Radio not activating local FM repeater (small subtone deviation)

[OK2MOP]

Hello,
I just wanted to share here the findings about a local repeater which works with many other RDA1846 radios (mostly the FM ones) but did not activate on both OpenMDUV380 and stock firmware in RT3s, DM-1701 and Ailunce HD1 versions (stock firmware only). Since it did work with DM-X (MD1702) stock firmware, I had a look with disassembler and PDF datasheet on the registers and found the culprit. Interestingly, after the localization of the problem, I also found out that switching the OpenMDUV380 into wide FM mode actually activated the repeater (but it would break channel width specified by IARU).

To make the long story short, the culprit is register 0x59 lowest 6 bits setting in hardware/AT1846S.c. The firmware has its value set to 0x10, but in my case at least 0x13 was required, with save margin 0x14 was chosen. In stock firmwares, this value can be typically calibrated separately for different TX frequencies and different CTCSS tone frequencies, or at least centrally set to a specific value. Maybe it could be included in radio calibration screen as well (but this would require changes in several places of the code)?

For now, replacing:

Code: Select all

{0x59, 0x0B, 0x90},

on lines 60 and 108 of hardware/AT1846S.c with

Code: Select all

{0x59, 0x0B, 0x94},

seems to do the trick (or binary patching the bin files), wide 25kHz variant has already {0x59, 0x0B, 0xA0}, so it has value 0x20 for this setting and that is why it works. I am not sure if this change could reduce quality of the sound or have some other unintended consequences, otherwise it would be easiest to change it in the source.

[VK3KYY]

Interesting

I will need to check whether this value is read from the official calibration data in the radio.

[G4EML]

As you say it also do not work with the original firmware on several radios then it would tend to suggest that the repeater in question requires a higher than normal deviation for successful CTCSS decode. The fact that it worked when switching to 25KHz spacing would tend to support that. Are you sure that it is not (or used to be) a 25KHz repeater?

Are you able to measure the CTCSS deviation your radio is sending? The levels used in the OpenGD77 firmware were selected to produce the ETSI recommended deviation levels, so on 12.5KHz channel spacing it should be between 300 and 500 Hz. (It was tuned for 400Hz). Any repeater should be able to reliably detect signals within that range of deviations. In practice most can detect much lower levels, so it is rare for tone deviation to be an issue.

Colin G4EML

[OK2MOP]

Hello,
I cannot check the setting of the input radio in the repeater (OK0D), output radio is clearly set to 12kHz or less. It is with high probability intentionally set to lower sensitivity because it is on the highest hill in the neighbourhood and several tens of meters from a TV tower with 100kW+ DVB-T2 MUXes and a multipoint communication relay for many technologies. So there were big problems with spurious repeater input activations by some digital noise in past.

If the value is based on some ETSI suggested value, then I would not modify it and if someone needs to, than it can be patched (I did not find any documentation about the calculation of spreading Hz from the register values). Actually, it also probably depends on higher bits for Voice+subtone spreading which differs per radio type in the sources. It may also be influenced by noise introduced by bypassed low and high pass filters in OpenGD77 firmware (register 0x58). Also I do not know if RDA1846 generates square, saw or pure sine wave.

I will try to measure the deviation with SDR and dummy load but it will be approximate.

BTW, the factory firmwares (say of HD1) have the subtone setting between 13 and your 16 but the common value can be higher (RT3s uses some other calibration value somewhere between 80-120).