Capacity ramp planner
Plan a fab or line ramp month by month: tool install and qualification, wafer starts, cycle time, yield and the output that reaches customers.
Manufacturing, fab and supplyFree to useHow it worksSourcesWhere it helps
Fills the inputs below with typical figures. Replace any of them with your own.
- Month demand is met
- Month 11
- Capacity covers demand from this month to the end of the plan
- Peak work in process
- 63,370 wafers
- By Little's law, in month 16
- First wafers out
- Month 7
- First good die out
- Month 7
- Good die out over the plan
- 214,059,825
- 572,911 wafers out of 646,233 started
- Capacity in the last month
- 25,500 wafer starts
- 6 tool sets installed; starts 24,000 a month
- Months held back by capacity
- 10
- 28% of the plan
Check these
- Starts are held back by capacity in 10 months of the plan.
- Wafer starts (solid line, circles)
- Wafers out (dashed line, squares)
- Demand (dotted line, triangles)
- Capacity (dash-dot line, diamonds)
Show the data as a table
| Month of the plan | Wafer starts | Wafers out | Demand | Capacity |
|---|---|---|---|---|
| 1 | 0 | 0 | 6,000 | 0 |
| 2 | 0 | 0 | 6,600 | 0 |
| 3 | 0 | 0 | 7,260 | 0 |
| 4 | 1,000 | 0 | 7,990 | 1,000 |
| 5 | 2,360 | 0 | 8,780 | 2,360 |
| 6 | 4,080 | 0 | 9,660 | 4,080 |
| 7 | 6,170 | 709 | 10,600 | 6,170 |
| 8 | 8,610 | 2,130 | 11,700 | 8,610 |
| 9 | 11,400 | 3,970 | 12,900 | 11,400 |
| 10 | 13,600 | 6,670 | 14,100 | 13,600 |
| 11 | 15,600 | 9,130 | 15,600 | 15,800 |
| 12 | 17,100 | 11,900 | 17,100 | 17,900 |
| 13 | 18,800 | 14,100 | 18,800 | 20,100 |
| 14 | 20,700 | 16,100 | 20,700 | 21,900 |
| 15 | 22,800 | 17,700 | 22,800 | 23,300 |
| 16 | 24,000 | 19,500 | 24,000 | 24,400 |
| 17 | 24,000 | 21,400 | 24,000 | 25,100 |
| 18 | 24,000 | 23,200 | 24,000 | 25,500 |
| 19 | 24,000 | 23,900 | 24,000 | 25,500 |
| 20 | 24,000 | 23,900 | 24,000 | 25,500 |
| 21 | 24,000 | 23,800 | 24,000 | 25,500 |
| 22 | 24,000 | 23,800 | 24,000 | 25,500 |
| 23 | 24,000 | 23,800 | 24,000 | 25,500 |
| 24 | 24,000 | 23,700 | 24,000 | 25,500 |
| 25 | 24,000 | 23,700 | 24,000 | 25,500 |
| 26 | 24,000 | 23,700 | 24,000 | 25,500 |
| 27 | 24,000 | 23,700 | 24,000 | 25,500 |
| 28 | 24,000 | 23,700 | 24,000 | 25,500 |
| 29 | 24,000 | 23,600 | 24,000 | 25,500 |
| 30 | 24,000 | 23,600 | 24,000 | 25,500 |
| 31 | 24,000 | 23,600 | 24,000 | 25,500 |
| 32 | 24,000 | 23,600 | 24,000 | 25,500 |
| 33 | 24,000 | 23,600 | 24,000 | 25,500 |
| 34 | 24,000 | 23,600 | 24,000 | 25,500 |
| 35 | 24,000 | 23,600 | 24,000 | 25,500 |
| 36 | 24,000 | 23,600 | 24,000 | 25,500 |
- Little's law (starts rate x cycle time) (solid line, circles)
- Counted wafer by wafer (dashed line, squares)
Show the data as a table
| Month of the plan | Little's law (starts rate x cycle time) | Counted wafer by wafer |
|---|---|---|
| 1 | 0 | 0 |
| 2 | 0 | 0 |
| 3 | 0 | 0 |
| 4 | 3,250 | 1,000 |
| 5 | 7,460 | 3,360 |
| 6 | 12,600 | 7,440 |
| 7 | 18,500 | 12,900 |
| 8 | 25,300 | 19,200 |
| 9 | 33,000 | 26,500 |
| 10 | 38,500 | 33,200 |
| 11 | 43,500 | 39,300 |
| 12 | 47,200 | 44,100 |
| 13 | 51,200 | 48,300 |
| 14 | 55,700 | 52,500 |
| 15 | 60,700 | 57,100 |
| 16 | 63,400 | 61,100 |
| 17 | 62,900 | 63,100 |
| 18 | 62,400 | 63,400 |
| 19 | 62,000 | 62,900 |
| 20 | 61,700 | 62,400 |
| 21 | 61,400 | 62,000 |
| 22 | 61,100 | 61,700 |
| 23 | 60,900 | 61,400 |
| 24 | 60,700 | 61,100 |
| 25 | 60,500 | 60,900 |
| 26 | 60,400 | 60,700 |
| 27 | 60,200 | 60,500 |
| 28 | 60,100 | 60,400 |
| 29 | 60,000 | 60,200 |
| 30 | 59,900 | 60,100 |
| 31 | 59,800 | 60,000 |
| 32 | 59,700 | 59,900 |
| 33 | 59,600 | 59,800 |
| 34 | 59,600 | 59,700 |
| 35 | 59,500 | 59,600 |
| 36 | 59,500 | 59,600 |
Sign in to save your inputs and come back to them later.
