Visual comparation of TDMA systems

Hi,

we're used to the TDMA everywhere but there are not too much people how have seen the TDMA.. the frame.

Here goes 4 screenshots of a time-domain (100 msecs) capture of diferent TDMA systems...

1) Ubiquity M5 (20 Mhz, ~40 CPEs, heavily loaded but  unable to move more than 12-15 mbit/s):  if anybody can see the "TDMA" please let me know....  the frame is absolutely non structured.. it is a plain WIFI (although airmax is enabled). So airmax more than a TDMA it may be a "round-robin" ?

2) ePMP1000 (20Mhz, ~30 CPes, lightly loaded <10-15 Mbit/s) : frame is 5 msegs,  The spikes are at the AP, then other signal with lower power are the CPEs

3) PMP450 (20 MHz, ~90 CPes, <40 Mbit/s) : frame is 5 msegs.

4) Albentia 802.16d (10 MHz in V polarization only, 40 CPEs, heavly loaded: ~30 Mbit/s). Frame is 10 ms.

Best regards

Antonio

14 Likes

Very nice! 

What equipment did you use to take those captures?

Could you try to capture how does Epmp gear looks like in "Flexible" mode?

1 Like

Hi,

these traces are captured by the albentia equipment itself, both AP and CPEs have the ability to perform local/remote spectrum and time analysis.

Time analysis is very useful to determine which kind of interference (pulsed, continous, etc.) or neighbours you have...

As per my understanding, the flexible mode is the same as this. The only difference is the UL/DL subframes inside the 5ms frame which are fixed to the split ratio in fixed mode or variable in the flexible mode but the frame is the same.

Best regards

Antonio

4 Likes

UBNT looks like a mess for gods sake. Are you sure that it is captured correctly ?


@doush wrote:

UBNT looks like a mess for gods sake. Are you sure that it is captured correctly ?


I think that the reason it looks like this is because UBNT is using flexible time frames and because UBNT doesn't have any ATPC controls, unless clients have their power manually tuned to the same level, hot clients will appear as spikes. It would be interesting to see if ePMP would look similar if set to flexible frames and ATPC disabled.

100% sure.

Airmax seems plain wifi; it may have some  round robin or so to solve the hidden node issue.

No TDMA structure in the way we are used to: frame duration must be fixed eg. 5 ms, and what can change is the part of the frame that is used for download or for upload. This can be also fixed or flexible but the is no visible frame duration.

The unstructured frame implies the very low efficiency of the cell and it increases as  CPEs are more far away from the AP and this is caused to the airtime wasted on travelling packets DL/UL (the gaps between tx/rx)

The spikes are the AP as it is colocated with the albentia BS which capured the data. CPEs are small spikes.

Hello Dear,

can you update  extremely interesting post to new ePMP 2000 and new Ubiquity PtMP products ?


@telecomunicazio wrote:

Hello Dear,

can you update  extremely interesting post to new ePMP 2000 and new Ubiquity PtMP products ?


Hi,

Frames structure and duration on ePMP2000 is the same with ePMP1000.

2.5 and 5 ms TDMA frames.

Thank you.

Can you make the same tests with Mikotik gear using NV2?

Hello, interesting post, I would like to see with the AP Rocket AC Prism with the same frame fize of 5 ms with cpe AC also, to see the difference of the line common mma airmax, would be very interesting as it would resemble the same conditions of equipment, such as epmp.

Hello, interesting post, I would like to see with the AP Rocket AC Prism with the same frame fize of 5 ms with cpe AC also, to see the difference of the line common mma airmax, would be very interesting as it would resemble the same conditions of equipment, such as epmp.

Great post! Has much changed?

I’d be curious to know how Wi-Fi 6 802.11ax compares with the proprietary TDMA protocols.

1 Like

https://en.wikipedia.org/wiki/Time-division_multiple_access