In Zero Retries 0175 — MMDVM-TNC is (Kind of) Real, I wrote:

I plan to do a future article in Zero Retries — Explaining the Use Case for Data Over Repeater, about why such a data capability is important.

Clinton Blackmore VE6CWB recently sent me an email, which I’m reprinting here with his permission. I’m including explanations to his questions because he asked some good questions which help me frame the longer discussion that will be Part 2.

It seems like you’ve mentioned MMDVM repeaters a fair bit recently in Zero Retries — enough that it’s caught my attention.

I’m happy to wait for your forthcoming explanation — “Explaining the Use Case for Data Over Repeater” — promised in ZR 0175. In preparation for that article, I have a few questions and I hope you’ll answer address them.

It sounds like you might add data transmission as a supplement to regular repeater activities. Would it just be using the repeater for data some of the time and voice at other times? Or would you imagine having extra frequencies (or wider bandwidth) for the repeater?

The very easiest mental model for the usage of a mixed mode FM voice / MMDVM-TNC data repeater is that hundreds or thousands of mixed mode repeaters already exist and are deployed and in use…* it’s just that the “data” capability of those repeaters is “fixed formatted” into digital voice (which in the end, is transmitted as data, not conventional FM voice). Example:

• Yaesu kickstarted the widespread deployment of mixed mode repeaters in the US with their DR-1X and now their DR-2X repeaters — FM voice and System Fusion digital voice (with a little bit of data).

• Icom’s current D-Star repeaters are mixed mode — FM voice and D-Star digital voice (with a little bit of data).

• Hytera’s DMR repeaters are mixed mode — FM voice and DMR digital voice.

All of these repeaters listen on the input frequency for FM voice and (their flavor of) digital voice, and then…

• When an FM voice signal is received, the repeater transmits FM voice.

• When a digital voice signal is received, the repeater transmits digital voice.

“Data capability” on D-Star, System Fusion, DMR, and P25

It’s a completely different discussion, perhaps meriting a detailed treatment at some point in the future, but D-Star, System Fusion, DMR, and P25 digital voice modes all incorporate some limited data capabilities. Thus with the installation of a new mixed mode D-Star, System Fusion, DMR, or P25 repeater, there (could have been) a data capability “built in”. Thus you might consider this entire discussion of mixed mode repeaters (FM voice / MMDVM-TNC data) to be moot.

However, in my studies of the data capabilities of all of these systems, the data capability in those systems is an afterthought at best, or only partially (poorly) implemented or in the case of System Fusion, deliberately limited. Not to mention the data capabilities in these systems is now far behind the current state of the art, such as no use of Forward Error Correction (FEC) for data.

Thus, to add good, reasonably fast data capability to a repeater¹, Amateur Radio has to invent new technology, such as adding an MMDVM and incorporating MMDVM-TNC data, or perhaps M17 or Opulent Voice. All of these new systems have a well-implemented data capability in addition to digital voice.

If I receive enough feedback that Zero Retries readers want to read a detailed critique of the limited data capabilities of D-Star, System Fusion, DMR, and P25, let me know.

Thus, what I’m proposing with retrofitting FM voice repeaters, continuing the primary use of an FM repeater’s existing voice operations, and adding an MMDVM to support a secondary use of MMDVM-TNC data is conceptually the same as what we’ve been doing for years on hundreds or thousands of existing mixed mode repeaters.

My observation over decades of promoting the conversion of little-used FM voice repeaters for 9600 bps FSK data (and creation of new such repeaters) is that no repeater owner is willing to do so. The idea of “data repeaters” is just completely foreign and not understood, and there are only historical examples to cite such as [The Puget Sound Amateur Radio TCP/IP Network](https://www.superpacket.org/2021/03/the-puget-sound-amateur-rad io-tcpip-network-circa-1995.html). Thus a scheme like adding MMDVM and the subsequent ability to also use MMDVM-TNC data, that can (theoretically) switch seamlessly from FM voice to MMDVM-TNC data will likely be better received.

Extra frequencies is an interesting idea, especially if the repeater could receive on another input channel, possibly on another band. I’m not currently promoting wider bandwidth, though I’m definitely promoting the preservation of current (non-narrowband) channels.

Also, I’m dying to know, as transmitting data over a repeater in an infrastructure-up situation doesn’t make a lot of sense (vs. downloading files from the internet) unless you hope people skill-up and can use it in an infrastructure-down situation. What sort of data might one transmit? (I seem to recall you saying something about FLArq in this vein, too).

