How double shot key legends hold up when the second colour bleeds

Lock the substrate pair before you cut a single legend

Two shots, two materials, and an interface that never quite goes away. That interface is where legends live or die. Before you approve artwork, before anyone cuts a tool, get the substrate pair fixed in writing: which polymer goes down first, which goes over it, and what happens to the bond line between them. Change either resin after the tool is cut and you are not adjusting a process. You are re-running a qualification.

Most remote-control programmes default to ABS over ABS for double shot key legends, or ABS over a softer TPE where the key needs a tactile crown. ABS on ABS bonds well because the second shot partly dissolves the first surface. That is also the problem: run your melt temperature hot and the same partial dissolution lets the legend colour migrate into the base. PC over PC gives a harder, more scratch-resistant key face and holds dimensions tighter, but both shots need higher barrel temperatures and a hotter tool, and the bond is far less forgiving of moisture. PC over ABS turns up a lot on premium keypads, and it is the pair most likely to fail a soak test if anyone skips the drying.

Settle three things in the specification, not in conversation:

• Resin grades by supplier and grade code for both shots, with the colour masterbatch named.
• Which shot is the legend and which is the body. On a light legend in a dark key, the legend is usually the second shot; on a dark legend in a light key, it is often the first. The order changes your tool layout and the direction the bleed runs.
• Whether the interface is mechanical (a dovetail or undercut) or purely chemical. Pure chemical bonding tolerates no contamination at all; a shallow mechanical key gives you a second line of defence at the cost of tool complexity.

Pass/fail for this decision: you can name both grades, both shot orders and the bond mechanism on one page, and your toolmaker has signed that page before electrodes are cut. If any of the three is still "to be confirmed" when the tool drawing is released, stop. The recovery cost is a new tool insert, not a parameter tweak.

Set the bleed allowance, then prove it with a cross-section

Bleed is the visible spread of the second colour past the intended edge of the legend. Every double-shot legend has some. The question is how much you allow and how you measure it at the tool, not at final inspection.

Specify the allowance as a dimensional band on the drawing: legend edge position ±0.08 mm, say, with a maximum visible bleed of 0.10 mm on any character, measured on a moulded part rather than on a rendering. Take a character stroke of 0.6 mm — roughly the minimum you can hold reliably on a keycap at this scale — and a 0.10 mm bleed eats a third of the stroke on one side. That is the number to argue about with your ID team early, because it decides the typeface weight and the minimum stroke in the artwork.

The executable check is a potted cross-section. Take three parts from the first shot of a trial run, section them through the worst-case character (usually a diagonal or a tight curve), polish, and measure the interface position under a toolmaker's microscope at 50×. Pass: the measured interface sits inside the drawing band on all three, and the second colour has not wicked past the mechanical edge into the first-shot surface. Fail: any section shows colour beyond the band, or shows a feathered interface instead of a defined line. Feathered means you have a temperature or viscosity problem, not a tool problem — keep your hands off the steel for now.

Keep the sectioned parts. They are your baseline. When a production batch is disputed six months later, the argument gets settled by comparing a new section against the potted original, not by opinions.

Hold the two levers: second-shot viscosity and mould temperature

Once the tool is cut, bleed is controlled by two variables that interact. Get either wrong and the section you approved will not repeat.

Melt viscosity of the second shot

Lower viscosity flows further into the interface before it freezes. Useful for filling thin legend channels, disastrous for keeping the edge sharp. Run the second shot at the upper end of the supplier's recommended melt range for fill, then back off until short shots appear, then come up two or three degrees. Record the actual barrel profile, not the setpoint — on older machines the two diverge. If your moulder cannot show you a verified barrel profile on the setup sheet, that is your first finding.

Mould temperature and the freeze window

The interface freezes from the tool surface inward. A hotter tool widens the window in which the second melt can move the first surface, which raises bleed; a colder tool shortens fill and risks a cold weld at the legend edge. Typical ABS double-shot tools run at 60–80 °C on the cavity side; PC needs more. The pass/fail criterion: hold tool temperature within ±3 °C of the approved setting across a full shift, logged, and re-check the cross-section at the start, middle and end of that shift. If any of the three sections drifts outside the band, you have a temperature-control problem — check the TCU capacity and the circuit layout before you blame the resin.

One more lever people forget: the delay between shots. If the first shot has cooled below its softening window before the second arrives, the bond becomes mechanical only and the edge behaves differently. Fix the delay in the process sheet with a tolerance, not as a "typical" value.

