CW testing

[ppsec123]

does any one have any information on the procedures involved in CW testing?
is it dont using TEMS or any other tool?

please help.

[s52d]

does any one have any information on the procedures involved in CW testing?
is it dont using TEMS or any other tool?

please help.

Hi!

CW testing means you put on carrier, and measure coverage.
We do it to assure proper coverage where we want it, and no interference elsewhere.
It is worth doing if we do not trust coverage predictions, for example indoor or in
mountains.

TEMS transmitter is nice: just small terminal, where you can program BCCH etc.
Low output power makes it suitable for indoor measurements and all sorts of
nice experiments. We even used it to find intermodulation problems.
For rural, 20 W TX is needed, and it is worth paying attention to proper frequency stability
if you want to use GSM terminals for measurements.

Plain measurements are made by any radio receiver capable of measuring signal strength.
For GSM, we like to modulate it as it is BCCH carrier with specific BSIC.
Make sure BA lists are properly tuned (list of NCC/BCCH to be monitored by terminals).

Thus, TEMS can decode it.
We often run BA list recording on it (reports from BSC), so all terminals are reporting coverage.
Some overshooting and indoor details are reveiled, normally not reported by TEMS.

In UTRAN, it is less used: probably because we already learned a lot on GSM.
Anyhow, you need to do it on spare ARFCN, otherwise you get a lot of dropped calls
and complains for bad coverage.

BR
s52d

[ppsec123]

thanks s52d...

How do you analyze this CW data? do you put this into a post processing data ? need some help in using CW data for model tuning..

[s52d]

thanks s52d...

How do you analyze this CW data? do you put this into a post processing data ? need some help in using CW data for model tuning..

Hi!

Now, model tuning is a story in itself: you need high quality receiver to get proper data,
(TEMS or other terminals can do, but I do not trust them to have good enough
accuracy and dynamic range for real work). Each radio prediction SW has its own
way how to tune it: you have to prepare data in proper format, and follow
your manual.
In short: given formulas, you want to tune constants in order to minimize standard
deviation of difference between prediction and measurements.


On testing of coverage where prediction tools, even if tuned properly,
can not be used:

For indoor: just monitor with TEMS terminal and verify coverage.
Thus, you can verify coverage even if details needed for calculation are not known (walls material etc).

When testing for new site, we use standard Mobile Measurement Report processing.
With E///, data formats are simple and well documented, so we just grab them and
look at the data for site specific questions.

(Maily: which cells could be interfered and how much traffic new cell might take)
As it is time consuming, we are not doing this often: but sometimes we simpy have to test
in order to get coverage where we want, and no interference where it should not be.
In hilly terrain with reflections Okomura-Hata based predictions are not good enough for details.

So, no fancy-GUI-$$$-SW, just some scripts, maps and common sense.

BR
s52d

[RF engineer]

does any one have any information on the procedures involved in CW testing?
is it dont using TEMS or any other tool?

please help.

If you donot have any CW software you can do it by TEMS investigation just you need for the sites that CW is performed are Disable Power control, Disable DTX and Hopping then do it by TEMS which gives you option to disable HO with the neighbor .
BR
RF

[ppsec123]

Hi!

Now, model tuning is a story in itself: you need high quality receiver to get proper data,
(TEMS or other terminals can do, but I do not trust them to have good enough
accuracy and dynamic range for real work). Each radio prediction SW has its own
way how to tune it: you have to prepare data in proper format, and follow
your manual.
In short: given formulas, you want to tune constants in order to minimize standard
deviation of difference between prediction and measurements.


On testing of coverage where prediction tools, even if tuned properly,
can not be used:

For indoor: just monitor with TEMS terminal and verify coverage.
Thus, you can verify coverage even if details needed for calculation are not known (walls material etc).

When testing for new site, we use standard Mobile Measurement Report processing.
With E///, data formats are simple and well documented, so we just grab them and
look at the data for site specific questions.

(Maily: which cells could be interfered and how much traffic new cell might take)
As it is time consuming, we are not doing this often: but sometimes we simpy have to test
in order to get coverage where we want, and no interference where it should not be.
In hilly terrain with reflections Okomura-Hata based predictions are not good enough for details.

So, no fancy-GUI-$$$-SW, just some scripts, maps and common sense.

BR
s52d


that is true..i do have some data but i dont have any big tool...the coverage from the CW data is good using which i can design site placement..so i am not sure where the model being tuned?

thanks all for this discussion