New energy vehicles have moved to 800V architectures, and the parts that live inside that world, high-voltage connectors, busbars, the housings that keep current where it belongs, now operate under a completely different set of rules. Engineers chasing tighter packaging and higher energy density will tell you the thing they fear most is not a part snapping under load. It is the insulation failing when nobody is watching.
Most people know nylon absorbs water. What gets hand-waved away is what that water actually does to the electrical behavior. People say the performance goes down, and they leave it there. That vague word, down, hides the real story. Under the right conditions nylon does not just lose a little insulation. It opens the door to leakage paths that should never exist, and it does so quietly, while the car is sitting in a humid garage or after a coolant line weeps for a few months.
This piece digs into what happens to the electrical properties of polyamide, PA, once it moves from dry-as-molded to conditioned, and why the difference decides whether your high-voltage design is a robust solution or a future incident report.
The amide group is the weak spot, and it was there from day one
Nylon’s backbone carries amide groups, those -CONH- units, and they are genuinely hungry for water. That is not a defect you can tweak out with a nicer drying schedule. It is built into the chemistry. When a part leaves the mold it is dry, and everyone likes the numbers from that state. Give it a few weeks in ordinary air and it reaches a moisture equilibrium, the conditioned state, and the inside of the polymer is no longer the same material it was on the spec sheet.
So the first move is to stop talking about dry parts as if they were the real product. The part your customer drives with is the wet one.
Volume resistivity takes the dive nobody budgets for
Water that migrates into the polymer does not sit still. It becomes the carrier for ion movement. As humidity climbs, volume resistivity falls, and it falls hard. Push the material into a hot, damp environment and the change is not a gentle slope. The part drifts from behaving like an insulator toward behaving like a semiconductor. That is the moment leakage current stops being a rounding error and starts becoming the dominant number on the test bench.
For a 400V system this was annoying. For an 800V system it is a different league, because the margin you were counting on has already been spent by the moisture you ignored.
Dielectric strength, the floor under your safety case, gets eaten
The breakdown strength, measured in kV per mm, is the line you do not cross. It is the bottom of the safety argument for any high-voltage component. Ordinary PA loses a meaningful chunk of that capability once it is moist and warm. The curves for PA6 and PA12 tell the same uncomfortable story, the material that looked safe at room temperature in a datasheet loses a visible amount of its withstand voltage after it has been sitting in heat and humidity.
The danger here is subtle. A designer picks the material based on dry data, validates at the bench on dry samples, and ships. The field part runs warm, picks up moisture, and the real breakdown margin is thinner than the certificate implied. Nobody feels it until something arcs.
Dielectric constant and loss factor wander, and your signal pays for it
High-frequency work, onboard 5G and 6G antennas, high-speed data links inside the vehicle, cares about two more numbers. The dielectric constant, Dk, and the dissipation factor, Df. Water moves both of them. Each added percentage point of moisture shifts the phase delay and increases signal loss, and at high frequencies that shows up as dropped packets or a collapsed signal-to-noise ratio.
This is the part that surprises people. They expect water to hurt the high-voltage safety. They do not expect it to quietly degrade the communication link sitting right next to the power line, in the same humid, hot enclosure.
CTI is the one place water gets blamed unfairly
Here is a correction worth making out loud. The Comparative Tracking Index, CTI, the number that tells you how well a material resists forming a conductive path under wet, contaminated conditions, is not driven by how much water is inside the bulk of the plastic. CTI is set by the polymer’s chemical structure, how it chars, the temperature, the surface tension and roughness, and the pigments and additives, including flame retardants, that got mixed in.
So if someone tells you a material lost CTI because it absorbed water, that is the wrong mechanism. The moisture on the surface and the contaminants matter for tracking, yes, but the internal water content is not the lever. This distinction matters when you are comparing materials, because it points you at the right design knobs instead of the wrong ones.
Stop saying all nylon is the same, the data says otherwise
The trap is treating every polyamide as one material. The level of modification and the choice of base resin decide whether you built a high-voltage solution or a liability.
Conventional PA66 runs into a wall
Under ISO 62 at 23°C and 50% relative humidity, PA66 pulls in a lot of water. Its mechanical stiffness and its electrical behavior both swing with that moisture. Look at the retention curves and PA66 with 30% glass fiber, PA66-GF30, holds far less of its modulus after conditioning than PPA or PPS do. The electrical side follows the same pattern. You can spec PA66, and it will work, until the environment stops being the friendly lab condition it was validated under.
PPA pulls ahead on more than glass transition temperature
Pull the published data from PPA grades and put them next to PA66. A PPA such as the Grivory HT1VA-4 HY type runs a glass transition temperature roughly 55°C above PA66, and it picks up water much more slowly. That matters more than it sounds. In a coolant leak or a brutal temperature-and-humidity excursion, the slower uptake means the insulating structure stays intact longer. The higher Tg is not a line on a chart. It is the reason the part is still an insulator at the moment it matters.
PPS sits at the far end, with moisture absorption near zero and electrical properties that barely notice the weather. The trade is cost and processing, but for the spots where failure is not an option, that trade starts to look cheap.
How to actually avoid the water trap
Material selection for an 800V system cannot live on the initial values printed on a datasheet. It has to live on what survives after aging. Three habits separate the designs that hold up from the ones that make the news.
First, stop trusting the dry number. Almost every supplier hands you electrical data measured dry. The real part reaches a conditioned equilibrium before it is even assembled, and certainly before it has been on the road for a year. Ask for the electrical numbers in the conditioned state. If a vendor cannot produce them, that silence is your answer.
Second, watch CTI like a threshold, not a footnote. For high-voltage connectors the practical bar right now is CTI 600V. That is the number that tells you the part can resist a conductive path forming under wet, dirty, real-world conditions. Anything below it is a compromise you should accept with open eyes.
Third, demand the long curves, not the headline. Ask to see what happens after 1000 hours at 85°C and 85% relative humidity, or after soaking in the specific electrolyte your system uses. If the retention curve stays flat and holds above 80% after that abuse, you are looking at a material worth specifying. If it falls off a cliff at hour 200, the pretty dry-number on page one was a distraction.
The honest takeaway
Nylon’s thirst for water is a material fact, not an excuse for insulation failure. In the fight to make 800V platforms work, choosing the resin is choosing the reliability. Drop the attachment to plain nylon. Steer toward high-performance PPA, PPS, or purpose-built modified polyamides, chosen for the actual temperature-and-humidity curve the part will live in. That is the most responsible move you can make for electrical safety. The data has been sitting in supplier folders for years. The job is to ask for the conditioned version, and then to believe it.