What does BER really indicate?

[kc7ve]

I know the definition, so don't just repeat the definition. I'm actually an Electrical Engineer, so I generally understand most of this.

However, I'm confused about BER and how it pertains, or is relevant to Pi-Star hotspots and DMR. Why am I confused, you ask?

1) Well, when I transmit I ALWAYS have some percentage, other than zero. Why is my BER non-zero and nearly everyone else is 0.0%. No, I doubt mine is any different than anyone elses.

2) For almost everyone else, on the network, they are coming across with 0.0% BER, even when they are unintelligible. How is that possible they have 0.0% BER, but can't be understood?

3) And, occasionally, there are other times, when they are near perfectly intelligible, yet might show some BER. How is that possible?

4) I have two hotspots, and 2 radios. When I transmit on one hotspot, I might see something like 0.7% BER, but on the receiving hotspot, it shows 0.0% BER. I can switch radios, and do the same test, and get the same result. So it would seem the issue isn't the radio. Why do I get 0.7% on one, and 0.0% on the other, for the same transmission? It is starts out with 0.7% shouldn't it be received with the same errors?

5) If I see 0.5% on my transmission, can others who receive it see 0.0% for that same transmission? I ask because of #4 above.

And then there is the "how to use it?". I have no information on the offsets for my boards. But I have read that you can improve the BER by changing the offset, per the manufacturer (but I don't have the offset info). But when I make any changes, to the offset (RX or TX), it makes virtually no difference. I can change it incrementally, until my hotspot no longer sees my radio signal. Yet the BER doesn't change.

Can someone explain some or all of this? And it's not packet loss, since in all of these examples, the packet loss is reported as 0.

[G8SEZ]

The BER is recalculated at each stage of the process I think, so when your hotspot receives from your radio it notes and displays how many bits needed correcting then passes on corrected data to the network layers, which someone else sees as 0.0% as it goes through to their hotspot dashboard and the modem but the radio at the other end of the link may still hear some errors and possibly correct them so you can still hear some glitches in the audio. Naturally if the HAT cpu cannot correct the errors it still passes on uncorrected data but the MMDVM has not way of knowing this if there is not an explicit BER indication field and in any case it can't correct it even if there is.

If you really want to dive deep into this then you need the DMR specs and the code in the HAT cpu and the MMDVM and server code to allow you to see how the data stream is treated in each direction.

I think my explanation is accurate enough without doing that, it's based on what I've picked up over the last 5 years or so of using these systems.

[kc7ve]

That's kind of what I figured but 1 issue wasn't addressed and I'm not sure what to do about it.

Like I said, I don't know what the offset should be. I've tried to adjust it starting with small changes and gradually increasing it. But I don't see any changes in BER. I keep increasing the offset (both positive and negative) until the pi-star no longer sees my TX.

I would expect to see the BER change, until it's a it's lowest. But it always stays about the same 0.7%.