Is Heat Treatment for Metal a Special Process?

[mad223cal]

It IS a special process if you need to evaluate a characteristic that is changed by the process, but cannot be evaluated without some consumption of the product itself. If you have to use NDE (itself a special process) or a sample prepared by exposure to the process to perform the evaluation of the characteristic... then you have a special process.

 

[Caster]

Hmmmm

So you hardened the wear plate?

Perhaps the gate designer made the plate softer than the rollers because it would be easier, faster and cheaper to replace than the rollers?

Have we gone sideways on special processes? It means less than nothing IMO.

I always like to look beyond the words of the standard for the intent.

You need to control time and temperature to heat treat. Period.

Yes, you can measure some part properties after the fact, but if the process is off, it is too late and at best you may need to re heat treat, or at worst you may have made scrap.

Thus a thinking person would make sure that time and temperature we in control, making this into that scary, scary thing called a special process.

Don't even get me started on the art of blacksmithing or Damascus steel, or crystal skulls or ancient batteries, or pyramids made by UFOs, or any old ways were better bunk.

Indiana Jones showed us all how to deal with a sword!

Steel is steel, there are no magical properties in a Samurai sword.

 

[Jim Wynne]

The context is right there in 7.5.2. No word-smithing will get around it. With your train of thought there would never be a need for validation of processes.

No "word-smithing" is necessary, and my train of thought, though it might move slowly, is firmly on the tracks. Let me ask another question, since you avoided the one about the meaning of "subsequent," and didn't acknowledge that any destructive testing (monitoring and measuring) I choose to do will be "subsequent" to production.

Suppose you have a drawing with a specification that says, "Heat treat to x degrees (F) for y hours after machining." That's all she wrote. What do I check? How do I (and why should I) validate the process? Don't think for a minute that specs like that don't exist; I've seen 'em.

You can't say that validation of any process is necessary or required in all cases so long as a single case exists where it's obviously not necessary. And I gladly acknowledge that in the vast majority of cases, validation is necessary. But "vast majority" ≠ "100%."

 

[Big Jim]

No "word-smithing" is necessary, and my train of thought, though it might move slowly, is firmly on the tracks. Let me ask another question, since you avoided the one about the meaning of "subsequent," and didn't acknowledge that any destructive testing (monitoring and measuring) I choose to do will be "subsequent" to production.

Suppose you have a drawing with a specification that says, "Heat treat to x degrees (F) for y hours after machining." That's all she wrote. What do I check? How do I (and why should I) validate the process? Don't think for a minute that specs like that don't exist; I've seen 'em.

You can't say that validation of any process is necessary or required in allcases so long as a single case exists where it's obviously not necessary. And I gladly acknowledge that in the vast majority of cases, validation is necessary. but "vast majority" ≠ "100%."

I'm not going to break the sentence down for you. The context is clear.

Now getting back on track, you may remember that I agreed that in determining if something is a special process "it depends" so I'm not in the school that says there are hard fast rules on what is or is not a "special process". Perhaps better said is the hard and fast rule is as the standard states "The organization shall validate any processes for production and service provision where the resulting output cannot be verified by subsequent monitoring or measurement and, as a consequence, deficiencies become apparent only after the product is in use or the service has been delivered."

So the fact that monitoring or measuring took place is not enough when you still cannot for certain tell the outcome.

I know you know this stuff.

 

[Jim Wynne]

So the fact that monitoring or measuring took place is not enough when you still cannot for certain tell the outcome.

We work with uncertainty all the time. In fact I'll go so far as to say that there is almost no occasion ever when dealing with a large lot of products when we can "for certain tell the outcome" for all of the individuals. The best we can do in most cases is be able to predict with a reasonable amount of certainty how many bad ones there might be. Thus my point about statistically valid sampling plans--why are they considered OK for machining (e.g.) and not for heat treating? In both cases, conscientious process control will almost never account for every possible variable, and the best-laid plans will sometimes go astray.

I'm not against validation; in fact I think we should do more of it because validation is about understanding and controlling variation such that outcomes are reasonably predictable and risks are understood. The great irony to me is that customers, by and large, will accept the idea of validation without subsequent monitoring and measuring for heat treating (and other special processes) but not for non-special processes.

I know you know this stuff.

There's some stuff I know, but there's more stuff to learn, and sometimes I learn it in these conversations.

