eCPH and PBR Summary Report

• ECPH = Actual placed components per hour / benchmark machine CPH*100%
  • Q: For line, this is summed across all machines in the line, right? Yes.

ECPH is per machine, summed per line, aggregated up the hierarchy.

For ECPH, panels have no effect.

• PBR = (H1 + H2 + H3 +…+ Hn) / [(Hmax) * n] * 100%]

H1 = cycle time of PLACEMENT HEAD-1 in the line               

H2 = cycle time of PLACEMENT HEAD-2  in the line

Hn = cycle time of last PLACEMENT HEAD in the line           

Hmax = the longest cycle time from H1 to Hn

n = how many PLACEMENT HEADs in the line

PBR is calculated by PnP work area. 4 ASM PnP machines. Each has up to 2 work areas. If recipe demands both work areas work simultaneously and run across line, so could have 8 individual areas. Measure PBR by area.  So eight numerators and denominator is area with highest rate.

At lowest level, by model (recipe): have PBR calculated purely by the below formula. For a given line and recipe, that's how you calculate.

When you compare across lines, want to weigh the PBR ratios by # of panels produced. e.g. Line 1 produces 1000 panels and Line 2 produces 2 panels, for aggregation by area and upward, weigh by number of panels. Rolls up to area, site, region, globally.

NOTE: Can only do a SUMMARY report for eCPH and PBR due to Excel - side limitations; no PivotTable-style report.

---

#### eCPH (All calculations will be across time)

```
machine level = 100 * actual placements / peak cph
```

```
entities = {line, area, site, region, global}

<entity>-level eCPH = 	100 * (sum of actual placements from all machines in the <entity>)
							----------------------------------------------------------------
								(sum of peak cph from all machines in the <entity>)

```

---

#### PBR (All calculations will be across time)

```

model (product identifier) level PBR ==>

		for a line, for a product_identifier,

			h1 = median cycle time of machine1
			h2 = median cycle time of machine2
			...
			hN = median cycle time of machineN

			PBR = 100 * (h1 + h2 + ... + hN)
			      --------------------------
			       max(h1, h2,....,hN) * N

line level PBR ==>

		for a line,

			p1 = # of panels from model 1 (product identifier)
			pbr1 = PBR of model 1

			pbr_line = p1*pbr1 + p2*pbr2 + ...
			           -----------------------
			           p1 + p2 + ...


entities = {area, site, region, global}
<entity>-level PBR ==>

		for all lines in <entity>

				pbr_area = mean(pbr_line) (no weighting, lines are identical)

```