| Month | Tool sets installed | New tool sets productive | Capacity (wafer starts) | Demand (wafer starts) | Wafer starts | Cycle time (days) | Wafers out | Line yield | Die yield | Good die out | Cumulative good die | Work in process (Little) | Work in process (counted) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 1 | 0 | 0 | 6,000 | 0 | 110 | 0 | 90.0% | 30.0% | 0 | 0 | 0 | 0 |
| 2 | 1 | 0 | 0 | 6,600 | 0 | 106 | 0 | 91.2% | 35.3% | 0 | 0 | 0 | 0 |
| 3 | 1 | 0 | 0 | 7,260 | 0 | 102 | 0 | 92.3% | 40.0% | 0 | 0 | 0 | 0 |
| 4 | 1 | 1 | 1,000 | 7,986 | 1,000 | 99.1 | 0 | 93.1% | 44.2% | 0 | 0 | 3,254 | 1,000 |
| 5 | 1 | 2 | 2,361 | 8,785 | 2,361 | 96.2 | 0 | 93.9% | 47.9% | 0 | 0 | 7,465 | 3,361 |
| 6 | 1 | 3 | 4,083 | 9,663 | 4,083 | 93.7 | 0 | 94.5% | 51.3% | 0 | 0 | 12,575 | 7,444 |
| 7 | 0 | 4 | 6,167 | 10,629 | 6,167 | 91.5 | 709 | 95.1% | 54.3% | 192,533 | 192,533 | 18,545 | 12,865 |
| 8 | 0 | 5 | 8,611 | 11,692 | 8,611 | 89.6 | 2,134 | 95.5% | 57.0% | 608,438 | 800,972 | 25,346 | 19,242 |
| 9 | 0 | 6 | 11,417 | 12,862 | 11,417 | 87.9 | 3,970 | 95.9% | 59.4% | 1,179,886 | 1,980,858 | 32,961 | 26,519 |
| 10 | 0 | 6 | 13,583 | 14,148 | 13,583 | 86.4 | 6,672 | 96.2% | 61.6% | 2,055,087 | 4,035,945 | 38,541 | 33,168 |
| 11 | 0 | 6 | 15,750 | 15,562 | 15,562 | 85 | 9,125 | 96.5% | 63.5% | 2,899,185 | 6,935,130 | 43,474 | 39,273 |
| 12 | 0 | 6 | 17,917 | 17,119 | 17,119 | 83.8 | 11,932 | 96.7% | 65.3% | 3,894,073 | 10,829,203 | 47,159 | 44,056 |
| 13 | 0 | 6 | 20,083 | 18,831 | 18,831 | 82.8 | 14,138 | 96.9% | 66.8% | 4,723,562 | 15,552,765 | 51,231 | 48,298 |
| 14 | 0 | 6 | 21,889 | 20,714 | 20,714 | 81.9 | 16,064 | 97.1% | 68.2% | 5,478,287 | 21,031,052 | 55,730 | 52,465 |
| 15 | 0 | 6 | 23,333 | 22,785 | 22,785 | 81.1 | 17,677 | 97.2% | 69.4% | 6,138,041 | 27,169,094 | 60,697 | 57,069 |
| 16 | 0 | 6 | 24,417 | 24,000 | 24,000 | 80.4 | 19,450 | 97.3% | 70.6% | 6,861,687 | 34,030,781 | 63,370 | 61,088 |
| 17 | 0 | 6 | 25,139 | 24,000 | 24,000 | 79.7 | 21,400 | 97.4% | 71.5% | 7,655,654 | 41,686,435 | 62,872 | 63,127 |
| 18 | 0 | 6 | 25,500 | 24,000 | 24,000 | 79.2 | 23,170 | 97.5% | 72.4% | 8,391,930 | 50,078,365 | 62,434 | 63,370 |
| 19 | 0 | 6 | 25,500 | 24,000 | 24,000 | 78.7 | 23,910 | 97.6% | 73.2% | 8,754,953 | 58,833,318 | 62,046 | 62,872 |
| 20 | 0 | 6 | 25,500 | 24,000 | 24,000 | 78.3 | 23,868 | 97.7% | 73.9% | 8,824,462 | 67,657,780 | 61,705 | 62,434 |
| 21 | 0 | 6 | 25,500 | 24,000 | 24,000 | 77.9 | 23,830 | 97.7% | 74.6% | 8,886,379 | 76,544,160 | 61,403 | 62,046 |
| 22 | 0 | 6 | 25,500 | 24,000 | 24,000 | 77.5 | 23,796 | 97.8% | 75.2% | 8,941,569 | 85,485,729 | 61,137 | 61,705 |
| 23 | 0 | 6 | 25,500 | 24,000 | 24,000 | 77.2 | 23,766 | 97.8% | 75.7% | 8,990,792 | 94,476,521 | 60,902 | 61,403 |
