Custom modem cable.
My 2nd hand modem E272 came with very thin 58cm cable, similar to the cables supplied with digital cameras. It never worked with this cable. I experienced various problems from failing to register to the network (green light blinking twice - searching for network) to the complete reseting of the modem (indicated that OS was reinstalling drivers for the device). I checked this cable and found that the power line has resistance 1.1Ohm and ground wire 1.4Ohm, for the total 2.5Ohm. The modem is rated 500mA, this will generate voltage drop 1.25V which is far below USB specification!
E272 is a type of the "sausage on the rope" device, it needs a cable. I tested it with 80cm old Nokia DKE-2 cable and 100cm cable supplied with Garmin device. It works, but I found that connection is more stable with the borrowed original E220 Y-cable. I realised that I need much shorter cable, or apply a solution below which helps with dynamic load.
Chosing a cable. It must be a thick one, the resistance of the power and ground wire must be beyond recognition of typical digital multimeter. 1Ohm of total resistance gives 0.5V drop of the voltage on the modem side. This solution will not help if cable resistance is high, or if USB port on PC has got faulty polyswitch.
A peak load of the modem is much higher than specified 500mA. It is why I decided to apply in-a-cable a filter solution. There are already such solutions published on the net. Cut the cable, add come capacitors between power and ground line. Adding a Shottky diode is claimed to work better... maybe, but a reasoning is wrong. Besides, I don't like losing 0.5V on the diode. It is why it is much better to replace diode with a choke.
In this project chose a cable as above. Cut insulation (near modem end) to uncover individual wires. Try not destroy a shield. Do not touch data wires. Open insulation of the ground wire for soldering (usually black), but do not cut it. Cut the power wire (usually red). Attach it firmly to a prototype printed-circuit board with already solldered-in choke and capacitors. It is simple. If you don't have experience or do not understand what I am talking about, better ask a friendly electrician to do it for you.
Chosing a coke. This is critical to use a choke which is rated more than 500mA DC current. Be careful with a torroid type or other closed-field. They can easily saturate, meaning do_not_work. Chosing unknown choke from the scrap electronics usually do not work. To meassure saturation current you need something more than a typical RLC meter. It is why it is a tricky part. I chose a choke from a scraped ATX power supply. It sits in 5Vsb circuit and this output is rated 1A, so I know that 500mA will not saturate a choke.
Chosing capacitors. Your choice is not critical, except for the high capacity electrolitic (1000uF or more), which must be low-esr type. It can be found in output of any regulated switching mode power supplies. It has to be also small (around) 10nF ceramic at the location closest to the cutted wire and modem. My choice of capacitors is C1=1500uF/6.3V, C2=33uF/(unknown voltage) tantalum, C3=10OnF/50V MKP or MKP like, C4=820nF (similar - unknown, added because it was space), C5=10nF ceramic.
Conclusion: Everybody needs a cable. To take modem out from the space obscured by metal cabinets of your desk, bring it to the location of better reception. Standard USB extensions for power-hungry modems usually give other (various) problems, unless modem is attached to the self-powered USB 2.0 bub. I tested various modems connected to self-powered USB hub and 3 pieces of 1m extension cable to the PC (total 3m). Similar results. It works only when external power is supplied to the hub. The same configuration (3 extension cables) works when using my modified cable as a last piece. It doesn't need hub to work and - if modem is connected to the hub, the hub does not need external power.
My 2nd hand modem E272 came with very thin 58cm cable, similar to the cables supplied with digital cameras. It never worked with this cable. I experienced various problems from failing to register to the network (green light blinking twice - searching for network) to the complete reseting of the modem (indicated that OS was reinstalling drivers for the device). I checked this cable and found that the power line has resistance 1.1Ohm and ground wire 1.4Ohm, for the total 2.5Ohm. The modem is rated 500mA, this will generate voltage drop 1.25V which is far below USB specification!
E272 is a type of the "sausage on the rope" device, it needs a cable. I tested it with 80cm old Nokia DKE-2 cable and 100cm cable supplied with Garmin device. It works, but I found that connection is more stable with the borrowed original E220 Y-cable. I realised that I need much shorter cable, or apply a solution below which helps with dynamic load.
Chosing a cable. It must be a thick one, the resistance of the power and ground wire must be beyond recognition of typical digital multimeter. 1Ohm of total resistance gives 0.5V drop of the voltage on the modem side. This solution will not help if cable resistance is high, or if USB port on PC has got faulty polyswitch.
A peak load of the modem is much higher than specified 500mA. It is why I decided to apply in-a-cable a filter solution. There are already such solutions published on the net. Cut the cable, add come capacitors between power and ground line. Adding a Shottky diode is claimed to work better... maybe, but a reasoning is wrong. Besides, I don't like losing 0.5V on the diode. It is why it is much better to replace diode with a choke.
In this project chose a cable as above. Cut insulation (near modem end) to uncover individual wires. Try not destroy a shield. Do not touch data wires. Open insulation of the ground wire for soldering (usually black), but do not cut it. Cut the power wire (usually red). Attach it firmly to a prototype printed-circuit board with already solldered-in choke and capacitors. It is simple. If you don't have experience or do not understand what I am talking about, better ask a friendly electrician to do it for you.
Chosing a coke. This is critical to use a choke which is rated more than 500mA DC current. Be careful with a torroid type or other closed-field. They can easily saturate, meaning do_not_work. Chosing unknown choke from the scrap electronics usually do not work. To meassure saturation current you need something more than a typical RLC meter. It is why it is a tricky part. I chose a choke from a scraped ATX power supply. It sits in 5Vsb circuit and this output is rated 1A, so I know that 500mA will not saturate a choke.
Chosing capacitors. Your choice is not critical, except for the high capacity electrolitic (1000uF or more), which must be low-esr type. It can be found in output of any regulated switching mode power supplies. It has to be also small (around) 10nF ceramic at the location closest to the cutted wire and modem. My choice of capacitors is C1=1500uF/6.3V, C2=33uF/(unknown voltage) tantalum, C3=10OnF/50V MKP or MKP like, C4=820nF (similar - unknown, added because it was space), C5=10nF ceramic.
Conclusion: Everybody needs a cable. To take modem out from the space obscured by metal cabinets of your desk, bring it to the location of better reception. Standard USB extensions for power-hungry modems usually give other (various) problems, unless modem is attached to the self-powered USB 2.0 bub. I tested various modems connected to self-powered USB hub and 3 pieces of 1m extension cable to the PC (total 3m). Similar results. It works only when external power is supplied to the hub. The same configuration (3 extension cables) works when using my modified cable as a last piece. It doesn't need hub to work and - if modem is connected to the hub, the hub does not need external power.
Last edited:
