Short Range Device in EU, which frequency band?

[Mad ID]

Hi,
I'm working on a smart meter device which should communicate over ISM band. From what I see the simplest solution is to use a Non-Specific Short range Device frequency bands.

868-868.6 / +14dBm / <1% D.C. is what I started with but we will have a problem with 1% duty cycle since we won't be able to transfer all of the data.

Another interesting one is
433.05 - 434.79 / +10dBm / <10% D.C.

Is there any catch here? This looks much better with 10% duty cycle. Can we expect much worse interference with other devices in this band? Larger antenna is something we can manage.

Another question, what about ERC 70-03 Annex 2 TRACKING, TRACING AND DATA AQUISITION
It states that there are specific bands for meter reading, but I don't understand what this means for us during the certification? I see that APC is required, but cannot find how to implement it.

Any information on this matter is appreciated.

Thanks.

[tggzzz]

Can we expect much worse interference with other devices in this band? Larger antenna is something we can manage.

In any ISM band you have to accept 100% interference without complaint.

Larger antennas won't help with some of the effects observable in radio channels.

[Kilrah]

we will have a problem with 1% duty cycle since we won't be able to transfer all of the data.

Really? Do you have that much data that you can't spread the transmission?

Is there any catch here? This looks much better with 10% duty cycle. Can we expect much worse interference with other devices in this band?

There are tons more 433MHz devices in use than 868 for sure.

[Mad ID]

Really? Do you have that much data that you can't spread the transmission?

1% duty cycle means 36 seconds / hour. We will be using LoRa modulation where bit-rate is low when configured for good link budget (high sensitivity). We need to transfer up to 200kB/hours which is not possible within 36 seconds.

A good info is that 433MHz is much more crowded..thanks.

[Kilrah]

OK.

BTW the 1% requirement only applies to "dumb" transmission, it can be avoided if a protocol implemeting listen-before-talk and frequency hopping scheme is used.

Summary here, but you might want to look at regulatory office docs for more reliable info:
https://en.wikipedia.org/wiki/Short_Range_Devices#SRD860

[Jeroen3]

Where does it need to communicate to? If you need a reliable but relatively fast link you could use the cellular network.

[G0FTD]

>Another interesting one is
>433.05 - 434.79 / +10dBm / <10% D.C.

Just be prepared for your device to have problems then.

That frequency segment has a lot of radio amateurs on it, who can run as many kilowatts ERP as they want most of the time, so the receivers of your device and are gonna have some very strong signals obliterating stuff.

And I use that band a lot ;-)

Better off on the 868Mhz band.

Regards,

Andy

[Mad ID]

>Another interesting one is
>433.05 - 434.79 / +10dBm / <10% D.C.

Just be prepared for your device to have problems then.

That frequency segment has a lot of radio amateurs on it, who can run as many kilowatts ERP as they want most of the time, so the receivers of your device and are gonna have some very strong signals obliterating stuff.

And I use that band a lot ;-)

Better off on the 868Mhz band.

Regards,

Andy

That's what I though, thanks

[CJay]

Why not use devices like the Nordic NRF 2.4GHz ones or am I missing something obvious?

[Mad ID]

Why not use devices like the Nordic NRF 2.4GHz ones or am I missing something obvious?

You are, with LoRam devices can communicate from the basement to the roof through the whole building. With nRF you can communicate with the other side of the wall

[CJay]

Why not use devices like the Nordic NRF 2.4GHz ones or am I missing something obvious?

You are, with LoRam devices can communicate from the basement to the roof through the whole building. With nRF you can communicate with the other side of the wall

Ah, cool.

Can you point me at a description or FAQ about it so I can learn?

[Mad ID]

Ah, cool.
Can you point me at a description or FAQ about it so I can learn?

Sure, you can learn a lot from IC manufacturers like here:
http://www.semtech.com/wireless-rf/lora.html

Sensitivity can be up to -137dBm.

[harry4516]

the 868 MHz is split into various segments.
For this application you can use 869,4 to 869,65 MHz where a duty cycle of 10% is allowed (500mW max)
or 869,7 to 870 MHz where 100% TX is allowed (5mW max).

I use these bands and they are usually totally free of noise, even in big cities.

see also:
https://en.wikipedia.org/wiki/Short_Range_Devices
and scroll down to SRD860

I am using certified modules from Mipot, an Italian manufacturer.