| 24 | 0 | 6 | 25,500 | 24,000 | 24,000 | 77 | 23,739 | 97.8% | 76.1% | 9,034,718 | 103,511,239 | 60,695 | 61,137 |
| 25 | 0 | 6 | 25,500 | 24,000 | 24,000 | 76.7 | 23,715 | 97.9% | 76.5% | 9,073,935 | 112,585,174 | 60,512 | 60,902 |
| 26 | 0 | 6 | 25,500 | 24,000 | 24,000 | 76.5 | 23,693 | 97.9% | 76.9% | 9,108,965 | 121,694,138 | 60,350 | 60,695 |
| 27 | 0 | 6 | 25,500 | 24,000 | 24,000 | 76.4 | 23,674 | 97.9% | 77.2% | 9,140,267 | 130,834,405 | 60,208 | 60,512 |
| 28 | 0 | 6 | 25,500 | 24,000 | 24,000 | 76.2 | 23,657 | 97.9% | 77.5% | 9,168,249 | 140,002,654 | 60,082 | 60,350 |
| 29 | 0 | 6 | 25,500 | 24,000 | 24,000 | 76.1 | 23,641 | 97.9% | 77.8% | 9,193,271 | 149,195,926 | 59,971 | 60,208 |
| 30 | 0 | 6 | 25,500 | 24,000 | 24,000 | 75.9 | 23,628 | 97.9% | 78.0% | 9,215,654 | 158,411,580 | 59,873 | 60,082 |
| 31 | 0 | 6 | 25,500 | 24,000 | 24,000 | 75.8 | 23,616 | 97.9% | 78.2% | 9,235,682 | 167,647,262 | 59,787 | 59,971 |
| 32 | 0 | 6 | 25,500 | 24,000 | 24,000 | 75.7 | 23,605 | 98.0% | 78.4% | 9,253,606 | 176,900,867 | 59,710 | 59,873 |
| 33 | 0 | 6 | 25,500 | 24,000 | 24,000 | 75.6 | 23,595 | 98.0% | 78.6% | 9,269,651 | 186,170,518 | 59,643 | 59,787 |
| 34 | 0 | 6 | 25,500 | 24,000 | 24,000 | 75.6 | 23,587 | 98.0% | 78.7% | 9,284,017 | 195,454,536 | 59,584 | 59,710 |
| 35 | 0 | 6 | 25,500 | 24,000 | 24,000 | 75.5 | 23,579 | 98.0% | 78.9% | 9,296,883 | 204,751,419 | 59,531 | 59,643 |
| 36 | 0 | 6 | 25,500 | 24,000 | 24,000 | 75.4 | 23,573 | 98.0% | 79.0% | 9,308,407 | 214,059,825 | 59,485 | 59,584 |
| Case | Month demand is met | First wafers out | Peak work in process (wafers) | Good die out over the plan |
|---|---|---|---|---|
| 0 months | Month 6 | Month 4 | 63,370 | 223,407,058 |
| 1 month | Month 7 | Month 5 | 63,370 | 220,927,443 |
| 2 months | Month 9 | Month 6 | 63,370 | 217,952,186 |
| 3 months (your plan) | Month 11 | Month 7 | 63,370 | 214,059,825 |
| 4 months | Month 17 | Month 8 | 62,872 | 207,789,582 |
| 6 months | Month 19 | Month 10 | 62,046 | 192,009,699 |
| Case | Month demand is met | First wafers out | Peak work in process (wafers) | Good die out over the plan |
|---|---|---|---|---|
| −30% | Month 11 | Month 6 | 44,359 | 219,472,576 |
| −15% | Month 11 | Month 6 | 53,864 | 216,774,714 |
| Your cycle time | Month 11 | Month 7 | 63,370 | 214,059,825 |
| +15% | Month 11 | Month 7 | 72,875 | 211,319,696 |
| +30% | Month 11 | Month 8 | 82,381 | 208,498,406 |
How it works
A new line does not switch on. Tool sets arrive, are qualified, and then climb to full output while the process learns. This planner turns an install schedule, a qualification lag, an output ramp, a cycle time and yield curves into wafer starts, wafers out, good die out and work in process, month by month, and shows the month capacity first catches up with demand.
