Hub SN28 Trailer Addendum - When this Hub first went into the 14SS, it wouldn't power up when inserted into the slot. - The Hub worked as expected when IPMC bypass was installed. - Without the bypass, IPMC is not turning on the ISO_12V converter, so nothing else on the board will come up. 08-11 JUN 2020 - Started investigating IPMC issue with SN28 - Verified all resistors and caps listed in drawings 6, 7, 9, 13, 14, 30, and 31 - Verified all traces/vias/connections as shown in the same drawings and the nets list. (i.e., if Resistor R956 is listed as being tied to JMP5-2 and JMP6-2 one one side and OPT1-1 on the other, I verified that this is true.) - Took Hub SN16 out of the card file and used it as a comparison board with my Huntron current trace, with the IPMC module both installed and uninstalled. - Initially, saw differences in the boards with the IPMC module in the socket, which is why I removed it. Specifically, ICs U2951, U2952, U2953, and U2954. This is the TPS3808, an LM3880, a 7410, and a 7404, respectively. - Also involved are ICs U2955, U2956, U2958, and transistor Q2951. - I tend to discount these parts as problems, as the board will power up normally with the IPMC bypass installed. - Back at the IPMC socket, the following pins come into play: 224 - Handle Switch input 225 - Payload Enable to ISO_12V supply; also to U2954-13 to begin the startup sequence 226 - Alarm IN from Power Entry module 227 - same - Without the IPMCs in their sockets, all the differences/problem areas between boards disappeared. - Re-seated the IPMCs and looked again. Again, all the differences were gone. - Removed the IPMC from SN28 and gently cleaned the card-edge connector. - Removed unit's front panel, applied flux, and re-flowed pins 224-227. - Then I re-installed the IPMC in its socket. - Again verified all traces, routes, vias, and connections. - Again verified all resistors in the above mentioned drawings. - Again compared between SN28 and SN16. The boards are now almost indistinguishable via current trace. - The unit is ready to go back into a slot to test and see if it will power up in a shelf. 12 JUN 2020 - Tried inserting SN28 into Slot #1 of the 6SS. It still won't power up. 15 JUN 2020 - Everything leads back to the IPMC module itself. I have a spare IPMC (SN206) and tried swapping that in, and the behavior changed, but only slightly. - Using the nomenclature given in hub/hardware/details/hub_0_ab_led_front_panel.txt - With original IPMC SN139, no matter what, Mandatory LED 1 and Mandatory Blue come on and stay on, regardless of the handle switch state - indicating, the 12V supply is not being turned on and that the rest of the board isn't powering up. - With IPMC SN206, when the card is inserted, Mandatory LED 1 and Mandatory Blue come on like before, but Mandatory LED 1 quickly turns off and Optional LED 2 comes on, and Mandatory Blue does turn off. - Looking at the other Hubs in the 6 and 14SS, Optional LED 2 stays on during operation. - Normally, after Mandatory 1 and 2 turn off and Optional 2 turns on, at that point the 12V supply is brought up, and the normal startup sequence begins. This is not happening for some reason. - Tried using the Pigeon Point SM to diagnose the problem, but no errors are shown in the log. This indicates no error with the power entry module, at least. - The card does respond to SM's Activate and Deactivate commands, but that's all. - Thus far, I have: - Verified the handle switch; - Verified the handle switch wiring vs. other, working Hubs; - Verified the traces between the handle switch (WTERM71 & 72) and the IPMC (224); - Verified IPMC-224 is at 3.3V when the handle is open and 0V when it is closed; - Verified the optocoupler (OPT1), and the connections between it, resistor R956, and the ISO_12V supply GND (2) & Enable (3) pins; - Verified the trace from IPMC Payload_Enable (225) to the opto via JMP5 and R956; - Verified that the Power Entry module is not sending an alarm (PE-14 -> R954 -> IPMC-226/227); - In all cases, the PE module has -48V in and -48V, 3.3V, and 5V out. - Referring to Drawings 6, 7, 9, 13, 14, 28, 29, 30, 31, 48, 51, 59, and 60; and based on my notes from above, the following conclusions are true: - The problem is not with JMP5, R956, or OPT1, as the board will power up as expected with the IPMC bypass installed, and all these same components make that happen; - The problem is not with the handle switch, the switch wiring, or the trace from the handle switch to the IPMC module; - The problem is not with the Power System control logic, as without ISO_12V, none of that is even in play yet. (i.e., the 12V supply is not dependent on the control logic for its own startup.) 16 JUN 2020 - Just for the sake of thoroughness, I investigated the control logic for all the DC converters: - U2951, 2952, 2953, 2954. - Plus all the discrete components in those circuits. - Plus many of the traces from A to B to make sure that given connections were true. - I did change out the following comnponents, as I didn't like the looks of the signals that passed through them: - TPS3808 (U2951) - LM3800 (U2952) - 74HC10 (U2953) - 74HCT04 (U2954) But, honestly, they were probably OK to begin with. I stopped there, as replacing parts for no good reason can cause more problems than it solves. - Finally just re-installed the IPMC bypass just to see if it would power up. - It does. - Verified that all on-board supplies are on and true. - SN28 is going to have to stay this way for now. Bottom line: The Payload Enable signal is not coming out of the IPMC at Pin 225 to enable the 12V supply. 22-24 JUN 2020 - Verified all connections/traces/vias listed in hub/hardware/mentor_nets_files/atca_isolated_12v_nets hub/hardware/mentor_nets_files/atca_power_entry_nets hub/hardware/mentor_nets_files/ipmc_hw_adrs_handle_switch_ipmb_nets hub/hardware/mentor_nets_files/ipmc_power_ground_and_nc_nets 08-10 JUL 2020 - Removed unit's front panel. - Once again verified all components, traces, and vias from all the above mentioned drawings. - Verified IPMB connections to Zone 1 connector. Possible bad solder joint on 242. - Verified HWA connections from Zone 1 to IPMC & FPGA. - Powered on bench and verified that PE is not sending alarm status. - Reflowed every pin on IPMC socket by hand to ensure good connections. - Verified every pin used on IPMC module.Followed each trace to it's next connection as given in Nets file. - Without PWR/GND and no-connects, there are only like 40 pins used on the whole thing. - Unit is again ready to try live in one of the shelves. 11-13 JUL 2020 - Soldered taps onto Pins 224, 225, 226-7, and 243 so I could scope them when the unit went back in a shelf. - Unit powered up as it should. See BMP images of scope traces and description file in IPMC Test folder.