I’m in the US and I also use a multi-SIM. It is able to access the three major US MNOs: ATT, T-Mobile, Verizon. I have spent recent months investigating handover behavior. I manufacture cellular modems based on the Quectel EG91 module.
Let’s consider the two basic scenarios,
1. the cellular modem is deployed in a fixed location
2. the cellular modem is moving (e.g. in a vehicle)
and also the impact of the modem being in:
a. a rural region (relatively few cell, large cells)
b. in an urban or suburban region (relatively more, smaller cells)
Taken together, there are four situations: 1a, 1b, 2a, 2b
In the case of a stationary modem deployed in a rural region (1a) it is likely the modem will only be in range of one or two base stations from each MNO. Intuitively, one might suspect the signal strength at a stationary modem for any particular base station will remain nearly constant. In that case, if the modem is within two base stations and the signal from one is significantly stronger than the other, one would expect the modem to register on the base station with the stronger signal and stay there forever. If the two signals are of approximately the same strength, then it could register on either and might sometimes register on the one and sometimes on the other. Similarly, if the modem is within range of base stations of other MNOs, one might expect that one or another of the MNO’s signals would be significantly stronger than the others and the modem would register with the MNO with the strongest base station signal and stay there or, if there are several “strongest” signals it would sometimes register with the one and sometimes with the other.
Unfortunately, even for a modem deployed at a fixed location, signal strengths can vary significantly from moment to moment.
The question also arises: why do you want to connect to the cell with the strongest signal? For example, if one base station is received at -68 dBm and another at -65 dBm (twice as strong), does it matter with which you register? Both are very strong signals.
Now, consider the case of a mobile modem in an urban or suburban setting (2b). You might think handoff happens primarily when a mobile modem is moving away from the cell transmitter with which it is registered, so the signal is getting weak, and it is approaching another cell with a stronger signal. If that was always the case, one would expect, while driving along a straight road, the frequency of handoffs to be according to the spacing of cells, with the modem handed off from weaker to stronger cells along its path, but that’s not at all what happens.
When a modem is moving across or around an urban or suburban setting, at any given moment it is often within range of three or more cells of each MNO. Received signal strengths from base stations are highly variable, not because the power output of the cell transmitters is varying, but primarily due to “fading” effects. Multipath, for example, occurs when a receiver receives the same signal more than once because the signal arrives directly from the transmitter while it also bounces off one or more radio frequency reflectors (a bus, or building). The reflected signals travel a greater distance than the direct signal, so each arrive a slightly different times. Those signals may be perfectly in phase, in which case they will interfere constructively and strengthen the received signal, or out of phase, in which case they will interfere destructively and weaken the receive signal. Similarly, if a base station uses multiple antennas, and the signals from the antennas are aimed directly at the modem, they will interfere constructively, while at oblique angles they will interfere destructively. Multipath and relative antenna direction can both change dynamically and quickly for mobile devices. Consider, too, that cellular communications are line-of-sight, so hills and buildings and even nearby vehicles between the base station and cellular modem can attenuate signals significantly, and a vehicle in motion can have a clear line of sight to one cell transmitter one instant and encounter and obstruction a moment later.,
All of this, I hope, suggests that–especially in scenario 1b–the “strongest” signal can vary, moment by moment, from one one cell to another of any given MNO and from one MNO to another.
Now, suppose you know at every moment which MNO has the strongest signal. If you’re registered with an MNO that doesn’t have the strongest signal and you want to switch to the one with the strongest signal, what must you do? To force a switch, you must disconnect from the one and register on the other, which ensures a drop in connectivity… By the time you complete the switch, you may need to switch again.
There’s another wrinkle too… Handoffs as so far described, were to ensure coverage. That is, when the signal being received from a cell’s transmitter gets sufficiently weak, the modem must be handed off to a cell with a stronger signal or the connection will drop. But there’s another, more recently addressed (in LTE) reason for handoff: load balancing.
Every cell has a fixed amount of bandwidth which must be divided among those using the cell. If the cell is at its full bandwidth capacity, it cannot accept any additional users without dumping some current ones. Suppose, however, the MNO has several overlapping cells, and one of those cells is at or near its capacity while another has excess capacity. In that case, the LTE network may handover some users from the busier cell to less busy ones. Okay, suppose you know which MNO has the strongest signal and you try to connect to it, but that MNO has only one cell within range and that cell is at full capacity? After you disconnect from the cell you were attached to, you’ll fail to connect to the desired cell because it’s already fully loaded.
I’ve been investigating handover behavior recently. I manufacture my own cellular modems (based on EG91) and my own custom IoT device. I’ve also written the software in the IoT device and the custom servers they communicate with. I’ve written software to track and record all handovers. I’ve been investigating in Santa Fe, NM, population 90,000, varies from suburban to rural, and where the terrain is very hilly and cell towers are very short (to prevent blocking mountain views), so line-of-sight problems abound.
I record far more handoffs than I expected. Handoffs frequently take place while stationary (at a traffic light). When moving, it is not unusual to be handed of 3 or more times in 10 seconds. The majority of handoffs occur within 300 meters of the previous handoff. Handofsf frequently ping-pong between two or three cells, over and over, while traveling in a straight line…
Finally, I’ll mention that not all multi-SIMs behave similarly. MNOs generally do not offer multi-SIMs. Multi-SIMs generally come from MVNOs. Most multi-SIMs “steer” the cellular modem to a favored MNO (the one that costs the MVNO least). Some multi-SIMs allow the modem to be forced to a non-favored MNO, but some do not.
The bottom line, I think, is the idea of selecting the MNO with the strongest signal seems, at first blush, to be a good idea, but there are very practical obstacles, and I’m not sure it’s even a meaningful goal.