- Install and qualify the tool sets. A tool set is a block of tools that together give a capacity in wafer starts a month. Tool sets are installed on the schedule you enter (a steady pace, or a month by month list). Each one becomes productive after the qualification lag. An existing line, if you have one, is counted from month 1 at its full rate.
first productive month of a tool set = install month + qualification lag - Ramp the output of each tool set. A newly qualified tool set does not deliver its full rate at once. Its output rises in a straight line from the starting share to the mature share over the ramp time, counted from its own first productive month. Capacity in a month is the sum over every productive tool set.
output share at age a = start share + (mature share − start share) x min(1, a / ramp months); capacity = existing line + sum of (tool sets x wafer starts a month per tool set x output share) - Limit the starts by capacity and demand. A line starts what it has capacity for, up to what customers ask for. Demand is a level that grows by a monthly percentage (with an optional ceiling), or a list in which each figure holds until the next one. The month demand is met is the first month from which capacity covers demand in every later month.
wafer starts = min(capacity, demand) - Move the starts through the line. Wafers started in a month leave the line one cycle time later. Starts are spread evenly across their month, so the wafers leave over an interval one month long that is shifted by the cycle time, and it is shared between the output months it overlaps. If cycle time learning is on, the cycle time falls from its starting value towards its mature value, and the cycle time that applies is the one in the month the wafers were started.
cycle time in months = cycle time in days / 30.4375; wafers leaving in month m = sum over start months s of starts(s) x overlap of [s − 1 + cycle time, s + cycle time) with [m − 1, m); cycle time learning: cycle time(s) = mature + (start − mature) x exp(−(s − 1) / time constant) - Apply the yields. Line yield is the share of wafers that finish the flow without being scrapped. Die yield is the share of die on a finished wafer that work. Both are applied in the month the wafers leave, and each climbs from a starting value to a mature value with a time constant, the same shape the yield learning curve tool fits to data.
yield(m) = mature − (mature − start) x exp(−(m − 1) / time constant); wafers out = wafers leaving x line yield; good die out = wafers out x die per wafer x die yield - Count the work in process. Little's law says the average work in process in a stable line equals the rate wafers enter times the time they spend in it. The tool applies it each month to the month's starts rate. The table also shows the work in process counted wafer by wafer (starts in minus wafers out), which differs from Little's law while the line is still ramping.
work in process (Little) = wafer starts a month / 30.4375 x cycle time in days; counted work in process = previous + starts − wafers leaving
Assumptions
- The planner follows the bottleneck: one kind of tool set whose capacity limits the whole line. A real line has many tool types, and the scarcest one sets the pace.
- Wafers of one month's starts all take the same cycle time; there is no spread of cycle times between lots and no queueing model.
- Wafers that fail the line yield are scrapped as they leave the line; the work in process counts them until then.
- Yields are applied by the month wafers leave, not the month they start. The learning curve is calendar based, so a change in the plan's start date does not shift it.
- The existing line (if any) runs at its full rate before month 1 with the starting cycle time, so its work in process is full at the start.
- All figures in the presets are editorial estimates for a typical case. Replace them with your own tool counts, qualification times, cycle times and yields.
Limits and warnings
- Cycle time does not rise with load. In a real line, cycle time climbs steeply as utilisation approaches the limit; feed that in as a longer cycle time or a lower mature output share.