[Mad ID]

the 868 MHz is split into various segments.
For this application you can use 869,4 to 869,65 MHz where a duty cycle of 10% is allowed (500mW max)
or 869,7 to 870 MHz where 100% TX is allowed (5mW max).

I use these bands and they are usually totally free of noise, even in big cities.

see also:
https://en.wikipedia.org/wiki/Short_Range_Devices
and scroll down to SRD860
I am using certified modules from Mipot, an Italian manufacturer.

Hi, we will use the 869.4 - 869.65 and have a 125kHz LoRa channel on it. 10% duty cycle gives us 3x more data / hour than 1% duty cycle with 500kHz bandwidth.

Thanks.

[Mad ID]

the 868 MHz is split into various segments.
For this application you can use 869,4 to 869,65 MHz where a duty cycle of 10% is allowed (500mW max)
or 869,7 to 870 MHz where 100% TX is allowed (5mW max).

I use these bands and they are usually totally free of noise, even in big cities.

see also:
https://en.wikipedia.org/wiki/Short_Range_Devices
and scroll down to SRD860

I am using certified modules from Mipot, an Italian manufacturer.

Hi, In the ERC/REC 70-03 Appendix 5 defined duty cycle categories. For <=10% they say maximum transmitter on time of 36sec. I cannot find anywhere what is the off time?

In some 10years old app noted from TI and similar I find that it was 3.6sec back than, has this changes and if not, where is it defined?

Thanks.

[Kilrah]

I cannot find anywhere what is the off time?

Uh that's determined by your 10% duty cycle. You have to transmit at most 10% of the time, so if you used the longest transmit "block" you can of 36sec you then have to wait for 360-36 = 324 sec before transmitting again.

[Mad ID]

I cannot find anywhere what is the off time?

Uh that's determined by your 10% duty cycle. You have to transmit at most 10% of the time, so if you used the longest transmit "block" you can of 36sec you then have to wait for 360-36 = 324 sec before transmitting again.

Hmmm this cannot be right, 324 is too long. And if you use the smallest like 1ms then I would have to wait 359.999 sec

It something like 3.6sec of even 100ms I found at various places.

[PA0PBZ]

Hmmm this cannot be right, 324 is too long. And if you use the smallest like 1ms then I would have to wait 359.999 sec

No, if you transmit for 1ms you have to wait 9ms to get to the 10% duty cycle.

[Mad ID]

Hmmm this cannot be right, 324 is too long. And if you use the smallest like 1ms then I would have to wait 359.999 sec

No, if you transmit for 1ms you have to wait 9ms to get to the 10% duty cycle.

OK but what about this case.

I need to send 10 500ms packets. The minimum time between two packets is say 5ms for my system. Can I send them successive order all at once as fast as possible and then think of that time as being 5seconds long. Another burst could come after 45 seconds.

So in short, in the 5sec interval duty cycle would be 98%. Will this satisfy the standards? (I cannot find this information)

Thanks.

[Kilrah]

Just had a look at the doc and it's very clearly specified that the "observation period" is an hour, so for a 10% DC you may transmit for max 360 seconds per hour, whether in 10 36 second blocks (longest allowed ON time in one go), 3600 100ms blocks, 1000 360ms blocks, you get the drift and they can be anywhere during that one hour period.

EDIT: Typo. So yes your case would work.

[PA0PBZ]

... so for a 10% DC you may transmit for max 36 seconds per hour ...

 

[Mad ID]

... so for a 10% DC you may transmit for max 36 seconds per hour ...

 

It's a typo I guess

This is all very simple but I still don't have my answer, can I transmit 500ms - 5ms pause - 500ms - 5ms pause for 36 seconds and then wait 324sec.

This would allow me to save battery. 10% D.C. at all time would be 500ms - 4500ms pause 500ms - 4500ms.

What say you? Thank you for your time.

[PA0PBZ]

Forget about the 5ms pause, if you transmit for 36 seconds and then pause for 324 seconds you are on a 10% duty cycle, so yes you can do that.

[Mad ID]

Forget about the 5ms pause, if you transmit for 36 seconds and then pause for 324 seconds you are on a 10% duty cycle, so yes you can do that.

My PHY will turn the carrier off after it sends the full buffer, CPU will need few ms to prepare the new data.

OK, I'll transmit for 36 seconds with small pauses.

Thanks

[Kilrah]

And remember that by implementing the "LBT + AFA" techniques you could send pretty much continuously as much as you want.