TFT LCD Capacity Creep
July 6th, 2007By Charles Annis, Vice President of FPD Manufacturing Research
In Q2’07 unrestricted large area capacity swung into a tight environment after multiple consecutive quarters of being characterized by oversupply. Despite across the board rising panel prices in Q2’07 and Q3’07, panel makers have maintained a generally conservative attitude about new capital outlays. With the emergence of the loose supply in the first half and tight in the second, highly seasonal supply/demand pattern, TFT LCD producers are focusing on stretching current capacity to meet demand rather than building expensive new lines.
Stretching” capacity, or increasing productivity, without major investment is a achieved by increasing input capacity of a line or increasing yield. Input capacity may be increased by throughput improvement, mask reduction (i.e. 5 to 4), faster automation, purchasing single machines in bottleneck locations or reducing job changes (i.e. mass produce fixed panel designs and don’t change them). Output can be further increased by yield improvement. This may be actual yield improvement gained from reduced job changes, increased inspection/repair and improved processes. Or it could be virtual yield improvement gained from lower binning standards. When supply is tight and prices rise, panel makers can sell lower quality panels. A 2% shortage can be rectified by a 2% increase in panels binned to GOOD.
Panel makers may “overbuy equipment” when starting up a fab. That is they buy enough machines to assume a safety margin to reach a target capacity. One equipment maker recently suggested customers buy 1.6X the number of machines that their through-put and up-time would allow. This is probably on the extreme side, but simple calculations assuming very high uptime show that typically coater/developer and lithography lines have 15% or more capacity than they serve. One AMHS vendor suggested that fab automation systems are typically budgeted for 70% maximum capacity with a minimum 10% safety margin required to run in production. That is, as far as stockers and cranes go, panel makers can increase capacity maybe 15-20% and still keep cassettes moving around the fab.
As a fab matures and reaches high yield, panel makers may be able to speed up processing and may be able to reduce some inspection and repair steps. However, reduced downtime for preventive maintenance may become a concern. Backend processes such as cell inspection can also become bottlenecks.
All of these issues suggest that capacity is quite variable, can be increased over time, and can be increased to meet peaks in demand.
How far can capacity by stretched without major new investment? Industry-wide, it’s possible to imagine that total capacity could be stretched up to maybe 10%, while capacity on individual lines can be stretched to substantially higher levels depending on the type of productivity enhancements implemented. Component capacity may not be able to be stretched at the same rate as panel capacity. If panel demand is higher than expected it can cause concern of component tightness. Stretching capacity will affect the supply/demand balance in 2007 and help panel makers meet rising demand to some degree. Even so, capacity can only be stretched so far; in 2008 and 2009, we expect that equipment spending will increase 14% and then 13% above 2007 as panel makers need to build new fabs to meet continuously rising demand.
www.displaysearch.com