Hello OP25 devs,
I'm an amateur radio operator working on implementing an open source DMR tier 3 trunked radio base station based on a multi-carrier SDR structure. I have reached a point where I no longer know how to interpret the ETSI standard and I thought to ask for help here due to your expertise with trunked radio systems.
The outbound Reverse Channel PDU is used in DMR tier 3 for MS power control and MS de-key (transmit interrupt). Based on my interpretation of the standard, a couple of things are clear: 1. Reverse channel is only available if aligned timing is used 2. It is supposed to be embedded in voice burst F of a voice transmission on the alternate timeslot 3. The MS is supposed to monitor the alternate timeslot during transmit and derive RC information from it
There are a few scenarios where it seems to me these tier 3 services (power control, priority interrupt) cannot be available at all... 1. Alternate timeslot is a control channel 2. Alternate timeslot is idle 3. Alternate timeslot carries something other than voice payloads
The standard doesn't clarify these cases and I have doubt that these 2 services can even be reliably implemented.
Are you able to help me understand how the Reverse Channel is used in the real world by existing industry implementations of DMR tier 3?
Thanks, Adrian YO8RZZ
Hello Adrian
Sorry to say I'm not particularly conversant with DMR; anyone who can furnish more detail is encouraged to jump in here... Accordingly I'll have to limit this to generalities...
reached a point where I no longer know how to interpret the ETSI standard
Yes, I know the feeling. The TIA standards can induce the same feelings...
As for "aligned timing", I assume this to mean the capability of the MS to transmit TDMA fragments whose start and end times are in strict alignment with the slot boundaries demarcated by the outbound data stream transmitted by the BS. The feasibility of this is more questionable when the MS is implemented in PC software and separated from the hardware radio by a number of software layers and drivers each having its own associated latencies / unpredictable delays.
All of this goes double when the MS also has to retune the radio at precisely the timing needed to acquire reverse channel bursts sandwiched between its own transmitted fragments.
For all the above reasons I decided it's not currently feasible for us to try to implement a TDMA subscriber in terms of GNU Radio blocks. A proper job would seem to require an implementation at a level much closer to the RF hardware and/or the FPGAs.
There are a few scenarios where it seems to me these tier 3 services (power control, priority interrupt) cannot be available at all...
- Alternate timeslot is a control channel
- Alternate timeslot is idle
- Alternate timeslot carries something other than voice payloads
In each of these cases I wouldn't be surprised if there might be room enough to also sneak a PDU or two into the alt. timeslot somewhere along the line?
The standard doesn't clarify these cases
And it wouldn't be the first or the last time. In such cases there is always the option of performing RF captures and decoding them. This situation might require spacing/separation between BS, MS, and test receiver to enable everything to fit within the limited dynamic range of the test reeciver (or use two test receivers). This also requires access to a vendor system and setup, of course...
hth 73
Max KA1RBI On Sunday, January 14, 2024, 5:02:11 AM EST, Adrian Musceac kantooon@droiddv.org wrote:
Hello OP25 devs,
I'm an amateur radio operator working on implementing an open source DMR tier 3 trunked radio base station based on a multi-carrier SDR structure. I have reached a point where I no longer know how to interpret the ETSI standard and I thought to ask for help here due to your expertise with trunked radio systems.
The outbound Reverse Channel PDU is used in DMR tier 3 for MS power control and MS de-key (transmit interrupt). Based on my interpretation of the standard, a couple of things are clear: 1. Reverse channel is only available if aligned timing is used 2. It is supposed to be embedded in voice burst F of a voice transmission on the alternate timeslot 3. The MS is supposed to monitor the alternate timeslot during transmit and derive RC information from it
There are a few scenarios where it seems to me these tier 3 services (power control, priority interrupt) cannot be available at all... 1. Alternate timeslot is a control channel 2. Alternate timeslot is idle 3. Alternate timeslot carries something other than voice payloads
The standard doesn't clarify these cases and I have doubt that these 2 services can even be reliably implemented.
Are you able to help me understand how the Reverse Channel is used in the real world by existing industry implementations of DMR tier 3?
Thanks, Adrian YO8RZZ
Hi Max,
Thanks for the response, I think I found my answer meanwhile, funny enough your interpretation of aligned timing made me read again part 1 of the standard, and I realized that my interpretation of it was actually wrong :)
Aligned timing is used to describe the RF timeslot transmission arrangement between MS transmission and BS transmission. It actually does not involve TDMA frame sample alignment in time between MS and BS. That's sort of implied and the standard does not prescribe how to do that.
Personally I use GNU radio timestamps to have a deterministic transmission time for BS so MS can be received at the right time. The code for that is not very complicated for the BS side, if you want I can link here the GNU radio flowgraphs with timestamps that I wrote in case you ever want to implement the MS side in OP25, the code is very arcane but it works.
The MS side in SDR is an interesting discussion. The way I would do it for the MS is to keep a SDR-time reference with RX_TIME timestamps, search for BS outbound CACH or VOICE / DATA sync, and update the time reference by adding the number of samples from the center of the sync sequence to the end of the TDMA frame. BS-side is easier since BS sets it's own time reference. I haven't really investigated this since the main problem I see is switching a PA on and off fast enough. With pin diode RF switches, we can achieve tens or hundreds of nanoseconds switch time, but then there's also the PA ramp-up and ramp-down which should take less than tens or hundreds of microseconds. I don't know really how to achieve that, I have a LimeRFE PA board, but I control it via USB so I think it's too slow for that.
Coming back to the aligned timing issue, to me it seems the ETSI language is clear as mud. I would have expected aligned timing to mean that the BS is transmitting timeslot 1 while the MS is transmitting also timeslot 1. But no, in ETSI lingo, that is called offset timing. Aligned timing actually means BS is transmitting timeslot 2 while MS is transmitting timeslot 1. It became clear to me after I read you message that actually in aligned timing mode, the MS transmits timeslot 1, then switches to RX during timeslot 2 and hears its own transmission as the BS transmits timeslot 1 during MS RX period. So that also neatly explains why full duplex calls are not possible in aligned timing mode.
So problem solved, the Reverse Channel can always be transmitted in the default aligned timing mode, since it is always appended to the active transmission instead of the alternate slot transmission as I believed at first!
Thanks and have a great day, Adrian
Thanks for the detailed explanation, Adrian. This is a very interesting question, especially regarding how Reverse Channel is actually handled when the alternate timeslot is idle or used for control signalling. I’d also be interested to hear from anyone with practical experience of DMR Tier 3 implementations.
-
Adrian Musceac -
gedeonluc＠proton.me -
ikjtel