How to Order the Right Protective Window
Size is the easy part — you measure the old one with a caliper. The two numbers that decide how long the new one lasts are written nowhere on it.
Why This Is the Part You Are Most Likely to Order Wrong
A protective window is the cheapest item on the machine and the one you buy most often. But frequency is not the real reason it goes wrong. It goes wrong because it fails quietly: a disc of the right diameter and thickness in the wrong grade fits, seats and cuts — and for the first few hours it cuts better, because new glass transmits more. No alarm. You find out three weeks later, when it fogs or the sheet stops cutting through.
That delay builds a loop: buy the cheap one, it dies in a fortnight, assume the machine is at fault, buy the cheap one again. And most reorders copy the last invoice, which records what you bought, not what the machine needs — so the first guess is inherited forever, with paperwork that makes it look verified.
One test you can apply to yourself: if your windows die early and unpredictably, don't change suppliers. Count how many things you specified when ordering, besides size.
The Number That Is Not on the Old Part: Your Machine Power
Material sets the power a window can survive. Same diameter, same thickness, different material, different ceiling — and a low-grade window on a high-power machine does not fail honestly. It fogs early, the coating goes, occasionally it burns through mid-job.
So the order must carry a power figure — the machine's rated power, the number on the nameplate. Don't convert it and don't derate it by actual load: trade practice maps grade to rated power, and doing your own arithmetic only makes the supplier guess which convention you used.
The insight here is not how to calculate power. It is where power lives. Diameter and thickness are properties of the part in your hand; power belongs to the machine. An order copied from the old part can, at best, get the size right — the variable that decides survival was never in the room.
Size is on the old part. Power is on the nameplate.
So the order carries three lines: machine power, the material you cut most, and the thickest section. The last two are not for you to convert anything — they let the supplier judge the duty, because the same rated power on thin sheet and on thick plate is not the same load on the optics.
Same Size, Different Grade — the Part You Cannot See
Two windows, same diameter, same thickness, same surface, side by side. One may outlast the other several times over, and looking will not tell you which. Substrate grade, coating process and internal defects all sit below what the eye resolves.
So “it looks the same and costs half” is not a discovery; it is a risk with a friendly description. There is no incoming check on your bench that returns a grade. You can check two things, and neither is the part.
First, whether the supplier can name the grade and say what changes one step down. A one-grade supplier describes his grade as right for everything; a three-grade supplier tells you which one your machine belongs to. Second, service life in your own process — but only in a fair comparison, which is the next section.
Which leaves the uncomfortable conclusion: protective windows have no incoming inspection, only supplier inspection. Everything that decides survival happens in a factory you will never visit.
You cannot inspect a protective window. You can only inspect the supplier.
Do Not Judge a Window by How Long It Lasted
Service life is dominated by conditions, not by the part. The same window on thin mild steel at part power or thick stainless at full power, clean gas path or wet, seated or not, cleaned weekly or never — the difference is a multiple, not a percentage. So “the last batch hazed over in two weeks” cannot convict the window on its own. Rule out four things, in this order.
1. Installation. Seated, and is the seat clean? The window is coated on both faces with no front or back, so “installed backwards” is not a failure mode — that story usually hides something else. Not pressed fully home, or debris trapped underneath, are both real, and both invisible once the cartridge closes.
2. Gas path cleanliness. A spent filter, oil or water carryover, and the best window dies young. This is the most common reason a good part gets blamed for a dirty process.
3. Cutting parameters. Sustained full power on thin material, focus too low, pierce burn-back on thick plate.
4. Cleaning interval and method. Dry wiping, the wrong solvent, a cloth off the floor. A scratch ends a window faster than contamination, and it is the one failure the operator causes personally.
Clear those four and it is still dying early — now the conversation is about the window, and it is fair. Notice three of the four cost less to fix than a grade upgrade: buying up a grade to compensate for a wet gas line buys nothing, because the failure was never in the glass. This section protects both sides — the buyer stops blaming the part for the process, the supplier stops being judged on a number outside his control.
Three Buyers, Three Right Answers
There is no best protective window, only a right one for a use. All three below are legitimate.
|
What you need |
Typical situation |
The right answer |
|
Cheapest that works |
Backup machine, sampling, occasional thin sheet |
Lowest tier. A failure costs one window and a few minutes. |
|
Best available |
Continuous production, thick plate, high power, tight dates |
Top tier. Downtime costs more than the price difference. |
|
Cost per hour |
Most routine production |
Middle tier — usually wins once you divide by hours run. |
What decides your tier is not budget. It is the cost of a failure: one window and twenty minutes means buy cheap; a stopped line and a late order means buy good. Two shops with identical machines can land in different tiers and both be right.
For the middle tier the arithmetic is short: divide price by hours run, and compare. On two-shift routine work it usually wins — not better glass, just a failure interval long enough that the price difference has nowhere to hide. We supply all three, so our recommendation does not depend on what is in stock.
Only a supplier who can sell you all three can give you the fourth answer — “you don't need one that good”. A one-tier supplier always recommends that tier.
What a Good Supplier Does With Your Answers
A good supplier does not take your size, look up a price and reply. Four inputs go in together: machine power, material and thickness, how hard the machine runs, and which tier you asked for. Then one of two answers comes back, and both are professional:
- “For your use you do not need the top grade.” That saves you money.
- “At this duty the cheap one dies faster; per thousand hours it is the expensive choice.” That also saves money.
Which one appears depends on the duty, not on what he wants to sell. A supplier who only ever gives the second answer is running a sales pattern in the clothes of analysis. And watch for the cheaper signal: does a question come back at all? “What are you cutting, and how thick?” costs thirty seconds and prevents a return.
The cheapest window is only a bargain if it fails when you are not busy.
A Five-Line Order Template
Copy these five lines into any inquiry, to any supplier:
- Outside diameter × thickness — measured on the old window with a caliper.
- Machine rated power, and the power you normally run at.
- Main material, and the thickest section you cut.
- How long the last window lasted, and how it failed — fogging, burn-through, scratches, coating lift-off.
- A photo of the old window, or the part number printed on it.
Line four is the one buyers skip and suppliers value most: “hazed over slowly” and “burned through mid-plate” point at different causes. If you don't remember, say so — that is a usable answer. The supplier who answers all five is worth an order. The one who replies with a price and nothing else is selling something he is not certain about either.
Raysoar supplies protective windows in all three tiers, alongside nozzles, lenses and the rest of what a fiber laser consumes.