 

[w_grunfeld]

Suppose you have a drawing with a specification that says, "Heat treat to x degrees (F) for y hours after machining." That's all she wrote. What do I check? How do I (and why should I) validate the process? Don't think for a minute that specs like that don't exist; I've seen 'em. ."

I don't see anything wrong with that spec. All it requires that the piece be heat treated according to a specified recipe. Perhaps it's just for strain relief, or hydrogen de-embritlement or hardening without a specific Rockwell hardness required.
How and why should you validate?
First off, in previous posts you decribed validation as writing a spec,preparing samples, testing the samples, etc.- all implying that you have to validate your product and not the process.(The two shouldn't be mixed up. Validating your product that may or may not incorporate a heat treatment step in its manufacturing process is a totally different issue! )
In practice,companies don't do heat treatment in -house except if the volume justifies having a dedicated department for that. The usual case is that you go to a subcontractor that already has a heat trement process in place that is validated. Remember that the process has to be validated before you use that process on your part.(Likewise if it's in house the validation of the heat treatment process is done once ( ok ocassionally it needs re-verification) on typical parts used as vehicles for the validation of the process) so it's one time (or yearly) expense with just a negigible effect on the price of a specific product
How do you do that? simply , you visit the place, audit if their furnace has the necesary temperature range, control of temperature and time, that the instruments are calibrated and the operators are trained/certified. In addition you would also audit to verify that they have in place adequate material management control and records to ensure they don't mix up different batches of different clients.
I may have skiped a few points , but that's the general idea.
As opposed to what you implied in previous posts, one doesn't have to pay for that when you go to such a certified processor.
Once your parts have passed through a validated process as above,and receive the adequate paper work attesting to it, it's up to the specific application and criticality to decide how much subsequent testing and at what confidence (sampling plan) should be done on the parts.
And btw, I take your advice and next time I audit a samurai sword-smith shop for compliance to ISO9001, I'll consider skipping the validation of the steel hardening provided he can show me that he's finishd his 20 years long apprenticeship. All others- don't count on it.

 

[Jim Wynne]

I don't see anything wrong with that spec. All it requires that the piece be heat treated according to a specified recipe. Perhaps it's just for strain relief, or hydrogen de-embritlement or hardening without a specific Rockwell hardness required.
How and why should you validate?
First off, in previous posts you decribed validation as writing a spec,preparing samples, testing the samples, etc.- all implying that you have to validate your product and not the process.(The two shouldn't be mixed up. Validating your product that may or may not incorporate a heat treatment step in its manufacturing process is a totally different issue! )
In practice,companies don't do heat treatment in -house except if the volume justifies having a dedicated department for that. The usual case is that you go to a subcontractor that already has a heat trement process in place that is validated. Remember that the process has to be validated before you use that process on your part.(Likewise if it's in house the validation of the heat treatment process is done once ( ok ocassionally it needs re-verification) on typical parts used as vehicles for the validation of the process) so it's one time (or yearly) expense with just a negigible effect on the price of a specific product
How do you do that? simply , you visit the place, audit if their furnace has the necesary temperature range, control of temperature and time, that the instruments are calibrated and the operators are trained/certified. In addition you would also audit to verify that they have in place adequate material management control and records to ensure they don't mix up different batches of different clients.
I may have skiped a few points , but that's the general idea.
As opposed to what you implied in previous posts, one doesn't have to pay for that when you go to such a certified processor.
Once your parts have passed through a validated process as above,and receive the adequate paper work attesting to it, it's up to the specific application and criticality to decide how much subsequent testing and at what confidence (sampling plan) should be done on the parts.
And btw, I take your advice and next time I audit a samurai sword-smith shop for compliance to ISO9001, I'll consider skipping the validation of the steel hardening provided he can show me that he's finishd his 20 years long apprenticeship. All others- don't count on it.

I don't think I disagree with anything you've said here, but you've ascribed to me things I didn't say and you don't address anything that I did say.

 

[bobdoering]

You always have the option of asking your customer their interpretation of whether a process that affects their product is "special". Most will be happy to help out with that.

 

[Jim Wynne]

You always have the option of asking your customer their interpretation of whether a process that affects their product is "special". Most will be happy to help out with that.

You should be careful; that sort of thinking can lead to workable solutions.

 

[bobdoering]

You should be careful; that sort of thinking can lead to workable solutions.

My bad. It does assume your customer understands your processes, however.