In an “infrastructure up” (neat descriptor!) situation, I’m positing using a repeater in data mode in the same way we currently use a repeater for voice. In “infrastructure up” situations, we certainly don’t need to use repeaters for voice communications — we have mobile phones and mobile phone networks for any conceivable voice needs.

The “broadcast files” capability I referenced is actually flamp² which stands for Fast Light Amateur Multicast Protocol. flamp is one of the fldigi suite of data modes. Literally any useful file can be distributed via flamp — maps, weather bulletins, repeater / node frequencies (code plugs), local Amateur Radio events and nets calendar, club bulletins, photos, even rudimentary web pages — pure HTML is remarkably compact. If you have a receiver on the flamp frequency, even a computer as simple as a Raspberry Pi Zero 2 W is usable for this purpose, and there’s no transmit capability required. Thus Amateur Radio clubs could use flamp to distribute information to those interested in Amateur Radio. The basic point is that with flamp, we can use Amateur Radio to promote Amateur Radio, similar to the way we use nets and HF chats to promote Amateur Radio to folks who might be listening to learn what Amateur Radio is all about. And yes, this capability is “legal” — this sort of one way information transfer is well within the category of Amateur Radio “bulletins”`.

My other question about retrofitting repeaters is, um, well, again about frequencies and slots. It seems you are suggesting we could upgrade to digital radios from analog ones using existing repeaters without leaving behind users with older tech, but how?

Ben Foght N5AMD wrote a great article on the hardware installation of an MMDVM into a repeater — How to make a MMDVM Digital Repeater. For the purposes of adding data capability (MMDVM-TNC), the MMDVM hardware merely has to get updated to add MMDVM-TNC as another mode, just like you would add another digital voice mode to an MMDVM, such as M17. Then the repeater passes FM, or digital voice, or MMDVM-TNC data. MMDVM-TNC is still experimental — it’s not currently in the main software distribution for MMDVM, so actual experience with it in the real world is a work in progress, and this is one of the many projects I have queued up for my test repeaters in N8GNJ Labs to get some real experience using MMDVM-TNC and “mixing and matching” FM voice operations and MMDVM-TNC data so I can advise from actual experience.

I could see a repeater identifying an input signal as analog FM or some digital mode … but then it surely needs to transmit an FM audio signal out. (Admittedly, there’s a couple of people with poor signals, who, if they could use digital modes with forward error correction, would probably come through way better). I assume the digital radios don’t want to listen to analog FM signals, though. I could imagine a repeater using two pairs of frequencies, one for digital radios and one for analog, and mirroring messages between them, but that sounds like extra expense and gear and getting repeater area authorities to allocate additional frequency pairs — which, which not insurmountable is not an easy upgrade.

Speaking of “repeater identifying…”, a longstanding, very widespread “missed opportunity” in Amateur Radio repeaters is that current generation repeater controllers can “ID” and transmit voice bulletins with high quality digitized voice, either recorded human or pleasant sounding artificial (AI) voices. Thus the repeater can, perhaps hourly, provide a rotating set of voice bulletins about the repeater’s various capabilities, upcoming nets, etc. Personally, after four decades now of listening to very low quality artificial voices to ID repeaters, any repeater that I have any influence on will not use such crappy voice, with the subsequent crappy impression to prospective Amateur Radio Operators.

At the moment, we don’t have “digital radios”, at least for practical use (reasonable transmit power). Thus what we’re using is conventional FM radios with data modems (formerly known as TNCs) attached. What we’re calling “data radios” are conventional FM radios with a special input, called, variously, “flat audio”, “data jack”, “9600 input”, etc. that can be used to bypass the conventional voice pre-emphasis and voice de-emphasis stages of the radio that make human voice sound good on an FM radio, but distort higher speed data “audio” that is sent through those stages.

The data modems are endlessly patient, and will happily listen to a mixed mode (FM voice / MMDVM-TNC data) repeater, and only “spring into action” when it actually decodes an MMDVM-TNC transmission.

All of this would be much easier explained in an interactive block diagram, which I intend to do eventually. I have not yet spent time learning how to instruct an AI like ChatGPT to create such diagrams, even videos, but that’s on my long to-do list.

There are endless refinements possible for mixed mode repeaters.

For one, consider the possibility of adding a secondary receiver to a mixed mode repeater. Imagine a 146.76 MHz FM repeater, with its input frequency at 146.16 MHz. This repeater gets retrofitted with an MMDVM and a secondary receiver, perhaps even on a different band like 222.01 MHz. When someone transmits FM voice on the primary input frequency, FM voice is retransmitted like normal. When someone transmits MMDVM-TNC data on 222.01 MHz, that data is routed into the MMDVM and recognized as MMDVM-TNC data, and is retransmitted as MMDVM-TNC data.