- Product mix, rework loops, reticle and tool re-use, and tool-by-tool constraints are not modelled; use a full line simulation for those.
- Fab building, cleanroom and utility readiness are assumed to be in place before tool installation; they add to the real schedule.
- Demand is a wafer start demand. Convert a die demand to wafers with the yields before entering it.
- Months are average months of 30.4375 days; a plan shorter than six months or longer than five years is outside what the tool is meant for.
Sources
The methods and default figures rest on these public sources. The wording here is our own; read the source for its full text.
- A Proof for the Queuing Formula: L = λW (opens in a new tab), J. D. C. Little; Operations Research, volume 9, issue 3 (INFORMS), 1961. Work in process equals the arrival rate times the time in the system. The link is to Little's own 50th anniversary retrospective of the proof (Operations Research, 2011); the 1961 article details are as cited there.
- Factory Physics, third edition, Wallace J. Hopp and Mark L. Spearman; Waveland Press, 2011. Little's law as throughput = work in process / cycle time, and critical work in process in production lines. Edition and publisher confirmed from library catalogue records; the book was not read page by page.
- These Advanced 12-Inch Wafer Fabs Enter a New Stage (opens in a new tab), TrendForce, 3 December 2024. Company capacity targets in wafers a month for new 300 mm lines; used to set the scale of the phase 1 and block sizes in the presets.
- Nanya Tech Unveils Up to US$10.7B Fab 5A Investment Plan Through 2029; Targets 10nm-Class DRAM with EUV (opens in a new tab), TrendForce, 6 August 2026. A stated memory fab ramp: wafer starts in the second half of 2027, 30,000 wafers a month in 2028 and 35,900 in 2029. The memory preset ramps to a similar scale.
- 300mm fab equipment spending set to reach $227 billion by 2029 (opens in a new tab), Evertiq, reporting SEMI's 300mm Fab Outlook, 15 October 2026. The pace of new 300 mm fab and line start-ups that SEMI expects, the context for ramp planning. The figures are SEMI's; check the SEMI outlook itself for the full report.
Where it helps
The lifecycle stages this tool serves: 10 stages in 10 lifecycles.
More manufacturing, fab and supply tools
Method notev1Formula, ranges, sources and limits
Plans a fab or line ramp month by month from tool installs, qualification lag, an output ramp, cycle time and yield curves: wafer starts, wafers out, good die out, the month demand is met and the work in process.
Formula
first productive month of a tool set = install month + qualification lag; output share at age a = start share + (mature share − start share) x min(1, a / ramp months); capacity = existing line + sum of (tool sets x wafer starts a month per tool set x output share); wafer starts = min(capacity, demand); cycle time in months = cycle time in days / 30.4375; wafers leaving in month m = sum over start months s of starts(s) x overlap of [s − 1 + cycle time, s + cycle time) with [m − 1, m); cycle time learning: cycle time(s) = mature + (start − mature) x exp(−(s − 1) / time constant); yield(m) = mature − (mature − start) x exp(−(m − 1) / time constant); wafers out = wafers leaving x line yield; good die out = wafers out x die per wafer x die yield; work in process (Little) = wafer starts a month / 30.4375 x cycle time in days; counted work in process = previous + starts − wafers leaving
| Input | Unit | Range it holds for | Comes from |
|---|---|---|---|
| Plan length | months | 6 to 60 | Your figure |
| Tool set installs from | choice | linear, rows | Your figure |
| First install month | month | 1 to 60 | Your figure |
| Tool sets installed each month | tool sets | 0 to 50 | Your figure |
| Tool sets to install in total | tool sets, 0 for no limit | 0 to 500 | Your figure |
| Capacity of a tool set at full rate | wafer starts a month | 0 to 100000 | Your figure |
| Existing line already running | wafer starts a month | 0 to 500000 | Your figure |
| Qualification lag | months | 0 to 24 | Your figure |
| Output of a newly qualified tool set | % | 0 to 100 | Your figure |
| Output of a mature tool set | % | 0 to 100 | Your figure |
| Months to reach the mature output | months | 0 to 48 | Your figure |
| Demand from | choice | level, rows | Your figure |
| Demand in month 1 | wafer starts a month | from 0 | Your figure |
| Demand growth | % a month | -20 to 50 | Your figure |
| Demand ceiling | wafer starts a month, 0 for none | from 0 | Your figure |
| Cycle time | days | 1 to 365 | Your figure |
| Cycle time learning | choice | constant, learning | Your figure |
| Mature cycle time | days | 1 to 365 | Your figure |
| Cycle time learning time constant | months | 0 to 60 | Your figure |
| Line yield at the start | % | 0 to 100 | Your figure |
| Line yield when mature | % | 0 to 100 | Your figure |
| Line yield time constant | months | 0 to 60 | Your figure |
| Die yield at the start | % | 0 to 100 | Your figure |
| Die yield when mature | % | 0 to 100 | Your figure |
| Die yield time constant | months | 0 to 60 | Your figure |
| Die (or units) per wafer | die | from 0 | Your figure |
Outputs: Month demand is met (month); First month wafers come out (month); Peak work in process (Little's law) (wafers); Good die out over the plan (die).
