The difficulty in remote troubleshooting is not in the technology, but in the information: if you say "the link is not working" on the spot, both parties will guess for half an hour. The following four sets of data are available. Most problems can be identified on the same day as modules, optical paths, or equipment configurations. There is no need to send the parts back first.
The first group: receiving and receiving optical power of modules at both ends
Read the Tx Power, Rx Power, temperature and voltage of DDM. It makes sense to look at the four values together. If the received optical power is lower than the sensitivity threshold, the direction is to check the optical path and attenuation; if it is close to the overload threshold, usually a long-distance module is used for the short link, add an optical attenuator to push the power back into the range; if one end is normal but the other end is abnormal, first suspect that the TX/RX connection is reversed. Multi-mode links can be used first to confirm whetherSFP Gigabit Multimode Dual Fiber Optical ModuleEquipped with OM3 jumpers, it is recommended to leave a 3dB margin in the link budget.
Group 2: Error Count and Port Negotiation Status
- Whether the CRC and input error counts of the interface increase linearly with time
- Check whether the port speed and duplex are negotiated to the expected level. If half-duplex occurs, check the cable and configuration.
- Whether the device reports an unknown or unsupported module, identify the problem and write code to eliminate it
- The frequency and time distribution of flash interruptions. Regular interruptions are mostly related to the scheduling of the peer device.
The third and fourth groups: light path conditions and peer model are all asked in one go
On the optical path side, the jumper length and connector type, ODF terminal number, number of splice points and measured link loss must be provided on site, and there must be direct return of the OTDR curve. When there is no instrument, you can make an alternative judgment: change a known good jumper to test run, clean the end face and then retest the power. These two steps can often distinguish the two main causes of pollution and bending.
The peer information must include the brand model, wavelength and distance range of both sides, as well as whether coding has been done according to the equipment brand. Mixing factories at both ends is a common reason. In the data center interconnection scenario, it is also necessary to confirm whether it belongs toData center high-speed interconnection solutionThe same rate series in the network avoids parameter mismatch between 10G on one side and 25G on the other.
EB-LINK recommends making these four sets of data into an empty form that is maintained with the project and filled in once when reporting for repair; if you need a form template, please contact 0755-83179002.