We lost a big project back in 2018. The customer needed PEEK brackets for an APU mounting system, and the critical bore diameter had to be held to ±0.025mm. We ran our T1 samples, first 50 pieces came out good, everything looked fine. Then we started production runs and our Cpk dropped to 0.9. Turned out our mold temperature was fluctuating ±4°C when PEEK at that tolerance level really needs ±1.5°C maximum.
That cost us the contract and about $180K in tooling that ended up being useless.
I'm bringing this up because most supplier websites only show you the wins. But we learned more from that failure than from a lot of projects that went smoothly. These days our mold temperature control runs at ±1.5°C using in-mold thermocouples, and we don't quote ±0.025mm on PEEK anymore without doing a proper capability study first. Some jobs we just turn down. That's reality in this business.
Pricing - Real Numbers
Everyone wants to know about cost. Here's what we actually charged for aerospace projects in 2023-2024:
Fuel line bracket
Avionics housing
Hydraulic connector body
Structural clip
The hydraulic connector taught us something important. Customer originally specified ±0.015mm on the seal groove. We ran mold flow simulation and showed them the data - injection molding physics just won't let you guarantee that tolerance in production. They agreed to relax it to ±0.02mm with post-machining on the critical OD, and the project succeeded. If we'd just said "yes, we can do it" to win the PO, we'd have ended up with another 2018 situation.
Tooling costs vary a lot more than most buyers expect. That $26K clip mold and the $128K connector mold look like they're from different worlds, but the cost drivers make sense when you look at the design. The connector needed 6 hydraulic core pulls, hot runner with valve gates, and we had to use hardened H13 instead of P20 steel. When customers ask why a mold is expensive, we walk through the design together and it usually makes sense once you see where the money actually goes.
The PEEK Problem
About 70% of our aerospace inquiries specify PEEK. Maybe 40% actually need it.
PEEK earns its premium in three scenarios: continuous exposure above 180°C, direct contact with hydraulic fluids like Skydrol or jet fuel, or structural loads where you're replacing aluminum. Outside those conditions, you're paying 3-4x the material cost for properties you're not using.
Last quarter we had a customer come to us with overhead bin latches for aircraft cabins, spec'd in PEEK. I asked about the operating environment - room temperature, no chemical exposure, just moderate mechanical load. We suggested switching to ULTEM 2300. For this specific application, the mechanical performance is equivalent, but material cost drops from $38/kg to $14/kg. That saved them roughly $2.80 per part on 200,000 pieces annual volume.
$560K per year
they get to keep instead of spending on over-engineered material.
We're not pushing cheaper options just to win bids. We push appropriate options because over-specification creates problems beyond just cost. PEEK shrinks more than ULTEM, higher processing temperature means a tighter process window, more things that can go wrong. Matching material to actual requirements isn't just about economics - it's risk management.
For projects that genuinely need PEEK, we process Victrex 450G and 150G regularly. Also run Solvay KetaSpire when customers prefer it. Our dryers hold material at 150°C for minimum 4 hours, verified by moisture analyzer before every production run. If residual moisture is above 0.02%, we reject the material even if it means delaying production start. Moisture causes splay marks and degrades mechanical properties.
What We Can Actually Hold for Tolerance
±0.05mm general dimensions
We deliver this reliably with Cpk 1.33+ under standard production monitoring.
±0.03mm critical features
This requires process validation during T1 phase and statistical control charting in production.
Tighter than ±0.025mm?
We'll quote it, but we'll also tell you upfront it needs secondary machining allowance. Pure injection molding at this level depends on too many factors we can't fully control: material lot variation, ambient temperature in the facility, mold wear over time. Some suppliers will promise anything to get the purchase order. We'd rather set realistic expectations from the start.
Getting to consistent ±0.03mm capability took us about three years. The core elements: upgraded to SINGLE oil temperature control units with ±1.5°C stability, installed cavity pressure sensors on all aerospace molds, and built material lot tracking into our ERP system so we can correlate dimensional data with specific resin batches. When a dimension starts drifting now, we can usually identify why within the same shift.
Our Facility and Capabilities
We run 12 injection molding machines from 80T to 650T, all Haitian Mars series purchased between 2019-2022. In-house tool room with 3 Makino machining centers handles all our mold manufacturing. We control quality and schedule from steel cutting through T1 samples - no outsourcing on aerospace tooling.
We know AS9100 is the basic entry requirement for most aerospace programs. Some customers work with us now on commercial aerospace adjacent applications while we complete certification. Others wait until we have the certificate in hand. Both approaches make sense depending on your internal supplier qualification requirements.
Our quality documentation follows AS9102 format for FAI whether or not we have the certification yet. We adopted this standard because it's the right way to document aerospace parts, not because an auditor required it. Sample FAI packages are available under NDA.
Typical timeline from PO to production approval runs 16-20 weeks. That breaks down to about 2 weeks for DFM and design finalization, 8-10 weeks tooling manufacture, 2 weeks for T1 and dimensional verification, then 2-4 weeks for customer FAI review. We've compressed this to 12 weeks for urgent programs, but that required premium pricing and schedule risk we don't normally accept.
What Doesn't Work for Us
High volume commodity parts
Two million ABS clips at $0.15 each? We're not competitive. Our overhead structure and quality systems add cost that makes sense for aerospace precision work but prices us out of pure commodity business.
Materials we haven't run before
First-time processing of a new resin takes learning curve time. If your project requires a specialty compound we've never touched, we'll either decline or quote additional process development time. We don't experiment on customer programs.
Unrealistic schedules
In order to let you know more about our products and whether they meet your
"Tooling in 4 weeks" means cutting corners on steel hardening, skipping proper mold trials, and shipping problems downstream. We learned this the expensive way.
How to Work With Us
Send drawings if you have them - even concept sketches help. Tell us about operating environment, annual volume, and the specifications you need to hit. If you're converting from machined parts, share your current part cost so we can show you where the breakeven point falls.
We'll respond with DFM observations, preliminary pricing, and an honest assessment of whether it's a good fit. Sometimes the honest answer is "this isn't right for us, you should try calling [another supplier]." Better to point you toward the right supplier than to struggle with a job that doesn't match our strengths.
Contact:
✉ mike@abismold.com
✆ WhatsApp: +86 137 5101 7242
If you're visiting Shenzhen, facility tours are available. The shop floor tells you more than any website can.