Don’t know if it’s the best way, but this is how I do it…
I have a counter that counts downtime for the shift. This is logged once per shift so you have easy access to total downtime.
I also have a “running status” counter and a timer based on takt time. Works like this (ours are assembly processes so your mileage may vary a bit):
[ul]The counter values are assigned a status:
[list]6-Break (or lunch)
3-Small Stop 2
2-Small Stop 3
1-Small Stop 4
Now it’s a race.
[ul]If a part is finished first, the counter is reset to 5, and the takt timer is reset.
If the takt timer elapses first, the counter is decremented, and the takt timer is reset.
We allow a little bit of small stop time for things like changing packaging, etc., but basically, if the no part have been produced for 5 cycle times, then the machine is down.[/ul]
I log the counter value every 10 seconds (our takt times range from 20-30 seconds, so the 10 second interval gives a nice resolution).
The upshot of this is that you can now use a status chart to graphically display your downtime and when it happened. So when an operator says that productivity was low because of machine downtime it’s easily verified.
Hope this helps!