Best Glue for ABS Plastic: What a JSP Quality Manager Learned From a $22,000 Mistake

An Email That Began With Glue

Last February, a maintenance manager named Drake sent us a message through the jsp app. The subject line was a question we see a lot: best glue for ABS plastic. My first instinct was to answer and move on. Then I looked at the photos he attached.

I am the quality manager at JSP. I review every replacement part order before it goes to production—roughly 250 unique jobs a year, from a single sheet of polycarbonate to a full custom molding run. In 2024, I rejected about 11% of first submissions. Most were documentation issues. This one was not documentation, and it started with a glue question.

Drake was not a stranger to us. A few months earlier, he had ordered a clear panel from our polycarbonate store for a different machine. That order took him five minutes. The ABS guard question was not going to be that simple.

Drake's ABS Plastic Guards

The guard protected a row of filling heads at his plant and kept washdown spray off the electronics underneath. It had been molded from ABS plastic years ago, by an equipment maker that no longer offered it. Every few months, a crack appeared at the same corner near the right mounting ear.

His team glued it, clamped it, and put it back into service. Then the crack came back, usually right beside the old repair.

By the time he emailed us, Drake had tried three different products he found in searches for best glue for ABS plastic. One epoxy peeled. One solvent cement grabbed hard but did not stop the crack from growing at the edge of the patch. The online videos made it look simple. The plant floor was less forgiving.

Here is the thing about ABS: solvent welding can work very well. A proper solvent cement does not just sit on the surface. It softens the plastic, and when the two pieces are pressed together, the bond can be nearly as strong as the parent material. That is why so many people assume the answer to a cracked ABS part is a better adhesive. But the adhesive is rarely the weak link.

Drake's guards were cracking because the inside corner near the mounting ear was too sharp. Every vibration and every temperature swing loaded that corner. The crack was the plastic telling us it could not handle the stress. Gluing it was like taping over a warning light. Workable for a while, until it is not.

If the same part keeps cracking in the same place, the glue was never the problem. The geometry was.

I told Drake that if the guard were a non-structural cover, he could keep repairing it. But on a machine running three shifts, the same corner would keep failing. What he needed was a replacement with a larger radius at that corner and a mounting ear designed to spread the load.

He asked if JSP made those. Yes. Most JSP products start this way: an old machine, a discontinued part, and a failure pattern that keeps repeating. We do not claim our guards are identical to the OEM version. We claim they fit and perform, and when the original design has a flaw, we fix it.

A Drawing That Looked Too Good

Drake ordered 400 guards. The old OEM drawing had disappeared years ago, so we used our standard replacement part process: the customer measures an existing part, uploads the dimensions, and our engineers review them before tooling starts.

He uploaded everything through the jsp app. There were photos, an ABS material callout, and one measurement I stared at too long without saying anything: the center-to-center distance between the two mounting holes. He listed it as 112 mm.

That number was a red flag. The photos showed a thin line of old solvent cement running across the same area. The crack had been pulled closed during a previous repair, which shifted the hole position. A repaired plastic part is not a reliable source of truth for dimensions.

I went back and forth with our production scheduler for most of an afternoon. Waiting for a field measurement meant losing three days, and Drake was already under pressure to get the line back to full speed. On paper, it was a no-brainer: the part was simple, the material was common, and the risk seemed low. My gut said otherwise.

We moved ahead anyway.

I had told myself for months that we needed a rule: never accept dimensions from a repaired plastic part without checking the mating metal surface. I did not write the rule down. I did not add it to the jsp app. It stayed in my head, which is exactly where process gaps live until they cost real money.

The $22,000 Redo

The molded guards looked right. The first article passed every dimension on the drawing. The material was correct, the color matched, and the surface finish was clean. The only problem was that the drawing was built on a number that came from a damaged part.

The actual distance between the mounting studs on Drake's machine frame was 118 mm. The old repair had pulled the guard slightly out of shape, so the 112 mm measurement carried that distortion into the new tooling. The first guard sat cocked on the frame. A technician loosened the bolts, wiggled it, and called us. The second one did the same.

Every guard in the first lot had the same flaw, because every guard came from the same wrong dimension. We pulled the order, measured the frame, and remolded the lot. The redo cost about $22,000 in material, tooling time, and expedited freight. JSP covered it. It was our process failure, not Drake's.

Drake was patient about it, which made it worse. He did not even ask us to cover the freight. I told him we would anyway. That is what the warranty on a process looks like.

The Rule I Finally Put in the jsp App

After that order, I did what I should have done months earlier. I wrote the verification rule into the jsp app so it could not be skipped.

Now, when someone requests a replacement part and indicates the original is damaged or has been repaired, the jsp app asks a required question. If the answer is yes, the form blocks submission until the customer adds field measurements of the mating metal surface—not the plastic part.

It sounds like a small change, and it prevents one specific error. But it forces everyone to be honest: a broken plastic part tells you where it used to be, not where it needs to be.

Best Glue for ABS Plastic: My Honest Answer

People still ask me for the best glue for ABS plastic, and I still give a real answer. If the part is not load-bearing, a solvent-based ABS cement is the best choice. Clean the edges, use primer if the manufacturer recommends one, clamp the pieces, and let the joint cure. Super glue can craze ABS, and most epoxies are only a mechanical stick. Solvent cement creates a chemical weld, which is what you want for a hairline crack in a low-stress cover.

But if the same part keeps failing in the same place, focus on why it failed, not on what to glue with. A bigger radius, a thicker wall section, or a different mounting design will do more than any adhesive. The best glue in the world does not fix a stress concentration.

I can only speak to industrial applications like Drake's. A bracket on a vibrating machine has different demands than an electronics enclosure that sits on a shelf. If your part is the second kind, use solvent cement and move on. Do not let a molded replacement parts company talk you into something you do not need. That includes ours.

The best glue for ABS plastic is the one that matches the job. Start with the failure, not the adhesive. When the failure is stress, and the replacement is genuinely needed, that is where JSP products come in. We will make the part, and this time, we will measure the steel first.