The Regrind Disaster That Cost Us a Weekend and a Customer

We had been sending our carbide drills to Great Lakes Tool Regrind for almost six years. They were one of those shops that just worked — decent pricing, reasonable turnaround, never gave us trouble. A few hundred dollars a month, routine stuff, nobody thought twice about it. That is probably why when it went bad, it went so bad so quietly.

It was a Tuesday in late February, and I had a rush order sitting on my desk from a guy named Pete over at Midwest Aerospace. Forty-eight pieces of 15-5 PH stainless, through-hole, 0.3125-inch diameter, 2.5 inches deep. Nothing exotic. We had run this exact job maybe a dozen times over the past two years. Standard operating procedure: spot drill with a 120-degree center drill, then peck with a 3-flute OSG carbide drill at 220 SFM and 0.004 IPR, clear the floor with compressed air, and send the parts out for deburr. Cycle time was eight minutes per hole, roughly three hours per part. It was supposed to be a Friday-to-Saturday push.

I grabbed a set of twelve 0.3125-inch OSG carbide drills out of the regrind bin. They had just come back from Great Lakes three days earlier. Looked beautiful — fresh gold TiAlN coating, crisp cutting edges, the little regrind sticker still on the shanks with the batch number handwritten in marker. I loaded the first one into a CAT40 hydraulic holder, indicated it in at less than 0.0002 TIR, touched it off on the Haas VF-4 with a Haimer 3D sensor, and hit the green button.

The first few holes looked good. Chips were coming off clean, spindle load was steady at 28 percent, surface finish looked right. I ran twenty holes without stopping, checked the first part with a plug gauge, and it passed. Everything looked normal.

About thirty holes in, I heard it. That sound every machinist knows in their gut — the high-pitched squeal that says your drill is rubbing instead of cutting. I hit feed hold, popped the spindle up, and looked. The hole looked fine. The drill looked fine. But the chip was wrong. Instead of those nice tight little figure-eights we usually got in 15-5, it was coming off as dust. Fine, gray, metallic dust.

I figured maybe I pushed it too hard — happened before with a different material where I got greedy on the feed. I grabbed a fresh drill from the same batch and started a new part. Same story. Twenty-five holes, that squeal, dust instead of chips. I slowed the feed down to 0.003 IPR and dropped the RPM to 180. That bought me maybe ten more holes before the squeal came back. I checked the edge under the microscope again. Still looked clean. That was the most frustrating part — I could not see the problem even when I knew there was one.

By noon I had scrapped three parts and burned through four drills. The parts had oversized holes at the bottom — every one of them was pushing 0.3145 at depth, out of print by a mile. The print called for 0.3125 +0.001/ -0.000, so 0.3145 was three and a half thou over the max. Pete’s parts were due Monday morning. I was already behind.

I called my lead hand Tony over and we stood there staring at the last drill under the microscope. The cutting edge looked clean. The web looked centered. The margin looked normal. I have been doing this twenty years and I could not see what was wrong.

Tony said, “Let me check the edge prep.”

He grabbed a brand-new OSG drill out of inventory — same part number, never used — and put it on the comparator next to the reground one. Side by side at 50x, the difference was obvious. The new drill had a sharp, clean cutting edge with maybe 0.001 inch of hone. The regrind had almost no hone at all — the edge was essentially sharp, but inconsistent. In some spots along the lip it looked like the corner radius had been ground off entirely. In others, there was a tiny burr folded over the cutting edge, invisible to the naked eye but clearly there under magnification.

That burr was the culprit. When you drill 15-5 PH at any kind of production rate, edge integrity is everything. A burr on the cutting edge creates localized heat that work-hardens the material right in front of the drill. Then the drill rubs instead of cuts, the heat spikes, and the carbide starts breaking down at the microscopic level. You get oversize holes and dust instead of chips. Classic symptoms.

I called Great Lakes and asked for the shop manager, a guy named Jerry I had spoken to maybe twice in six years. I told him what we were seeing — the dust instead of chips, the oversize holes, the fact that I had scrapped three parts and burned through four drills. He listened without interrupting. Then he put me on hold for five minutes. When he came back, he said, “We had a new guy on the grinders last week. He was running the wheels too fast and not backing off for the finish pass. It has been happening on about fifteen percent of the batch.”

