TDMA Repeater?

[Nerumph]

Hello all, so I don't know if this is the right place to discuss this, but it is about digital protocols so...

I was looking around the internet and catching up on some of the latest advancements with regards to repeaters and just general radios. A couple weeks ago I came across the term TDMA (time division multiple access) as opposed to FDMA (frequency division multiple access). It didn't take long for the idea to pop in my head of: why are repeaters not utilizing TDMA more than FDMA to repeat their signal?

I can only think of one reason. Lets say, on the repeater side, Rx would be on T1 and Tx on T2. Would T2 drown out T1's next incoming transmission byte? Would there be enough time to add a >millisecond delay to the Tx to compensate?

Obviously there would be equipment to worry about (mainly on the mobile radio side). I was exploring that possibility around too where TETRA radios did something similar to this.

I can see this being a real repeater revolution in the market if it can get past those hurtles; finally no cavities and filters to worry about! Not to mention finally being able to switch channels without re tuning fragile equipment!

What do you all think about this? Is it viable? Is it the future?

 

[N4GIX]

You haven't quite grasped the concept of TDMA. You speak of TS1 and TS2 as though they are different frequencies. They are not...

They are simply two halves of a single frequency pair. The major difference between a conventional analog repeater and a DMR (TDMA) repeater is that where the analog can only handle one user at a time, a DMR repeater using TDMA can handle two users at the same time.

There are many DMR (TDMA) amateur repeaters in use here in the U.S., currently >1,000 on the DMR-MARC network, and many others on the Hytera network. I have no idea how many are in use by the commercial and public safety sectors.

All the above aside however, the infrastructure for a TDMA repeater is identical to that required by a conventional analog repeater. Duplexers, circulators, isolators, and other necessary hardware is commonly required.

BTW, one other difference between analog and digital repeaters is that the digital repeater is full-duplex. The repeater is in constant "conversation" with the user's mobile or HT radio whenever the user is transmitting.

 

[Nerumph]

Oh no I fully grasp the benefits to TDMA in context of dual conversations. That one talkgroup can use the same frequency of a second for a completely different private or TG-wide communication at the same time.

My theory, though, would be that instead of using frequency TX and frequency RX to repeat the signal, thus requiring a duplexer, why not use timeslot 1 (TX) and timeslot 2 (RX) on one frequency? I understand that in high RF traffic environments one would still need to use bandpass filters to cut back on desense and yes a feedback prevention device such as a circulator, but for places where there is little RF crosstalk, would it not be lighter and more compact to just have one duplex radio/ antenna doing the relaying?

I mainly want to know what the physical limitations and barriers to this form of repeating would be. And if those physical limitations could be circumvented, it would really only be a matter of updating firmware on TDMA capable radios to include the new method of repeating.

 

[toastycookies]

Oh no I fully grasp the benefits to TDMA in context of dual conversations. That one talkgroup can use the same frequency of a second for a completely different private or TG-wide communication at the same time.

My theory, though, would be that instead of using frequency TX and frequency RX to repeat the signal, thus requiring a duplexer, why not use timeslot 1 (TX) and timeslot 2 (RX) on one frequency? I understand that in high RF traffic environments one would still need to use bandpass filters to cut back on desense and yes a feedback prevention device such as a circulator, but for places where there is little RF crosstalk, would it not be lighter and more compact to just have one duplex radio/ antenna doing the relaying?

I mainly want to know what the physical limitations and barriers to this form of repeating would be. And if those physical limitations could be circumvented, it would really only be a matter of updating firmware on TDMA capable radios to include the new method of repeating.

TS1 and TS2 are one frequency.

The repeater TX on one frequency and RX on another one close by, usually around a 5 mhz split in the uhf band.

That is the reason you need a duplexer.

 

[902]

why are repeaters not utilizing TDMA more than FDMA to repeat their signal?

1) Legacy equipment. Not everyone ditched analog or their other formats and there are still plenty of analog/ "FDMA" radios around.

2) Marketing and purpose. Some equipment manufacturers push their idea better, and in some parts of the country, some manufacturers' reps push a product line better than others. There are places where TDMA is used, others where very narrow FDMA is used, and still others where it's analog FM. And, for VHF high band, very narrow FDMA is probably the best solution, considering other waveforms will spill into the adjacent channel.

There were also difficulties creating TDMA devices initially. They weren't seen until relatively recently.

TDMA and the more advanced waveforms are based on time and synchronization in addition to frequency, so the receiver hears everyone because not everyone is transmitting at the same time. They transmit in packets, or bursts of data in their assigned timeslot, whether that's 2/TDMA like DMR or 4/TDMA like TETRA. On a DMR system, for example, a receiver is listening to timeslot 1 - 2 - 1 - 2 - 1 - 2 - 1 - 2... not both at the same time. It's because of fast computers and vocoding that it seems like it's at the same time.