Read the first-off sample with a loupe, not a glance

First-off inspection on a double-shot keypad is a ten-minute job done properly and a five-second job done badly. Do it in this order:

1. Lay the part on a white card under a 10× loupe with raking light from one side. Raking light shows the interface step; straight-on light hides it.
2. Measure legend edge position on the five worst characters against the drawing band. Any single character outside the band fails the sample, whatever the average says.
3. Check for a visible step at the interface with a fingertip, then a profile projector. A step you can feel will collect dirt in service and read as wear within months.
4. Section one part from the sample and compare it to the approved potted baseline. Note the interface position on the setup sheet.

The 0.05 mm rule: if the interface step exceeds 0.05 mm on any character, reject the sample. Below that, the step is generally invisible at arm's length and does not trap soil meaningfully. Above it, you will see it in the field and your customer will see it in a returns bin.

Do not sign off a first-off sample from a single cavity on a multi-cavity tool. Run all cavities, mark each part with its cavity number, and keep the set. Cavity-to-cavity variation in the interface position is the single most common cause of a "the tool was fine in trial" argument later.

Run the bleed test to failure before you release the tool

Two tests, both destructive, both cheap compared to a field failure.

Abrasion. Take ten moulded keys and run them on a reciprocating abrasion rig against a defined cloth, counting cycles until the legend edge shows measurable loss of definition. There is no universal pass number — it depends on your key material and your customer's expected life — but you must set one before you test. A reasonable starting point for a consumer remote is 500 cycles with the legend still legible and no bleed line widening beyond the drawing band. If your first sample fails at 200, the bond is mechanical only or the second shot is under-packed; go back to the section and look at the interface line.

Solvent soak. Immerse five keys in isopropyl alcohol for 24 hours at room temperature, then dry and re-inspect under the loupe. Pass: no colour lift at the legend edge, no softening or clouding of the second shot, no visible change to the interface. Look specifically for two failure modes — a faint halo around each character (the second colour has started to release at the edge) and a chalky film on the key face (attack on the base resin). Both mean the resin pair or the moulding condition is wrong, and neither shows up in a drop test.

Write both results on the qualification record with the actual cycle count and the soak duration. A record that says "abrasion test passed" without a number is not a record.

Where legends bleed in production, and what it looks like on the line

The failures cluster in a few predictable places. Learn to recognise them by sight.

• A halo on the first few shots after a colour change. Purge is incomplete and the previous masterbatch is still in the barrel. The operator sees a slightly darker fringe on the legend edge for perhaps a dozen cycles, then it clears. The cure is a written purge routine with a scrap count, not operator judgement.
• Bleed that appears only on one side of the tool. Almost always a temperature imbalance between circuits, or a blocked or short-circuited line in one half. It shows as a gradient across the part: sharp on the left, feathered on the right.
• Bleed that starts after a weekend shutdown. Cold tool, first shots run slow, and the interface behaves differently until the tool stabilises. Specify a scrap allowance at start-up and a warm-up routine; do not let the first hour of a Monday shift reach the customer.
• Feathering on the deepest character only. Short shot in the second colour. Gas trap or insufficient pack pressure at the far end of the legend channel. It will look acceptable in the hand and fail a section every time.
• Colour shift rather than geometric bleed. The legend is in the right place but looks muddy. Masterbatch dispersion or overdosing at the machine. Check the dosing record before you touch the tool.

The pattern to watch for: bleed that is stable within a shift but drifts across a week is a process-control problem. Bleed that varies within a single shift is a machine or material problem. Those two need different responses, and mixing them up wastes a week of tooling time.

Verify with a signal, not an opinion

Pick one measurable that the line reports every shift: the interface position on a sectioned part from a fixed cavity, measured against the drawing band. Plot it. A process in control shows a stable band with normal scatter. A process drifting toward failure shows a slow shift in the mean before any single part fails — that is your early warning, and it is the whole point of measuring.

Back it with two secondary signals: the scrap count at start-up after a colour change, and the number of loupe rejections per thousand parts at final inspection. If both are flat and the section data is stable, the process is in control. If the section data is stable but loupe rejections climb, your inspection criterion has drifted from the drawing — recalibrate the inspectors against the potted baseline before you change a single machine parameter.

If you are sourcing this line from an ODM and want the qualification record reviewed against your own drawing, send the specification to peter.sun@serow-tech.com. Replies come back within one working day. Bring the tool drawing and the resin grades; without those, the conversation is guesswork.