor. I could go higher than 10 feet at the CPQ end. The client wanted it installed at that location. In the next county over from us an ISP uses Tranzeo radios and I have seen lots installs with fresnel zone infractions working fine at distances of 10 miles plus. I guess I just got lazy thinking if I use the same radios at a much shorter distance it is bound to work. Lesson learned.

Ooops not sure where I came up with Aironet, they are actually Alvarion BreezeACCESS 900 units.

I asked Tranzeo support and they said the have tested the PoE CAT5 to 300 feet. That would be long enough to get past the trees.

Good point I guess it would only prove that there is nothing wrong with radios.

Thanks for the advice.

I hope they get it sorted out I've become quite comfortable with the WRT54G routers.

Again thanks for all your help. I 've learned some good lessons from this little project. I had previously setup a couple of one mile links, one using two WAP54G's in point to point mode and one with a WRT54G and WET54G bridge and was getting a little cocky.

Ian