 

[buddrousa]

Conversation 1 is on input 455.100 time slot 1 and output 450.100 Time Slot 1
Conversation 2 is on input 455.100 time slot 2 and output 450.100 Time Slot 2

 

[Nerumph]

Those of you that are telling me how conventional repeaters work, thank you. However, I already have an understanding on how they work in the conventional sense: I've already built one of my own and thus know how the offset system works for conventional/ digital repeating systems.

My question in here is not on how conventional repeaters work, but how feasible a TDMA, single frequency repeater system would be.

Again, outlined in the way that buddarousa described how standard TDMA works: repeater 1: input is 455.100 input (rx) is on timeslot T1 and output is on timeslot T2 on THE SAME FREQUENCY and preferable with a fully duplex radio (yes they do make those now).

Is this a feasible, theoretical way to do repeating of signal? What are the PHYSICAL LIMITATIONS (as in waveform propagation issues or RX desense issues) that may arise as a result of this sort of new TDMA implementation?

 

[mmckenna]

Oh no I fully grasp the benefits to TDMA in context of dual conversations. That one talkgroup can use the same frequency of a second for a completely different private or TG-wide communication at the same time.

My theory, though, would be that instead of using frequency TX and frequency RX to repeat the signal, thus requiring a duplexer, why not use timeslot 1 (TX) and timeslot 2 (RX) on one frequency? I understand that in high RF traffic environments one would still need to use bandpass filters to cut back on desense and yes a feedback prevention device such as a circulator, but for places where there is little RF crosstalk, would it not be lighter and more compact to just have one duplex radio/ antenna doing the relaying?

I mainly want to know what the physical limitations and barriers to this form of repeating would be. And if those physical limitations could be circumvented, it would really only be a matter of updating firmware on TDMA capable radios to include the new method of repeating.

I understand what you are saying. Single frequency repeater/trunking. Someone had a product out a few years ago that would do this, but I can't remember who it was.

Single frequency, let's say 460.500. With the ability to key it rapidly enough and the ability for the hand held/mobile radios to synchronize with it, TX into the repeater could be handled in a time slot and the RX out of the repeater could be handled on a second time slot.

It's got some possibilities, but there are more efficient ways to do it. The synchronization of the time slots between uplink and downlink would be difficult to nail down due to multipath, but I'm sure that's something that could be solved.

On the other hand, there may not be too much of a market for this. The LMR industry is pretty well stuck in frequency pairs for repeaters. Band plans for UHF, 700, 800 and 900MHz are already set up for paired frequencies. Narrow banding and refarming will free up more frequencies. Trunking really solves a lot of "lack of frequency" issues if the money is there.


Still, it's an interesting idea, and may have a market in some places.

 

[nd5y]

Again, outlined in the way that buddarousa described how standard TDMA works: repeater 1: input is 455.100 input (rx) is on timeslot T1 and output is on timeslot T2 on THE SAME FREQUENCY and preferable with a fully duplex radio (yes they do make those now).

I read a few years ago where somebody in Japan was developing a system like that but never heard any more about it.

Is this a feasible, theoretical way to do repeating of signal? What are the PHYSICAL LIMITATIONS (as in waveform propagation issues or RX desense issues) that may arise as a result of this sort of new TDMA implementation?

The physical limitations are issues related to RX to TX and TX to RX turnaround time and time slot length. Using only one frequency and one user at a time on two time slots would be different than full duplex systems with two users at different distances from repeater talking at the same time on four time slots.

I don't know if TETRA, DMR and P25 phase II have staggered TX/RX times or advance and retard the slot timing to compensate for differing user distances from the base station like what is done in GSM.

 

[mmckenna]

Is this a feasible, theoretical way to do repeating of signal? What are the PHYSICAL LIMITATIONS (as in waveform propagation issues or RX desense issues) that may arise as a result of this sort of new TDMA implementation?

I don't think RX desense is an issue since many systems do this already, however with enough separation between TX and RX frequencies. The key is getting it all synchronized, like I mentioned above.

I'll do some searching and see if I can find what that product was that I remember for a few years ago.

 

[N4GIX]

I don't know if TETRA, DMR and P25 phase II have staggered TX/RX times or advance and retard the slot timing to compensate for differing user distances from the base station like what is done in GSM.

DMR TDMA certainly does change the timing based on the propagation distance in order to keep thing in synch. Fortunately in most cases the mobile will loose the repeater due to excessively low RSSI long before loosing synch will become a problem.

If the user is running a Motorola with auto-roaming enabled, it will switch to a closer repeater even before complete loss of signal.

 

[Nerumph]

I understand what you are saying. Single frequency repeater/trunking. Someone had a product out a few years ago that would do this, but I can't remember who it was.

