DIY BTS

This is merely a historical archive of years 2008-2021, before the migration to mailman3.

A maintained and still updated list archive can be found at https://lists.osmocom.org/hyperkitty/list/OpenBSC@lists.osmocom.org/.

David A. Burgess dburgess at jcis.net
Mon Nov 29 17:25:40 UTC 2010


We have tried a half-duplex BTS with OpenBTS, only transmitting on 3  
slots to avoid a duplexer, and it did not work very well.  Most  
handsets would not camp to it, even when it was the only signal  
available.  So you need to deal with the frequency duplexing problem  
and the antenna switch that you find in a typical MS will not work.   
There are also some small-but-important important changes in L1 and  
L2 that might not be possible in a highly integrated, optimized product.

On Nov 29, 2010, at 12:29 AM, Hans Witvliet wrote:

> On Sun, 2010-11-28 at 23:49 +0100, Sylvain Munaut wrote:
>> Hi,
>>
>>> My first reaction was one of sceptism, but according to the  
>>> people of
>>> huwei it is do-able, no hardware changes needed, just some (major)
>>> software rewrite.
>>
>> With a single dongle ? Sounds weird.
>>
>> The "client"/MS side of GSM have no duplexer but a TX/RX switch and a
>> BTS needs to be able to TX and RX at the same time.
>> (to be easily detectable by phones. It could 'possibly' work without
>> it and phone would stay camped onces they are there but it could take
>> several try for them to 'see' the network).
>>
>
> Indeed, i am still skeptic: seeing is believing.
> Got my HF-licence since 1978, so i know that it is very hard to do tx
> and rx on nearby frequencies. One needs extreme good (cavity) filters.
> Without those, even separate transmitters and receivers will not be
> sufficient, as your receiver input stage will be completely saturated.
>
> Eventhough "normal" bts must be capable of handling multiple  
> handhelds,
> while these dongles needs only to communicate with just a single  
> one, at
> extreme low power, there is not enough space inside for decent  
> filters.
>
> Hans (PE1CXJ)
>
>


David A. Burgess
Kestrel Signal Processing, Inc.








More information about the OpenBSC mailing list