# Downtime and OEE

> See when each machine stopped and why, give the reasons, and read OEE per shift, worked out by Proxus from the machine's states and counts.

Source: https://proxus.io/docs/monitor/downtime-and-oee

Proxus keeps a **downtime log** for every machine whose [machine type](https://proxus.io/docs/organize/machine-types#states) names its
states: when the machine stopped, for how long, in which state, in which shift and under which work order. People add
the reason. From the log and the machine's counts, Proxus also works out **OEE** for every shift. Downtime is in
**Monitor > Downtime**.

## Before you start

- The machine type has a **State Field** and its **States**, each with what it counts as (running, idle, planned stop,
  unplanned stop, fault). See [Machine types](https://proxus.io/docs/organize/machine-types#states).
- The device has the type, and the state field is mapped to a tag.
- For shifts, set them in **Organize > Shifts** (see below). Without shifts, OEE is worked out per day.
- For performance and quality, the type names its **Good Count Field**, **Reject Count Field** and **Ideal Rate**.

## The downtime log

Every minute, Proxus reads the state tag of each machine and writes a record for each time the machine was in a state
other than running. A stop still going has no end yet; it gets one when the machine runs again. A value the type does
not name is not counted, so a state list left half done never invents stops. The first time, Proxus reads one day back.

| Column | Meaning |
| --- | --- |
| **Device**, **Started**, **Ended**, **Minutes** | Which machine stopped, and when. Times are on your clock. |
| **State**, **Category** | The state's name from the type, and what it counts as. |
| **Shift** | The shift the stop started in. |
| **Work Order**, **Product** | The work order running on the machine when it stopped, and its product. |
| **Reason**, **Note** | Why it stopped, as people gave it. A stop without a reason (not a planned stop) is highlighted. |

## Give the reasons

1. Select one or more stops in the list.
2. Select **Set Reason** and choose the reason, for example *Unplanned › Breakdown › Mechanical*.

The reasons come from the **Downtime reasons** list of [forms](https://proxus.io/docs/organize/forms) (**Organize > Forms > Lists**),
the same list a downtime form uses. Change the list there; the reasons already given keep their text. Proxus records
who gave the reason and when.

## Shifts

**Organize > Shifts** holds the plant's shifts: a name, when the shift starts and ends (an end before the start is on
the next day, for example 22:00 to 06:00) and the days it starts on. A shift is for one site, or for every site that
has no shifts of its own. Shifts of one site cannot overlap. Times are on the site's clock (**Region and Time**).

## OEE

For every shift (or day), Proxus works out:

| Figure | How |
| --- | --- |
| **Availability** | Run time ÷ planned time. Planned time is the shift without its planned stops; run time is planned time without the other stops. |
| **Performance** | Units made ÷ (ideal rate × run time), at most 100 %. Needs a count field and the ideal rate. |
| **Quality** | Good units ÷ (good + rejected). Needs both count fields. |
| **OEE** | Availability × performance × quality. |

Count fields are counters that go up; a counter that resets (at a shift change, say) is fine. A figure that cannot be
worked out stays empty, and OEE with it: Proxus never guesses one. A shift is updated while it runs and closed a few
minutes after it ends.

OEE and downtime appear on the machine type's dashboard (**Create Dashboard**): an **OEE by shift** table and
**Downtime by reason** for the last 7 days, for the device chosen at the top. Ask the
[AI Assistant](https://proxus.io/docs/monitor/copilot) "how long did the labelers stop on shift A today, and why?" or "what was the OEE
of line 3 yesterday?": it reads the same records and works out the totals.

## What's next

- [Machine types](https://proxus.io/docs/organize/machine-types): states, count fields and the ideal rate.
- [Forms](https://proxus.io/docs/organize/forms): the downtime reasons list.
