In what context? For TTN there is no recommendation/agreement - basically not used for the community network. In the context of LoRaWAN and The LoRa-Alliance there is a specification for use of the EU433Mhz band (e.g. RP002-1.0.3) but no specified channel plan as such and per 2.7.9 of the above
“There are no specific default settings for the EU 433 MHz ISM Band.”
Only a requirement under 2.7.2 a requirement for:
"The LoRaWAN® can be used in the 433.05 to 434.79 MHz ISM band in ITU Region 1 as long as the radio device EIRP is less than 12 dBm. The end-device transmit duty-cycle SHALL be lower than 10%.The LoRaWAN® channels center frequency can be in the following range:
• Minimum frequency: 433.175 MHz
• Maximum frequency: 434.665 MHz
EU433 end-devices SHALL be capable of operating in the 433.05 to 434.79 MHz frequency band and SHALL feature a channel data structure to store the parameters of at least 16 channels. A channel data structure corresponds to a frequency and a set of data rates usable on this frequency. The first three channels correspond to 433.175, 433.375 and 433.575 MHz with DR0 to DR5 and SHALL be implemented in every end-device. For devices compliant with TS001-1.0.x, those default channels SHALL NOT be modified through the NewChannelReq command. For devices compliant with TS001-1.1.x and beyond, these channels MAY be modified through the NewChannelReq but SHALL be reset during the backoff procedure defined in TS001-1.1.1 to guarantee a minimal common channel set between end-devices and gateways of all networks. Other channels can be freely distributed across the allowed frequency range on a network per network basis."
(Note: final sentence emphasis mine - no standardisation? Also note need for 16 channels vs simpler 868 on 8, I believe the additional channels needed as a potential mitigation against (2) below?)
So you can use if you want to on your own network, IIRC TTS in its various guises except deployed TTS(CE) the TTN Community network can support, but think you have to define your own plan.
A few observations
-
Market for 433Mhz much more limted so choice of COTS sensors and nodes more restrictive - note not all will have fiemware that meets outlines requirement or will have bothered with LoRaWAN certification.
-
433Mhz band often much more noisey that other bands - fewer available channels and long legacy of e.g. key fob, alarms/controllers/garage door openers, central heating systems etc,., especially in more urban environments, etc.
-
antenna length for 433Mhz = 2x that for 868Mhz hence much bulkier implementations
-
Max TX power - EIRP - lower for 433Mhz vs 868Mhz, hence small benefit of increased free space range and potential built environment penetration is then offset/lost - a zero sum game?
I want to use it for long distance. See TTN have as option to configure the band when you put new end device. Loriot have option with 5 channels.
See (4) above!
Google ‘free space loss’ for the various frequencies and you will see the relative signal fall offs as frequencies increase. e.g. (IIRC) moving from the ~900Mhz bands to ~2.4Ghz you loose ~10db of signal. LoRa gives a ‘coding’ gain vs conventional (legacy?) modulations such as FSK of ~2.5db per SF step (again IIRC) so we see a typical range from ~7 to ~20db (at SF12) gain, meaning that if you use LoRa at higher frequency you can offset the free space loss, similarly if you go down in freq, the loss decreases and you get better signal, with potentially longer range.
All this is at constant TX power…and the problem you have at 433Mhz vs 868 is that where the EU868 limit is 16dbm EIRP, as above you are limited on TX power at 433Mhz - by approximately the same reduction as the reduction in FSL. So you dont get actual significant realworld change in open LOS range. Therefore, conclusion would be stick with 868 'cause e.g. (1) through (3)! You might see increase built environment penetration from the lower frequency…one of th reasons why meter applictions - especially water metering, with meters in pits -or applications where the node (meter) is deep in a building will often see use of e.g. 169Mhz, that is where range as a proxy for penetration starts to break down.
As noted though at 433Mhz the potential small improvement in penetration might actually be offset by the degradation in SNR from the noiser environment in the urban/built environment - YMMV!