Assumptions
- The planner follows the bottleneck: one kind of tool set whose capacity limits the whole line. A real line has many tool types, and the scarcest one sets the pace.
- Wafers of one month's starts all take the same cycle time; there is no spread of cycle times between lots and no queueing model.
- Wafers that fail the line yield are scrapped as they leave the line; the work in process counts them until then.
- Yields are applied by the month wafers leave, not the month they start. The learning curve is calendar based, so a change in the plan's start date does not shift it.
- The existing line (if any) runs at its full rate before month 1 with the starting cycle time, so its work in process is full at the start.
- All figures in the presets are editorial estimates for a typical case. Replace them with your own tool counts, qualification times, cycle times and yields.
Uncertainty: The arithmetic is exact; the result is as good as the install dates, qualification times, cycle times and yields entered. A real ramp slips more often than it runs early, so test the plan with a longer qualification lag and a longer cycle time.
Known limits
- Cycle time does not rise with load. In a real line, cycle time climbs steeply as utilisation approaches the limit; feed that in as a longer cycle time or a lower mature output share.
- Product mix, rework loops, reticle and tool re-use, and tool-by-tool constraints are not modelled; use a full line simulation for those.
- Fab building, cleanroom and utility readiness are assumed to be in place before tool installation; they add to the real schedule.
- Demand is a wafer start demand. Convert a die demand to wafers with the yields before entering it.
- Months are average months of 30.4375 days; a plan shorter than six months or longer than five years is outside what the tool is meant for.
Where the method comes from
The Hub's own month by month ramp arithmetic, using Little's law for work in process and the usual exponential learning shape for yield and cycle time. Presets are editorial and sit near the scale of the capacity targets named in the sources. (Editorial)
Reference example, checked by test: Worked example in this method note and in the tool's unit test: two tool sets of 1,000 wafer starts a month installed in month 1, qualified after one month at full output, demand 1,500 a month, a cycle time of exactly one month, no yield loss in the line, 50% die yield and 100 die per wafer, over six months. Month demand is met 2; First month wafers come out 3; Peak work in process (Little's law) 1,500; Good die out over the plan 300,000.
Versions
- v1 (2026-10-25): First version, built from the tool idea.
Terms on this page
- Fab
- The cleanroom facility where wafers are processed, together with its subfab utilities, chemical distribution and automation.
- WSPM
- The standard measure of fab capacity, counting wafers entering the line rather than good die leaving it.
- Cycle time
- The elapsed time a lot takes from wafer start to wafer out, including all queueing, not just processing.
- Yield
- The share of manufactured units that meet specification, counted at a given step: line yield through the fab, die yield at wafer sort, and assembly and final test yield after packaging.
- Wafer
- A thin, polished disc of single-crystal silicon on which hundreds or thousands of identical circuits are built at the same time.
- Die
- A single rectangular block of processed silicon cut from a finished wafer, carrying one complete circuit.
- WIP
- The total quantity of wafers currently released into the line but not yet completed, held at various process steps.
- Line yield
- The fraction of wafers started that emerge from the fab intact and eligible for test, lost to breakage, misprocessing, scrap and hold decisions rather than to die-level defects.
- Die yield
- The fraction of die on a wafer that pass electrical probe test, the headline yield number for a product.
- Yield learning curve
- The empirical trajectory of yield versus time or cumulative volume for a technology, typically steep early and flattening as random defect limits are approached.
- Learning curve
- The steady fall in unit cost as cumulative production grows, driven by yield improvement, cycle-time reduction and equipment productivity.
- Utilisation
- The fraction of available placeable core area consumed by cell and macro instances, typically quoted as a percentage.
Need help acting on this?
Tell us what you're after and leave your email: we'll connect you with the right person on the Hub.