# least GSM power in dBm required to send an SMS

I deploy an application in the forest using an M2M module (SIM968). I have read through its datasheet but I am not able to find out what is the least GSM power that should be incident inside the forest to perform a simple SMS transmission or a GPRS transmission. (I guess both use 2.5G technology, please correct me if I am wrong)

Attached is the 2.5G coverage map of the network provider and I am interested in knowing if I would be able to transmit am SMS or perform a very small (of 10-15 bytes) TCP/IP transmission when the coverage is at -105dBm.

p.s.

I had a conversation with a network engineer of a respective provider, he advised their cut off point is at -102dBm.

• Depending on where you are, the least power required is the maximum power your device can offer. – PlasmaHH Feb 16 '15 at 14:01
• I am talking with regard to incident GSM radiation power. Not the power m2m can liberate. – Denis Feb 16 '15 at 14:25
• @QwertyCoolGuy I think PlasmaHH is right, as you are trying to use the upstream link, it will highly depend on the output power your device can offer. You would have to test it in the forest. – gstorto Feb 16 '15 at 15:07
• @kurtovi Exactly what I am doing right now :) But I need to include a proper analysis of GSM coverage in my research paper. – Denis Feb 18 '15 at 11:04
• I suspect the colors on that chart are akin to contours, i.e. in the middle of one of the white regions, the power available could be significantly lower than -105 dBm. Also, tree coverage will affect the signal on a local scale. – pericynthion Feb 19 '15 at 3:34

"I had a conversation with a network engineer of a respective provider, he advised their cut off point is at -102dBm."

That is indeed the minimum receiver sensitivity figure for a class 4/5 GSM900 mobile station in GMSK mode. If you wish to read the full performance specs, you may consider downloading the 3GPP TS 45.005 pdf (Old ETSI TS 145.005 ref). 3GPP download.

Chapter 6 deals with receiver performance. Remember that a test if the receiver conforms is typically done in a lab environment with a radiocomtester like an R&S CMU200 or Agilent 8960. These instruments simulate a GSM base station including signal fading and (Gaussian) noise conditions like present in a real-life situation (patterns & figures specified in the document above). Nevertheless real conditions can still be worse.

Without knowing the actual system details you can always have a very coarse performance estimate of a random radio receiver with the following equation:

$$L = - 174 + 10.log(BW) + NF$$

• The BW of a GSM channel is 200kHz.
• A realistic receiver noise figure (NF) is 8dB

$$L = -174 + 53 + 8 = -112 dBm$$

Channel coding, SNR and other factors further influence this figure, either positive or negative.

• Question 1 : Since mobile stations need to abide my international laws, can there be a mobile station that uses not GMSK but any other? in that case, would out mobile phone know that modulation scheme? or is it compulsory that GMSK be used for all transmissions in GSM telephony? – Denis Feb 18 '15 at 4:35
• Question 2 : what type of a receiver is used to measure these dBm measurements? Because at a particular point, amount of GSM radiation in dBm is also highly dependent on my antenna characteristics too. Am I correct? Or do these values are assumed to be true for mobile phone size antennas? My tracker application going to be deployed in forest has a much better antenna. Does that mean I will have slightly better performance than depicted from this map? – Denis Feb 18 '15 at 4:48
• Could you please post a direct link to the GSM specification, I am having trouble finding the document in their website. thank you. – Denis Feb 18 '15 at 5:16
• See details in my next post. – Ambiorix Feb 18 '15 at 9:26
• You're welcome, but I think, I misunderstood your question 2. You're referring to the coverage level measurements instead of of MS receiver sensitivity levels, right? I explain this later, is done in a different way, obviously. – Ambiorix Feb 18 '15 at 10:49
• "Question 1 : Since mobile stations need to abide my international laws, can there be a mobile station that uses not GMSK but any other?"

You don't need to worry about this. The system determines this (GPRS protocol). On a radiocomtester (see below) you can set this manually but a public network doesn't allow that.

• Question 2 : "what type of a receiver is used to measure these dBm measurements?"

This is not a receiver but a radiocommtester like this one or this one. This instrument has all the functionality of a BTS and also includes a fading and noise generator capable of generating real-life conditions. This comtester can answer a call from, or initiate a call from the MS in the exact same way as a normal cellular network except that data is generally not speech/SMS but usually bit patterns (PBRS). To perform a receiver sensitivity test the MS is generally put in loopback mode via an AT command (sent via a PC through the serial port of the device). The MS sends the bit pattern generated by the radiocomtester back via the uplink. The radiocomtester compares the original bits with the received bits a calculates the BER, FER, etc. By reducing the output level of the comtester you can determine the level where the communication breaks down.

RF engineers who are still in the phase of developing a frontend also use what is called a Vector Signal Generator, Vector Signal Generator. This instrument can generate a GSMK (or other) signal modulated with a fixed data pattern or PRBS. They usually have their own way of determining the BER or FER since only a downlink path exists.

• "Question 2 : "Because at a particular point, amount of GSM radiation in dBm is also highly dependent on my antenna characteristics too. Am I correct?"

Yes, but such tests are typically done in what is called 'conducted mode' where the antenna of the MS is replaced by a small connector connected to the comtester via a short stretch of coax cable.

Radiated versions of such tests exist, but generally take place in a anechoic chamber, usually by a full conformance/performance test system way beyond the capabilities of an average application developer.

• "Question 2 :"Does that mean I will have slightly better performance than depicted from this map?"

No, the figures you see in the document are receiver input levels. An antenna is a passive device and can never amplify the signal. Antenna gain is not real gain but the relative gain with respect to an omnidirectional antenna. A better antenna factor can increase the level at the receiver input with respect to the prevailing field strength at that location, but never the field strength itself, of course.

• "Question 3: Could you please post a direct link to the GSM specification".

Finding a doc in there is indeed looking for a needle in a haystack. Here.

You can use one of the following devices as measurement receiver:

1. A standard mobile device with integrated measurement functions, like this one. You have to contact the manufacturer, Walmart won't have these :). Data can be read-out via USB or serial port, generally AT commands.
2. A specific measurement receiver like this one. Comes with LAN, RS232 or GPIB, usually via SCPI commands.
3. Some general purpose measurement receivers or vector spectrum analysers also have the ability to sync to a GSM power ramp, nevertheless set-up is far more complex and at least requires some understanding of the GSM/GPRS physical layer to perform a proper measurement.

You have to keep in mind that these devices measure the level (dBm/dBμV) at the receiver input. If you wish to obtain the actual field strength (dBV/m) you have to scale the measurement values with the antenna factor and path losses. This article explains the conversion. Plenty of other, more comprehensive sources, out there.

If you wish to take the cable/connector loss into account either take the insertion loss data from the manufacturer (cable, connectors) or better, calibrate the cable with a RF generator and RF power meter or Spectrum analyser.

• Thank You @Ambiorix I will read these carefully and get back to you if I get stuck at an unclear place I hope you would be able to help. thanks again. – Denis Feb 19 '15 at 3:01