I asked him how long the new guy had been on the grinder. He said three weeks. Three weeks of bad regrinds going out the door, and nobody caught it because the drills passed the visual inspection.

Fifteen percent.

I had twelve drills in my bin. That meant at least a couple of them were probably bad. Maybe more. And we had no way to tell which ones without putting every single one under the microscope.

I spent the rest of Tuesday afternoon and all day Wednesday inspecting every reground drill in the crib. We had seventy-eight drills total from the last Great Lakes shipment. Here is what the inspection found:

Twenty-three of them had the burr issue. That is nearly thirty percent. The new guy had been running the wheel speed at 4500 SFM instead of the 3200 they spec’d for the carbide grade, and he was skipping the finish pass to hit his piece count. Management did not catch it because the drills looked fine to the naked eye, and the packer just looked at the sticker count. No one was looking at the edge under magnification.

I spread the bad drills out on the bench and looked at them under the microscope one more time. Knowing what to look for, the burr was obvious at 50x. But without the new drill sitting next to it for comparison, I would have missed it again. The difference was subtle — a fraction of a thousandth of an inch of folded-over carbide that completely changed how the tool cut.

Thursday morning I drove the bad batch back to Great Lakes myself. Jerry met me at the counter, apologetic, said they would regrind the whole lot free of charge and do 100 percent microscope inspection going forward. I appreciated it, but the damage was done.

Here is the real cost breakdown:

Pete’s parts did not ship until Tuesday. He had to air-freight his assembly to his customer, which ate his margin. He told me he could not trust our delivery anymore and moved the next three repeat orders to a shop in Toledo. I did not blame him. I would have done the same thing in his position. A shop that cannot deliver on time is not a reliable supplier, regardless of the reason.

After that, we started inspecting every reground tool at 50x before it hit the machine. I built a simple go/no-go checklist: edge condition, corner radius, margin wear, coating integrity. The checklist lives on a laminated card taped to the tool crib bench. Takes about thirty seconds per tool. We reject maybe one in fifty now and send them back. Most shops I have talked to think it is overkill. But I would rather spend thirty seconds looking at a drill than three days explaining to a customer why his parts are late.

The inspection station is simple — a 50x microscope with a built-in camera, a comparator for checking geometry, and a good light. I trained every setup guy on what to look for: the edge should have a consistent hone line, the corner radius should be uninterrupted, and there should be no visible burrs folded over the cutting edge at any point along the lip. If any of these fail, the tool does not go into the machine. My article on choosing a gun drill regrind service covers the inspection criteria I now use, and I have also written about the measuring process in more detail in tool wear patterns and what they tell you.

You trust a supplier for six years, and you get complacent. The drills come back looking right, so you assume they are right. But a regrind is only as good as the guy running the wheel that day, and that guy has a production target and a boss breathing down his neck. He does not care about your 15-5 PH job. He cares about his piece count.

It took losing a customer for me to remember that a regrind is not a new tool, and it should not be treated like one. It is a repair. And every repair needs to be verified before you bet a rush order on it. Nineteen years in this trade, and that lesson still cost me a weekend and a customer to learn.

I still think about that Tuesday sometimes. If I had checked the first reground drill under the comparator before loading it into the holder, I would have caught the burr in thirty seconds. I would have saved the parts, saved the weekend, and Pete would still be sending us work. But I had gotten comfortable. The drills came from Great Lakes for six years without a problem, so I stopped looking. That was the real mistake — not the regrind quality issue, but my assumption that I did not need to verify. Regrind inspection is now the first step in every job setup, and it has paid for itself a dozen times over in drills caught before they ever touch a part.

I checked in with Pete about six months after everything settled. He said the shop in Toledo had their own regrind problems six weeks in and he was thinking about coming back. I told him we would be happy to have his work again, but I also told him about our new inspection system. He said, “That is exactly why I am calling you.” He has been back with us for two years now, and every drill that goes into his parts gets checked at 50x before it touches steel. He does not know it, but I check his regrinds first. Some lessons you only need to learn once.

That is the thing about shop-floor stories. The ones that stick with you are not the jobs where everything went right. They are the ones where something went wrong and you figured out why, and you changed the way you work so it does not happen again. I still have a couple of those bad reground drills in a drawer in my toolbox. I pull one out every time I train a new guy on the inspection station. It reminds me — and him — that thirty seconds with a microscope is cheaper than a weekend of rework.