Single frequency, let's say 460.500. With the ability to key it rapidly enough and the ability for the hand held/mobile radios to synchronize with it, TX into the repeater could be handled in a time slot and the RX out of the repeater could be handled on a second time slot.

Eeeexactly! Single frequency, not paired. Still duplex, just no duplexer (and hopefully only only one transmitter).

It's got some possibilities, but there are more efficient ways to do it. The synchronization of the time slots between uplink and downlink would be difficult to nail down due to multipath, but I'm sure that's something that could be solved.

Hmmmm... ya multi prop is something I hadn't considered, but I would think it would be a problem with DMR regardless of if it was repeating or just dualslot conversing? And because of that, would already have compensation techniques in place? (I mean, otherwise TDMA would be useless). Would there be too much difference from standard TDMA to this proposed iteration of it?

On the other hand, there may not be too much of a market for this. The LMR industry is pretty well stuck in frequency pairs for repeaters. Band plans for UHF, 700, 800 and 900MHz are already set up for paired frequencies. Narrow banding and refarming will free up more frequencies. Trunking really solves a lot of "lack of frequency" issues if the money is there.


Still, it's an interesting idea, and may have a market in some places.

I do agree that right now, while the 12k bandwidth limits helped to open the airwaves, it would not be too much of a viable market place for this sort of system. But with new repeaters going in every day and the ever increasing number of communications spreading across the world, it would only be a matter of time before this might be seen as something worth investing in. Hell, by the time that happens we might get even more efficient code usage of 4, 5, 6, 8, 10 TDMA systems! Imaging how that'd sound on the scanner without tracking software!

----------------------

The physical limitations are issues related to RX to TX and TX to RX turnaround time and time slot length. Using only one frequency and one user at a time on two time slots would be different than full duplex systems with two users at different distances from repeater talking at the same time on four time slots.

It wouldn't technically be 4 timeslots though would it? The TX and RX radios would, so long as they were standard, non duplexing radios, be sync'd with a global clock and not be able to do full multiplex, right? As in, they would technically be pulling in two timeslots on the RX frequency and rebro-ing the same two timeslots out on the TX frequency?

I don't know if TETRA, DMR and P25 phase II have staggered TX/RX times or advance and retard the slot timing to compensate for differing user distances from the base station like what is done in GSM.

I would think this is already in place with DMR as multipath would be an issue, as said above, no matter the iteration of TDMA. Or is there a factor not taken into account here that would make one frequency TDMA repeating special in this account?

------------------

DMR TDMA certainly does change the timing based on the propagation distance in order to keep thing in synch. Fortunately in most cases the mobile will loose the repeater due to excessively low RSSI long before loosing synch will become a problem.

This is what I am thinking would be in place regardless, so it would just really require a firmware upgrade on radios to implement repeating on a single frequency, right?

If the user is running a Motorola with auto-roaming enabled, it will switch to a closer repeater even before complete loss of signal.

Yea, this is something that might be tricky. But then again, if these repeater stations were sync'd already with eachother so as to not ghost eachother's signal, wouldn't they be able to keep global clock of TDMA in the same manor as well?

 

[mmckenna]

I do agree that right now, while the 12k bandwidth limits helped to open the airwaves, it would not be too much of a viable market place for this sort of system. But with new repeaters going in every day and the ever increasing number of communications spreading across the world, it would only be a matter of time before this might be seen as something worth investing in. Hell, by the time that happens we might get even more efficient code usage of 4, 5, 6, 8, 10 TDMA systems! Imaging how that'd sound on the scanner without tracking software!

I think it would depend on location. As more and more public safety systems go to region wide trunked systems, the need for spectrum decreases. The market for LMR radios for non-public safety users really took a hit when NexTel and other mobile carriers hit the market.
To make a product like this valuable you'd need to either beat the cellular carriers at their own game, or find an untapped market.
While duplexers and antenna systems are expensive, duplexers last a long time. Antennas and feed lines have a limited life span, but on a single antenna system, like you'd find on a conventional repeater, you'd be replacing those anyway.

As for more efficiency, you won't get much useable audio quality if you break down a 12.5KHz channel much more. As it is 6.25KHz NXDN sounds pretty good, but you probably wouldn't want to go much more than that.
So, even with TDM you aren't going to squeeze a lot more traffic into a 12.5KHz channel unless some pretty wild multiplexing schemes are used.

 

[K5MPH]

Yeah,no duplex just separate the Tx and Rx with digital time slots is this not like the GSM system for cell phones now not for sure might be wrong,this would be great if really could be done......