Disclaimer: FWIW I use and have used 433Mhz on LoRa P2P in a few instances with ok, but not stellar results, but I dont touch for LoRaWAN 
Then you can submit a proposal for a frequency plan to TTN or use your own backend. Keep in mind the existing network uses 868/915MHz so you won’t be able to use any other parts of the community network if your proposal is accepted and made part of the TTN setup. You will also be the cause of yet another VHS vs V2000 dilemma for users…
Long term TTN users that already invested a lot of resources and effort into the existing frequency plans (gateways are cheaper these days but still not free) might not encourage splitting the community by using a currently non standard frequency.
you live in europe not same as India we far from towns
australia use 2 band
So you need to investigate any local interferers and other close proximity RF users for the band…We are obviously commenting in the context of OP and thread subect - Europe! with local experience 
Have some prior customer history with India (again largely from smart metering perspective and there are a few gotchas IIRC). With lower urban deployment you ‘may’ find (2) less of an issue for your specific case, but in populous areas you may need to conduct RF surveys to determine actual situation and viability… again YMMV!
You also need to investigate the 433Mhz band limits wrt TX power in your area… again we have discussed in EU433 context vs any IN equivalent (IN865 & EU868 close enough for condideration of 1st order effects in this context).
And that has been an issue for users from day one. With the move to TTN v3 it was decided to harmonize on one band to try and resolve the issues.
Actually just checked those same L-A docs and NO 433Mhz option - its:
But that didnt work if I recall - was on some of the early discussions where harmonization was proposed and discussed - I think they still carried on with dual deployments based on AU915 & AS923 options - 'cause many were already invested as you noted above and there is then reluctance to change… havent heard much on that lately as the group discussions collapsed to AU locals only with us remote parties dropped from the invites
Think for wider Indian Sub-Continent some EU433 options (but no predefined plans) - Bangladesh, Pakistan and (IIRC) Sri Lanka, but not India? - 
military considerations per other territories?
If you wish to rebut a comment it would be polite only to do so if the person was aware of the facts at hand.
Your topic opener said Europe. Someone provides high value technical advice and you dismiss it on the grounds you think it doesn’t apply on the basis of information you did not disclose.
It also begs the question why on earth you are interested in a channel plan that does not cover your jurisdiction. If you go and get permission from your authorities, ie they enact it in law, then please feel free to come back and request the channels to be included.
Under the right circumstances you do get longer distances using 433MHz, I use it for High Altitude Balloons, but it’s not on LoRaWAN. If you want to pursue LoRa P2P, this is not the forum for you and you will have to ask in-country if the frequencies are permissible.
Because of the differance butween the ERP limits, 10dBm @ 434Mhz and 14dBm @ 868Mhz, the range differance is small. A 434Mhz system would only go around 25% further.
And given that the 434Mhz band can have quite a lot of ‘interferance’ from other users, you might actually get less range, not more.
It would make sense to carry out real world tests, at the legal limits, of the actual achieved range differances, and not just assume the change is worth all the effort.
2 Likes
I have both systems in use. One for LoRa-APRS on 433 MHz, the other one on 868MHz for LoRaWAN. The antennas of the gateways are nearly on the same position.
I can confirm that there’s no significant difference in coverage between 433MHz with 14dBm (I am allowed to use more than 10 dBm) and 868MHz at 14dBm erp.
2 Likes
Yep, basically what I said
I was @sofia2019 who opened, and hasn’t commented since. @Piet_Pillay ‘contributed’, guess interested in same potential, but advice given, comparatives provided by others, real world is where rubber hits the road…. Or EM signals hit the atmosphere….so down to users to take advice, ignore or experiment for their own benefit/interest….we can but comment and put ideas forward 
if 433 and 868 on distance 20km and same transmit power and same antenna gain you have loss of -111 on 433 and -117 on 868
what difference it make on SF with 6db difference? ignoring fact of SNR in this.
not ignoring you, 868 use here as much as 433 where i live we have very ffew new cars or garages that us 433 transmitters
Yes, but remember\quote all of what what I said earlier, in most places you are limited to 10dBm on 434Mhz and 14dBm on 868Mhz, so the practical difference is only 2dBm, not 6dBm.