Experience from Co-curing

Tom Johnson, Eric Anderssen, Mario Cepeda, Gil Gilchreise, Carl HaberLawrence Berkeley National Lab

11-Nov-2010 AUW

Outline

• Introduction• Build process• Cable metrology• Stave metrology• Plans and Ideas

Introduction• Early tests with lamination of bus cables to

stave cores were successful but difficult• Concern with QC on a large number of parts and

loss of expensive components• Have studied co-cure of bus cables to facings

followed by core lamination• Shape metrology and bus cable metrology• Have used these in subsequent electrical studies

as well

Composites Facility

Configuration• Lay up is 90-0-90, better for thermal conductivity

across the module and reduced local curvature• Overall twist can be handled with fixtures

Cable Inspection

• Inspect Cable for Damage prior to curing.

• Broken traces• Wrinkles• Laminated dirt• Discoloration

Clean Top Surface and Apply 1 mil Flashbreaker Tape

Surface Prep of Backside Prior to Bonding

- Use ScotchBrite pad to Abrade Backside of Cable. - Clean with Alcohol

1 2 3

4 5 6

-Part is layed up on a glass plate covered with A4000 release plastic.-0 deg ply on bottom and 90 deg ply on top.-The white is the paper backing of the prepreg.-Once layed out they are taped together to hold in place and covered with another layer of A4000. (4,5)-Plys are then compacted under vacuum to stick the plys together. (6)

- After compaction, flip part so the 90’s are down against the glass. - Remove the paper backing from the 0 ply. (1)- Stack the next 90 deg ply the same as the previous 90 ply. (2)- Tape together, A4000 release plastic and compact (3,4)

1 2

3 4

- After compaction remove the exposed paper backing.-Apply a single layer of pealply to this side. Pealply gives the surface the desired texture and absorbs extra resin

12

3

4

5

- Flip part again so the pealply is now against the glass and remove the paper backing. (1)-Position cable and apply it to the stack. (2)- Add a small piece of pealply to the remaining exposed carbon on the end. (3)- Cover stack with A4000 release sheet. (4)- Compact. (5)

- After compaction, Flip the entire stack so the cable is now against the flat glass surface. (1)-Cover with a fresh piece of A4000 making sure to have no wrinkles. (2)- Add layer of breather (no wrinkles) and put under final vacuum bag for cooking. (3)

12

3

Stave Assembly• Stave assembly proceeds similar to the usual case• Special 1.5 meter vacuum fixture was prepared to

hold co-cure facing/cable flat for lamination to core

-The honeycomb needs to have one side ground/sanded flat prior to bonding onthe first co-cure cable.Can be ground on a surface grinder using thick “Airtech” double sided tape.Or as we did if a grinder is not available at the time.Flat Mic6 plate and two spacer blocks. Sandpaper is applied to one side of the mic6 .Then honeycomb is sanded flat

Each piece of honeycomb (HC) is dipped into a thin pool of epoxy and let sit for 30 seconds to allow the epoxy to wick onto the HC. Its then positioned onto the co-cure cable that is securely held down flat by the vacuum chuck. An aluminum plate with a silicon sheet is placed on top the HC and left overnight to cure. You can see the aluminum plate w/silicon and epoxy pool in the next slides.

Red Silicon and aluminum plate.

Left to cure overnight.

This first side with bonded HC is held flat by vacuum. Second side HC (exposed) is then surface ground or sanded.

Completed Honeycomb side . Epoxy Pool for dipping Honeycomb.

The entire stave will be set into the pool for 30 sec then placed on top the 2nd co-cure facesheet that will be held flat by the vacuum chuck

Align top co-cure to bottom co-cure. Place aluminum plate with silicon sheet on top. Add additional weight as needed.Red silicon can be seen under the alum plate.

After minimum of 24 hour cure can be remover from fixture. Hysol EA 9396 epoxy has 80% cure at room temp in 24 hours. 100% cure room temp 72 hours.

Post Processing• Completed stave is cleaned with an atmospheric plasma

system• Cleanliness can be checked with a water droplet system• Good wire bonding performance observed along the

entire length

Full stave with 24 BCC boards mounted and bonded, EOS card bonded at the left

Cable Cleaning• Find that wire bonding

ease and reliability improves with plasma cleaning

• Have procured and installed a plasma cleaning and scanning system

• Plasma generator from SurfX

• Custom large area scanner

Droplet System

Metrology

• Optical metrology (OGP) was performed on free cables, cable+facing co-cure, and completed stave

• Metric was module-to-module feature, nominally 98 mm

• Touch-probe metrology was performed on the completed stave to measure full profile

Optical Metrology of Cable and Stave

full stave

Data on Full Length Parts

Findings

• Free bus cables (or co-cure to facing) are consistently short by ~100 m per segment– Cummulative error is ~ 1 mm

• If true we would need to adjust in artwork• On complete stave some of this error is

recovered, particularly near the ends.

Stave Metrology Flatness

• Thickness measured with calipers• Flatness measured on a CMM with a touch

probe scan.• Stave supported flat (maximal sag orientation)

at the Airy points• Consistent built-in bowing measured from

both sides: 735 microns

Thickness

• Well correlated side-to-side: 10-20 m• Systematic ally thicker at far end: 50-60 m• Maximum deviation point-to-point = 102 m• Potentially better if HC had been ground rather than sanded

6.188 6.195 6.159 6.173 6.191 6.204 6.251 6.230 6.236

6.200 6.183 6.162 6.154 6.200 6.183 6.256 6.212 6.200 mm

CMM Scan Data

Discussion• What drives this bow?• Is it related to the cable differential linear

measurements on the stave?• How much CTE difference would drive this?• Does it matter? – barrel mounts would compensate• In the future could co-cure both cables and facings

on the same sheet and at the same time– Ensure identical stack and thermal profile