What dBm actually means on a signal meter
Signal bars are a marketing decision. There is no standard that says how many bars correspond to how much received power, so the same spot in the same room can show four bars on one handset and two on another sitting beside it. The number underneath is not a matter of opinion: dBm is an absolute unit of received power, defined the same way by every chipset in every network. Once you can read it, the bars become what they always were — a rough summary drawn by the phone's manufacturer — and you get to work with the measurement itself.
A logarithmic unit, and why it is negative
The unit means decibels relative to one milliwatt. Zero dBm is exactly one milliwatt of received power. Because radio power arriving at a handset is a minuscule fraction of that, every value you will ever see indoors is negative, and the closer to zero the stronger the signal. Minus eighty is far better than minus one hundred, which trips up almost everyone at first. The scale is logarithmic, which is what makes it usable: an increase of three decibels means double the power, ten decibels means ten times the power, and twenty decibels means a hundredfold. A move from minus one hundred and ten to minus eighty is not a thirty per cent improvement, it is a thousand times more energy landing on the antenna.
The range that matters in practice
A phone standing under a mast can see minus fifty or better. The floor of usefulness sits somewhere near minus one hundred and twenty, where the wanted signal disappears into thermal noise and the radio can no longer decode anything reliably. Everything interesting happens in the seventy decibels between those extremes, and the practical thresholds are surprisingly sharp: voice survives on very little, while high throughput needs both a strong signal and a clean one. That is why a value near minus one hundred often still carries a call perfectly but makes a video stall — the modem drops to a slower, more robust coding scheme long before the connection fails outright.
Reading the bands
The table below gives the working thresholds for reference signal received power on a modern network. Treat the boundaries as soft: a value sitting on a line will behave like the better category on an empty cell and like the worse one at eight in the evening in a crowded street. One more caveat belongs with the table: the same figure means different things on different frequencies. A reading of minus one hundred on a low band that has already fought its way through two concrete walls is a healthier situation than the same reading on a high band with a clear view of the mast, because the low band still has margin left for the next obstacle while the high band does not. Read the number together with the frequency whenever you have both.
Why two phones disagree
Three effects explain almost every disagreement. Antenna design differs between models, and a couple of decibels of gain in the same casing position is entirely normal. Reporting differs too: some chipsets average over a longer window, so one phone reacts to a step behind a wall in a second while another takes five. And the bars mapping is unique to each manufacturer, so the same minus ninety-eight can be drawn as two bars or three. Comparing raw dBm between handsets held in the same hand, at the same height, facing the same way, is the only comparison worth making.
The other numbers on the same screen
Older networks report a plain wideband power figure, so on third and second generation service you will see a received strength that includes everything in the channel. Fourth and fifth generation service reports reference power instead, measured only on the known pilot symbols, which is a cleaner measurement and typically reads about twenty decibels lower than the wideband figure for the same conditions. That offset is the single most common source of confusion when someone compares a screenshot of a 3G reading with a 4G one. There is also a small integer called the arbitrary strength unit, kept for compatibility, which on modern networks is simply the reference power plus one hundred and forty.
Turning the number into a decision
The practical use of an absolute number is comparison across places and times. Stand where the call dropped and note the value. Walk to the window and note it again. If the difference is ten decibels or more, the fix is furniture, not the operator. If the whole flat reads within a few decibels of the same poor value, no amount of moving will help and the answer is Wi-Fi calling or a femtocell. If the reading is strong but the connection is still slow, the problem is not strength at all but quality and load, which are measured by different fields entirely.
| Reference power (dBm) | Verdict | What it feels like |
|---|---|---|
| −44 to −80 | Excellent | Full rate data, instant call setup, video without buffering |
| −80 to −90 | Good | Everything works; throughput depends more on how busy the cell is |
| −90 to −100 | Fair | Calls fine, downloads slower, occasional stalls on video |
| −100 to −110 | Poor | Voice usually holds, data crawls, handovers become noticeable |
| −110 to −120 | Very poor | Edge of coverage; drops, long call setup, frequent loss of data |
| Below −120 | No usable service | Signal is at or under the noise floor |
Frequently asked questions
Is minus 70 dBm better than minus 100 dBm?
Yes, by a very wide margin. The scale is negative and logarithmic, so minus seventy is a thousand times more received power than minus one hundred. Anything closer to zero is stronger.
What is a good signal strength number to aim for?
Better than minus ninety is comfortable for everything. Between minus ninety and minus one hundred is usable but sensitive to load. Below minus one hundred and ten you are living at the edge of the cell and should expect problems under any traffic.
Why does my phone show full bars but the internet is slow?
Bars follow strength alone. A strong signal on a congested cell, or one drowning in interference from neighbouring cells, gives poor throughput. Quality and signal-to-noise fields describe that, and they can be bad while strength is excellent.
Why is my 3G reading so much stronger than my 4G reading?
They measure different things. The older reading is total power across the whole channel; the newer one is the power of reference symbols only. For identical conditions the newer figure typically lands about twenty decibels lower, which is normal rather than a loss of coverage.
What does the ASU number mean?
It is a small integer kept from older interfaces. On modern networks it maps directly onto the reference power: add one hundred and forty to the reference figure in dBm and you have the arbitrary strength unit.
How many bars is minus 95 dBm?
There is no answer that holds across handsets, because the mapping is chosen by each manufacturer. That is precisely why a meter shows the number instead of redrawing someone else's bars.
Does holding the phone differently really change the reading?
It can, by several decibels. Covering the antenna strip with a palm, resting the phone on metal or sliding it into a pocket all attenuate the signal, and a meter makes the effect visible while you experiment.
Why is the strongest value I ever see around minus 50?
Because that is roughly what standing under a mast with a clear path gives you. Values closer to zero belong to laboratory conditions with a cable rather than an antenna, and a phone reporting them in a living room is reporting a fault.
See what your phone's radio actually reports. CellBar shows live signal strength in dBm, the network generation, the band and downlink frequency behind the channel number, the serving cell and the neighbours around it. Nothing is invented and nothing is exaggerated, and no app can make a signal stronger. Knowing the real number is what lets you move two metres and fix the call.
Get CellBar on Google PlayAbout CellBar