Wireless Service Providers → [Equipment] TDMA noise

Unfortunately this is a common misconception in WISP community, most probably due to robustness of Canopy FSK modulation in bad RF environments.

Complex modulation [TDMA] systems work best in good RF environments. Standard 802.11 protocol with CSMA work better "by design" in poor RF environments.

That is a necessary compromise in all [TDMA] systems which gives us more system capacity and other benefits.

[mod note: OT subthread - moved to its own topic. Title edited to better reflect content.]

Anyone have ideas for a more robust approach?

Motorola is like Nokia, they set the industry standards.

If there was something better than Canopy FSK/TDMA with GPS Sync, Motorola would have built it.

But that doesn't answer my question.

CW is a more reliable approach. On a channel 1 Khz wide I can talk to the UK from Hawaii with 23dBm transmit power at 15 meters on an HW-8 radio.

I can work that signal through QRM that is within 3 dB of my RSSI from the UK signal and communicate reliably.

That is vastly more reliable than Canopy FSK.

Now your question is answered.

Not really, but I don't speak HAM unfortunatly

All I really said is that the simpler the modulation scheme the more reliable the communications stream will be.

So if you communicate with morse code on an HF channel using CW it will require very little power, have great range, and extremely high reliability.

That is *why* Motorola Canopy BFSK/TDMA stands head and shoulders above any other system when it comes to reliability in difficult RF environments. BFSK is a very simple modulation scheme compared to even what 802.11b uses. That means that data rates are constrained in order to obtain range and reliability.

There is never a substitute for the KISS principle. You can modulate your signal with a quadrillion bits per second and all it will do is reduce your range and reliability. Good engineering will ensure that you get a good balance of reliability and signaling rate.

I do hope that answers WHTs question.

Do it did not, and you know that. Please try to stay focused on WISP deployments.Canopy and SkyPilot both have limited throughput. So the question should be, what is more robust for two different throughput venues?

This point is so important for WISPs but unfortunately starts so many debates.

I originally wanted to clarify that switching to (any) TDMA protocol will not improve performance in noisy RF enviroment after others wondered about that aspect:

"Whenever you have low signals or CCQ s, all the CPEs connected to the AP gets affected in a very bad way ... not the behavior you would expect from a TDMA system ... Switch back to normal 802.11n in the similar environment, everything runs smoothly"

Perhaps we should stop discussing merits of various modulation schemes and let people discuss merits of the "Airmax" TDMA variation.

Now you want to ask me to stay focused upon "WISP deployements" instead of answering the question that you asked which was and I paraphrase, "Does anyone have any ideas for a more robust approach to delivering broadband than Motorola Canopy BSFK/TDMA?"

Simple modulation schemes are more reliable than complex modulation schemes. FSK (Frequency Shift Keying), and BFSK (Bi-level Frequency Shift Keying) are more simple than almost any phase shift keying modulation.

The easier it is to detect a change on the receive side, the more information is preserved in any discrete transmission.

Motorola has chosen a very simple method of modulation. BFSK. They have done that because it is reliable. They then leverage that reliability on top of a TDMA system.

That is why Motorola Canopy knocks the socks of of systems like Ubnt "AirMax".... well actually it is only the most basic reason.

You think? I actually like Ubiquiti. They have some pretty good marketing folks. Their packaging ideas are good (if not original) and I like them.

If we actually discuss the merits of the "AirMax" TDMA crap, I am afraid I will loose friends at UBNT because they don't do Ap to Ap synchronization and you and I understand what that means. Sadly if we try to explain it here we are going to spend six to ten hours arguing with people "who have done it so piss on the scientists".

.. but you know, we may just have to do that ... We may have to debunk Ubiquiti because there is a whole lot of stupid crap floating around with respect to that vendor these days ...

Hmmm..... While Airmax isn't Canopy, it also isn't the same price. If the Airmax stuff allows me to have even 1/2 as many users on its AP's that Canopy does, I will laugh all the way to the bank.

On the other hand, it may prove that the extra $$ for Canopy may damn well be worth it. I will let you know in a few days, weeks, whatever?

I personally would LOVE to have an adult conversation about the differences of the various modulation schemes and why they do/ don't do what they are talked up to be.

Comparing Ubiquiti to Canopy is not an apples-to-apples comparison. For example, Canopy runs 10MHz channels usually and the default for Ubiquiti is 20MHz. Ubiquiti is also far more frequency efficient. Ubiquiti is also innovating very quickly and the firmware is moving forward almost as fast. If I was doing a direct comparison, I would use M series radios with a 5MHz channel.

It doesn't do any good to deploy a Canopy radio in a high noise environment since downshifting the modulation means a much lower throughput. It's better to improve your design to optimize for the best throughput. Motorola makes quality equipment, I don't argue that point. However, Canopy users are going to find Ubiquiti is going to continue to intefere with Canopy and Canopy is not going to win that battle.

