At the machine
Suppose a counter moves from 104 to 105 when a part leaves a mill. Watch one completed cycle and compare the control display with the value in Edge. Establish whether the counter covers one spindle or the entire machine, and whether it includes rejected parts. A counter that resets at each job needs different handling from a lifetime total.
On the local Edge device
In the local machine profile, identify the signal that carries the count and how it maps to a production event. Check the observed value, timestamp and signal quality. If the cloud connection drops, Edge can keep collecting and queue events locally while storage is available.
In Qynetic
Edge opens the outgoing HTTPS connection and sends its queued events. The heartbeat in Integrations shows when the Edge device last contacted Qynetic. To verify production data, follow an actual count change for the same machine and confirm its event was accepted. A rejected event and a machine that has simply stopped making parts call for different checks.
On the shift dashboard
Compare the dashboard count with the number of parts completed during the selected shift. Planned production time and running time give availability. Ideal cycle time multiplied by total output, divided by run time, gives performance. Good parts and total parts give quality. Their product is OEE. A single counter alone cannot establish all three factors.
For example, a shift starts with the counter at 104 and ends at 137. With no resets or missing intervals that is 33 parts. If two failed first-pass inspection, quality is 31 / 33, or 93.9%. A reset during the shift breaks the subtraction. Reconcile the counts on each side of the reset with the job records.