 

[nd5y]

It wouldn't technically be 4 timeslots though would it? The TX and RX radios would, so long as they were standard, non duplexing radios, be sync'd with a global clock and not be able to do full multiplex, right? As in, they would technically be pulling in two timeslots on the RX frequency and rebro-ing the same two timeslots out on the TX frequency?

I would think this is already in place with DMR as multipath would be an issue, as said above, no matter the iteration of TDMA. Or is there a factor not taken into account here that would make one frequency TDMA repeating special in this account?

What you are talking about running TDMA on one frequency simplex would take at the minimum two time slots, one tx and one rx. Physical limitations would depend on the how long the time slots are, propgation delay (how far apart the stations are), how fast the transmitter turns on and ramps up to full power, how fast the reciver turns on and is able to receive valid data, and how long any latency is that is caused by the CPU and audio processing and buffering times. Then you need to figure out the channel bandwidth and data rate required.

High speed wide band high data rate simplex data is done all the time, that's how wifi and bluetooth work, but I don't know if they use actual time slots that are the same lengh and contain the same amount of data all the time.

 

[nd5y]

this not like the GSM system for cell phones now not for sure might be wrong,

About 15 years ago I worked for a telecom manufactuer and had GSM training. GSM (at least on the phones I was trained on in the 1900 MHz US band) isn't full duplex as far as the RF part is concerend. The phones were not capable of transmitting and receiving at the same time on 1900 MHz and didn't have a duplexer for that band like they did on 850 MHz. They transmit and then receive at different times on two frequencies just like a subscriber radio using a repeater. The transmit and receive time slots are staggered so that the phone could key up and transmit on slot 1, for example, then unkey and receive slot 1. Slot 1 transmit and receive don't happen simultaneously like on a DMR repeater. Then the rest of the time the phone could be checking the control channel and neigbor cell signal strength during the other time slots. They can also use more than one time slot to transmit data if needed. If I remember right GSM as 8 time slots that are 400 microseconds long in a 200 kHz wide RF channel. They also had frequency hopping so each time they transmit could be on a differnt RF channel.

 

[slicerwizard]

My question in here is not on how conventional repeaters work, but how feasible a TDMA, single frequency repeater system would be.

Great. So now any radios near me, that had been TXing a good distance from where I'm RXing (like 824 MHz vs 851 MHz, for example), are now TXing in the same band I'm trying to hear. So as I try to hear my repeater on 851.0125, some guy across the street is TXing to his repeater on 851.0500 (or even closer)

Do you think this is going to work? Or were you planning to somehow time-sync the TX/RX slots on completely unrelated systems?

 

[Project25_MASTR]

So it's not a new idea. The idea has actually been bounced around many Tier III manufacturers. Now Simoco has a feature called Simple Repeater Dual Slot...which is really a misleading way for saying pseudo-trunking but it gets you going (and it's been heavily discussed at Simoco for adding this feature).

Now, could it be done with DMR, yes. Between each timeslot is a 3 ms long sync frame. The subscribers are physically capable of performing the switch quick enough (a repeater would have to be mobile based and not off a repeater chassis).

Unfortunately, I don't think you'll ever see a manufacturer try to build it though for DMR. Reason being, the primary market where this feature would be used is currently dominated by P25. Could we see it in in someone's Phase 2 offerings one day? Maybe but I don't think we'll see it in DMR.

 

[ElroyJetson]

So you want to do full duplex on a single frequency using TDMA technology to separate TX from RX.

OK, fine, I get that.

Now please explain the benefit of this proposed configuration. What does it give us that we don't already have, without having to invest in the R&D costs required?

That is the question that must always be asked. For the cost of development, will the benefit be a worthy upgrade?

Is what you propose feasible? Yes. The RF technical challenges are not insurmountable. Timing issues should be of no more concern than they are in any DMR system.

But ultimately the question is one of cost vs. benefit.

And I'm not seeing any benefit that isn't equalled by existing solutions.

 

[902]

So you want to do full duplex on a single frequency using TDMA technology to separate TX from RX.

OK, fine, I get that.

Now please explain the benefit of this proposed configuration. What does it give us that we don't already have, without having to invest in the R&D costs required?

That is the question that must always be asked. For the cost of development, will the benefit be a worthy upgrade?

Is what you propose feasible? Yes. The RF technical challenges are not insurmountable. Timing issues should be of no more concern than they are in any DMR system.

But ultimately the question is one of cost vs. benefit.

And I'm not seeing any benefit that isn't equalled by existing solutions.

Obviously I'm not him, but I see one right off the bat - instead of searching for a frequency pair for a repeater, a repeater could be implemented with one channel. It could be very valuable for unpaired spectrum, like VHF. It would probably work very nicely in a sterile environment. The problem in the real world would be receiver blocking from co-channel and adjacent channel systems. Another situation could be MO3 operations, where one frequency employed in 2/TDMA could get full control and break-in capabilities.