Hi, your timing is excellent. I am still playing with these and researching them myself, but this is what I have learned thus far. Many of these modem batteries from equipment have been becoming available. Most of the ones I see use Lithium ion batteries most often using 18650 cells. After watching DamonHD's use of Lithium batteries over the years, I just ordered up 75 of them as scrap to see if any of the cells are salvageable.
There are some general considerations for each of the individual Lithium ion cells connected in series and parallel to make up a larger battery.
1. The maximum voltage must not exceed 4.2V or damage to the cell may occur. In fact, leaving them sit at 4.2V is not very good for most of these cell designs from what I have read in the manufacturer's literature. Significantly exceeding this cell voltage may not only damage the cells, but in some cases may lead to venting, bursting, shorting and fire.
2. The cell may be damaged and permanently lose some capacity and cycle life if the voltage falls below 2.5V. In general, no battery chemistry I am aware of likes to be run dead and the deeper you cycle a battery the fewer cycles you will get in general. For these modem packs, the condition of the batteries will be degraded if they have been removed from equipment and the cells have gone flat while they are sitting in storage. From what I have read, (I have never allowed the cells I have to go flat so I do not have personal experience with this outside of laptop batteries) the longer they are left at lower voltages the more damage and less capacity they will have when revived.
3. They can only be charged and discharged safely at rates below a certain current which is dependent on the particular cell design. I have some cells from old power drill batteries that were originally only 1300 milli-Amp Hour capacity, but can are rated to discharge current at a rate of up to 10 Amps of current. The modem ones seem to be more for energy storage and often are around 2200-2600 milli-Amp Hour capacity and are only rated to discharge at one to a few amps. Charging rates are generally lower than discharging rates. I do not have empirical data yet to back this up, but from manufacturers literature it appears rather clear that charging them more slowly will increase the number of cycles that you will get out of the batteries. Charging or discharging them too fast will damage the batteries, and if excessive can cause them again to vent, burst, short, and catch fire.
4. Charging the cells below 0 Degrees and above roughly 55 Degrees Celsius can damage the cells which of course can also under some limited circumstances result in shorting and fires.
Now, for building a bigger battery out of those modem packs that you have, each of them will generally have a board built into the battery pack which will try to limit the voltage and current to or from each cell and may also limit the temperature at which the pack can charge or discharge for the above reasons. Connecting these in series and parallel as packs with individual boards can cause problems if some of the packs cut out and others do not. In general you are going to have to remove the batteries from the packs and build your bigger battery and add a battery management system often referred to as (BMS) for that pack.
For a 48V Ebike, for example, as you stated you plan to build, you may wish to connect 14 of these in series. That would give you a nominal voltage of 51.8V, a maximum of 58.8 though I would never go that high, and a flat battery at 35V, though I would again never run it that low. Most of the available power of the pack would be between 44 and 54 volts and that is probably the range I would try to keep it in, maybe even raising the cut off voltage to 46 if I did not need the minor extra capacity of discharging it deeper. This voltage range would work with most 48V systems designed to run with lead acid batteries. Now, if you just hooked 14 2500 milliAmpHour cells together you would just have a 2.5 amp hour roughly 48V nominal battery, which likely would not give you much range and might not safely provide enough current for your bike.
Connecting these cells in series and parallel presents other problems for you to overcome in designing your battery. These are magnified by having cells that are likely in differing conditions because of there prior lifetime of use. The first thing to do is to make some measurements of the condition of each of the cells.
I have been measuring how much energy they discharge between charged voltage and discharged in milli-Amp Hours (mAH). I have also been attempting a measurement of their internal resistance, and have been trying to make a measurement of their self discharge rate. To assist in this I did purchase a charger which has built in diagnostics, a Liitokala Liiengineer-500 model, which has a test function to measure the cell capacity in mAH, and the internal resistance. To ensure that they do not self discharge too fast, I measure the voltage after a couple of days of them sitting around after I test them.
If their voltage discharges to less than 4V I do not use them, if their internal resistance measurement exceeds 100 milli-Ohms, I do not used them, and if their capacity is less than 70% of the original rating I do not use them. These are ballpark parameters I use to toss out cells that I predict may cause me problems with my pack down the road. If I had a good digital thermometer handy, I would toss out any cells that get hot when charging or discharging on the tests as well.
Once you have selected cells, you will want each 3.7V section with parallelled cells to have as close to the same capacity to all of the others as possible given the cells that you have available. If they vary in capacity, then the voltages between the sections may vary considerably during charging and discharging, especially as the battery is nearing full charge upon charging and full discharge upon use. Additionally if you are connecting a larger number of cells in parallel you want fusing on each cell as a failure in one cell could dump all of the current that the paralleled cells can provide into that failed cell, which could again cause venting, bursting, and fire.
Once you have the pack wired up and fused, you will need a BMS, fortunately there are plenty of BMS boards available for a home built pack on Ebay and Amazon among other places. You may want to measure the voltage where it cuts the battery off on discharge and where it stops on charging as well as what voltage it equalizes the cells at as I have read that some cheap BMS boards sold online do not function in a way that will properly protect a battery. You will also still want to fuse your battery, probably at a rate less than or equal to two thirds of the rated capacity of each cell multiplied by the number of cells in parallel.
I plan to post on my battery building exploits in the coming week or two, once I have all of my thoughts together, but this is just my thoughts on it for now. Good luck with your modems, Rich