Another potential refinement for mixed mode operation is that most current repeaters can be configured to transmit a subaudible tone when the repeater transmits in FM voice mode, and a FM user radio can be configured to mute the audio unless that subaudible tone is received. A digital voice radio will generally automatically switch between FM voice and digital voice.

Part 2 Upcoming - First Principles of Mixed Use FM voice / MMDVM-TNC Data Repeaters

In Part 2 that I’ll publish in the next week or two, I’ll discuss the following ideas (“First Principles”) in more detail:

• Amateur Radio capabilities, during normal (non emergency) times is generally inferior to the capabilities of commercial (and public safety / government) Internet / cellular / satellite services. What Amateur Radio can do better than Internet / cellular / satellite is personal experimentation and training / learning with radio technology, some research, and some recreational activities.

• Why do we in Amateur Radio VHF / UHF operating want to use repeaters at all? Because repeaters allow wide-area communications by limited power / limited antenna stations over widely varying terrain. In a phrase, using a repeater “levels the playing field” for all stations, including portable radios, mobile radios, and base stations. It’s generally easier, more effective, and provides a more satisfying experience to use a repeater instead of simplex communication on VHF / UHF.

• Why do we need data capabilities in Amateur Radio at all, including repeaters? Again, personal experimentation and training / learning with radio technology, some research, and some recreational activities. Some would argue that we need to be able to use data capabilities in Amateur Radio for communications in emergency conditions that increasingly require data capability rather than voice. Example, distributing the occupancy list of a mass casualty shelter when a hurricane has wiped out terrestrial infrastructure including power and mobile networks. It just doesn’t work to try to read hundreds of names, accurately, via voice.

• Why not use (single channel, store and forward) digipeaters for data operations instead of repeaters? Generally, for the same reason we don’t use single channel store and forward (simplex) voice repeaters - the experience is poor in comparison to a realtime full duplex repeater.

• Isn’t using a repeater a “single point of failure”? Yes, but we don’t seem to consider that an issue for voice operations… and we have a lot of (redundant) repeaters to switch to if one fails. Not to mention we know how to quickly stand up a temporary or emergency repeater.

• In the 2020s, increasingly many… arguably most, repeaters have gotten quieter and quieter with less and less usage³. In past decades, we used repeaters a lot, for example, for autopatch (some repeaters were built exclusively for autopatch). But now, we don’t use repeaters nearly as much. Thus these “almost entirely quiet repeaters” aren’t widely used. Thus does it make sense to continue supporting repeaters that are barely / rarely / almost never used? Wouldn’t it be a better use of such a repeater to adapt it for data use with an MMDVM and add MMDVM-TNC data capability? Ditto for repeaters that are unused during certain times such as late evening / early morning? Computers and data communications are “patient” and can wait.

• Why do a disruptive hardware change of installing an MMDVM when a repeat owner could allow an unmodified repeater to be used with VARA FM or conventional 1200 bps AFSK packet radio, or perhaps a newer packet radio mode such as what’s possible with a NinoTNC? That’s a fair point worth discussion… but generally such operations are disruptive to normal operations on the repeater — FM voice users will hear “a lot of garbage” on the repeater when data modes are in use.

• What’s so special about MMDVM-TNC? Why is it worth considering converting a repeater to support it? I posit MMDVM-TNC is worth considering because it is a well-designed, fast data mode that can be run from an MMDVM, and MMDVMs have already been successfully implemented by existing repeaters (initially, for digital voice operations). MMDVM-TNC uses a robust modulation method, uses an advanced Forward Error Correction (FEC) system (IL2P), and offers different speed tiers, with a minimum 9600 bps, and potentially as fast as 38400 bps, exceeding the 25 kbps possible with VARA FM.

• What if we do go through a disruptive upgrade to add an MMDVM and “something better comes along” and we have to do it all over again? That’s quite possible, and it’s likely that “something better will come along” in the next few years. But that “new thing” will likely be a highly capable software modem that will connect the same way as the MMDVM, include all the existing MMDVM functions including MMDVM-TNC, and much more.

• And much more, including some examples of what we can do with a real Amateur Radio data infrastructure. Hint — look at what’s already being done with user-accessible Amateur Radio microwave networks such as HamWAN and AREDN.