So what if UBNT does introduce sync like Canopy? Do you think this is beyond UBNT to consider this?

There is no perfect radio. I would love to see a unit with GPS synch like Canopy, Polling performane,pcb design, and ethernet robustness like Trango, programable features like Mikrotik, and Pricing and packaging like Ubiquity.

I just wish the debates didnt go to what is the best unit--there is none--there are flaws with all of them. It is more constructive to discuss what the flaws in each system are and how to overcome them. Like it or not Ubiquity has entered the market with a price/performance ratio that is hard to beat. Yes they have many flaws but hopefully they will get better and they do work good. We cannot compare Ubiquity with Canopy the price difference is huge when you include provisioning tower with all of the goodies.

In my opinion the worst thing that happened was the push to OFDM for everything. Signaling methods used by Trango and Canopy such as BSK/FSK are much more forgiving and handle interference better. Give me a 20 meg per cpe capable system that does not use ofdm and I will not install another mikrotik, ubiquity, or wimax radio.

Wow I missed this thread, sorry for coming in on the 9th inning.

=======
Edited post - Seems when mods split topic it didn't quote properly
=======

I must have missed that memo. Do you have any citation or reference materiel?

Seems the original start of the topic was not as it should be during the topic split. Mods said it wasn't quoted properly.

Edited my post.

As for Airmax not being TDMA,
Last I knew Ubnt was claiming it was a mix of TDMA and some "extras" and no further info as to just how much TDMA it really is, or what's been changed on it. Can't exactly call it true TDMA as it's not exactly that.

Also last I knew they never released the exact details of what was different with it. I know it does follow TDMA closely but even I remember Ubnt crew saying it wasn't 100% TDMA but a mixture.

I agree with you 100%. You may notice that I tried to generalize a very important technical fact without mentioning specific vendors.

" Complex modulation [TDMA] systems work best in good RF environments. Standard 802.11 protocol with CSMA work better "by design" in poor RF environments

That is a necessary compromise in all [TDMA] systems which gives us more system capacity and other benefits"

Let me offer some common sense observations which are applicable to any past, present or future TDMA system.

Assume 100 active CPEs are assigned 10 millisecond sized TDMA slots (100 x 10 ms = 1 second) during which 100 average sized packets can be sent at the currently selected optimum link rate based on signal vs noise.

Some packets from AP gets clobbered by noise/interference but CPE can't immediately request a resend for that packet because the process would cross its slot boundary. So CPE has to wait until next upload slot assigned to it.

With standard round-robin style TDMA slot assignment, that may correspond to a long delay of (99 slots x 10 milliseconds) or almost 1 second.

Some high level applications could be affected by such long long delay. Applications could time out or request a whole bunch of packets to be resent.

Often external interference is very regular: SCADA, another TDMA system, etc. With deterministic slots, some unlucky CPEs will be affected much more than others.

Some advanced TDMA systems randomize slot assignments among active CPEs so more CPEs encounter interference but less often. This approach increases jitter.

Some advanced TDMA systesm use tiny slots (1 ms or less) but they need to break up average sized Ethernet packets into many chunks. That allows only corrupted chunks to be resent but requires an FPGA for packet processing. Even then, there is extra overhead per chunk which take away some of the already reduced throughput in noisy environment.

There are many such issues encountered by "hard" or "soft" TDMA WISP system designers and their "solutions" usually make compromises in some other performance area.

Than how can we know what type of TDMA system does a manufacturer use ? For example UBNT ..

well as far as i know they do nothing else than use some Features
the AR91xx and the AR92xx mimo chipset's can do if you have access to the licenced driver from Atheros. It behaves near a TDMA system right .. no hidden node problem etc but it's no real TDMA protokoll .
If the ubnt guys build a true TDMA system why in the world we can't
see a funktion like "Specify the time slice of TDM Coordination" .
How ever Ubnt's Airmax is working and it has it's place for some setups but i don't think that we will ever see something like AP to AP coordination and some other things simply because its not possible on such systems the added airtime would be hughe and they loose more than they win.

Like what? GPS sync?

If you want so yes how ever AP to AP coordination don't have to be done
necessarily with GPS . As long as you find a way to control what Ap is using what time slice/slot with in the same area as the other AP .Cluster mounts with multible AP's are one example .

Lutful and some others here pointed it out many times complex modulation
needs a very good RF environment. Since you can't control external nois
you all you can do is make sure that your own TDMA devices don't affect each other. There are not many things you can do to archive this proper bandplan and a coordination between the AP's are some of them.

How ever there are many cases where ubnt Airmax devices work really good as long as you can keep the signal in the mid 60's to low 70's and if you life in a country where they legal it's a really good option.