Out of 12 Cisco/Linksys SD216 10/100 16 ports Ethernet switches, 7 have now failed. I just got time to look at these. To open the box, first remove the front by depressing the 3 tabs on top of the plastic, then slide the top metal part of the housing to the back to disengage the four ‘L' hooks, see pictures.
[A] 5 units [out of the 7] were ‘Rev. 1'; 4 had the same failure mode: with 12VDC applied, the PWR LED would light up, then all the LEDs would blink very fast at a very low power. I noticed that the 2 primary caps C3/C4 were very hot e.g. 80deg C. These caps are ‘STONE' 100micF 25V 105degC, and are connected to the power socket J1 wired to the input socket 12VDC. The remaining 1st stage of the power circuit converts 12VDc to 3.3VDC using an LM3485 buck controller U5, a 4403 MOSFET Q1 and a Shottky diode D3; schematic is straight out of the Texas Instruments LM3485 data sheet. The 3.3VDC is further converted to 1.81VDC for the 2 10/100 8 ports IC's hidden by the large heatsinks. I replaced the 2 caps by a single 330micF 25VDC, and all is OK –for now- The removed caps measured 18, 19, 38 & 41micF.
[B] 2 units [out of the 7] were ‘Rev. 2” with a redesigned power circuit and new switch ICs. 1 unit died because of the poor quality of the 12VDC socket wired to the jack J1: the center pin was ‘floating' and caused intermittent contacts. Replacing the socket fixed the issue.
When I have time, I will look at the box of failed SD208, 8-ports in the same compact chassis.
Pics: SD216 box, Front tags, Side hooks, PCB Rev1, Power circuitsviews 1&2, 'hot' Caps, Fix Rev1, PCB Rev2, Replacement socket.
[A] 5 units [out of the 7] were ‘Rev. 1'; 4 had the same failure mode: with 12VDC applied, the PWR LED would light up, then all the LEDs would blink very fast at a very low power. I noticed that the 2 primary caps C3/C4 were very hot e.g. 80deg C. These caps are ‘STONE' 100micF 25V 105degC, and are connected to the power socket J1 wired to the input socket 12VDC. The remaining 1st stage of the power circuit converts 12VDc to 3.3VDC using an LM3485 buck controller U5, a 4403 MOSFET Q1 and a Shottky diode D3; schematic is straight out of the Texas Instruments LM3485 data sheet. The 3.3VDC is further converted to 1.81VDC for the 2 10/100 8 ports IC's hidden by the large heatsinks. I replaced the 2 caps by a single 330micF 25VDC, and all is OK –for now- The removed caps measured 18, 19, 38 & 41micF.
[B] 2 units [out of the 7] were ‘Rev. 2” with a redesigned power circuit and new switch ICs. 1 unit died because of the poor quality of the 12VDC socket wired to the jack J1: the center pin was ‘floating' and caused intermittent contacts. Replacing the socket fixed the issue.
When I have time, I will look at the box of failed SD208, 8-ports in the same compact chassis.
Pics: SD216 box, Front tags, Side hooks, PCB Rev1, Power circuitsviews 1&2, 'hot' Caps, Fix Rev1, PCB Rev2, Replacement socket.
Comment