Page 1 of 1
Posted: Fri 09 Mar 09 2018 3:27 pm
Google Drive images for inspection:-
https://drive.google.com/drive/folders/ ... rcj6HVtpoh
Photos appear to be of 3 different boards:-
- 2 PCI boards
- 1 cased board
One of the PCI boards appear to be an industrial SBC (HICORE).
Posted: Fri 09 Mar 09 2018 3:34 pm
1. Full length PCI card
2. Lots of dust ingress - uncertain what kind of dust: conductivity?
3. Some signs of oxidation
4. Some scuff marks indicate the board has been out for a while and not kept in protection
5. Most of the caps appear to be MLCC
6. A few large electrolytics 82uF/20V, 220uF/?? and tantalums 100uF/10V
7. Chipped inductor L5 (img 20180305_183030.jpg) valued at 1R5 (1.5uH)
8. Board includes a Class 1 laser module - uncertain function: optical coupling? fibre-optic comms?
9. Board design from 1999
10. Bodge wire at RN3 (resistor network) under board
11. Excessive dust accumulation at lower left corner of board (img 20180305_182300.jpg, 20180305_181514.jpg)
12. Diagnostic power bus LED indicators D8 and D9 for 3.3V and 5V bus respectively
13. Some scratches on traces (img 20180305_182031.jpg)
14. Unidentified stuff on the board (img 20180305_182016.jpg, 20180305_182006.jpg): liquid ingress? glue? paint?
15. One of the boards appear to be a i386 PC architecture (PhoenixBIOS 1999-2004) with Lithium battery on board
16. 1500uF/10V and 1500uF/6.3V electrolytics on board - no obvious signs of leakage or bulging, may need inspection
17. SIMM memory stick not properly seated (img 20180305_181456.jpg)
18. Laser module shows some flux residue. Could this have been reworked? (img 20180305_182307.jpg)
Posted: Fri 09 Mar 09 2018 4:07 pm
1. Power supply issue - this typically happens with old equipment that are run for long hours
- supply voltage testing
- electrolytics or tantalum failures
- need to check the bus voltages to see if expected 3.3V and 5V are present
2. Laser module failure
- laser LEDs tend to have limited operational life, may fail over time
- flux might indicate that this part has been reworked, or see previous failure
- need method to test function
3. No obvious signs of component failure
- although MOSFETS were not shown clearly
- lots of dust ingress, may or may not have affected function
4. Unlikely to have FPGA or CPU failures
- these should be typically safe unless there is a power line issue
5. Random component failure
- assuming there is a pattern to the failure, we are unlikely to be looking at random component failure
- more likely a usage issue
1. What are the failure patterns?
- do the boards work well initially and fail after several years? If so, it may be due to wear and tear, particularly of the laser module
2. What is the nature of the failure?
- are the on-board diagnostic LEDs indicating any power issue?
- is the onboard CPU booting up? Does the VGA output give any signal?
- is the serial port outputting any diagnostic info?
Posted: Fri 09 Mar 09 2018 4:17 pm
- Intel Pentium III Socket 370 Tualatin SBC
- appears to be standard industrial equipment
- no reports of systemic or chronic failure
- no reason to suspect design flaw (either thermal or electronic)
This board can be purchased used on e-bay cheaply (USD270):-
https://www.ebay.com/itm/1pcs-Used-ARBO ... 2645492334
Posted: Mon 12 Mar 12 2018 1:41 pm
1. One of the parts is supposedly falling off the board?
- this could be due to mechanical failure if board is exposed to vibration or physical stress
- more likely a thermal issue of either temperature change induced material failure (solder joints)
- or high temperatures leading to solder melting
2. One of the boards (uncertain which one or if related to #1) getting charred?
- probably points to high temperatures, possibly related to #1
1. To check operating board to see if excessive temperatures are present
- may also want to check board ventilation and cooling
2. To understand installation to see if there are sudden temperature changes
3. To check if boards were exposed to vibrations or other physical stress (unlikely)
1. Is it always the same part that falls out?
2. For boards that fail - are they always or usually the same board involving the component falling out?
Posted: Mon 26 Mar 26 2018 9:41 pm
Part 2 photos observations
1. Boards on the backplane appear to utilize fibre-optic cables for some purpose. Two connections per board.
2. Image 731 shows what appears to be oxidation on underside of BGA part. This could indicate over-heating or corrosion from leakage.
3. Image 730 shows similar signs on what appears to be the same part but different location on the board.
4. Image 729 shows several electrolytics on the other side of the board? Could these be leaking?
1. Capacitor leakage on the power supply side
2. Identify the reverse side of the components with issues.
Posted: Wed 28 Mar 28 2018 4:28 pm
Summary of prime suspects
1. Backplane power-supply over-voltage leading to overheating of components and systematic eventual failure
2. High operating temperatures leading to multiple modes of failure: leaking electrolytes resulting in corrosion
3. Dust accumulation leading to poor ventilation and high heat, possibly leading to melting of solder
4. Temperature fluctuations leading to material failure
1. Ascertain what is making things hot. Weather? Equipment? The part itself?
2. Ascertain frequency and modes of failure. Always the same slot or part?
3. Measure backplane voltage. Get model number of backplane.