Heat Treat Radio host Doug Glenn continues his conversation with AMS2750F expert Andrew Bassett. This time the pair discusses Revision F changes to System Accuracy Tests (SATs).
Below, you can either listen to the podcast by clicking on the audio play button, or you can read an edited version of the transcript.
Click the play button below to listen.
The following transcript has been edited for your reading enjoyment.
DG: We are back today for our second episode of a three-part series with Andrew Bassett. Andrew is the president and CEO of Aerospace Testing and Pyrometry, headquartered out of Bethlehem, PA, with offices across the county. They do a lot in pyrometry services and related things. Andrew also had a seat on the committee that was responsible for – that owned – the AMS2750 revision F, so he can speak with firsthand knowledge of some of these changes.
If you are interested, you can listen to the first part, which dealt with the major changes in thermocouples and sensors, major changes in instruments, major changes in calibration, and then we also spent a little bit of time right at the end of the last episode talking about offsets.
AB: Yes, and the offsets were one of major changes that we, as a team, did a very good job of spelling out the new requirements for the two different offsets: modification offsets and correction offsets. So that’s a valuable tool to go back and take a look at.
DG: If you didn’t catch that first episode, you can certainly do that. You can go to heattreattoday.com, jump back into the radio section which is under heat treat media on our main navigation tab, and check that out. It would be very worthwhile.
Before we jump into the topic for today, which is the system accuracy tests (SATs), I wanted to ask you a question about this revision. Often, the AMS folks will come out with a minor modification, or not a huge modification, let’s say; other times, it’s pretty much a re-write, end to end. How would you classify this revision F? Where does it fall on that scale?
AB: It leans towards the side of a complete re-write. I think one of the big things that changed was obviously the number of pages of the document; it jumped from roughly 43 pages up to 54 pages. We expanded the number of tables that were from revision E, which had 11 tables, into 25. This was to do some more clarifications of the requirements, or to spell things out a little bit more. I would be leaning on the side of this as being more of a complete re-write. There’s going to be quite a bit in there that is the same old stuff from the previous revisions, but there is quite a bunch of new stuff.
I would lean towards saying that this was a complete re-write and that’s why there were no change bars associated with the spec. Typically, when these specs get revised, the change bars show you where the changes are, but since this was more of a re-write, we left out the change bars this time around.
DG: Instead of having someone go in and “cheat” and just look at the change bars, you’ve got to pretty much start from the beginning and go straight through.
Where do you see some of the major changes in rev F on the overall or the resident SAT?
AB: Not a whole lot completely changed on the resident sensors. We still allowed for the same sensors as we did in the previous revisions, where you are limited to different types of sensors based on the temperature ranges, that they were going to be seeing. For instance, if you’re above 500 degrees Fahrenheit, then you’re going to be limited to type N, S, R or B thermocouples, and if you’re above 1,000 degrees, they would have to be what’s called a nonexpendable thermocouple, the metal sheathed type thermocouples. We left that stuff alone. But one of the things we did allow for with the new resident sensors, which I believe is a benefit to the supplies that are using the resident sensors, is that we’re going to allow for some things. Let’s say you have an over temperature sensor, and you also want to use that as your resident sensor. Now you’re allowed to do that as long as you follow the guidelines that say a resident sensor has to be replaced. If it’s a base metal thermocouple it has to be replaced every 90 days, or on a quarterly basis. If it is a noble metal, one of the type R, S, or Bs, it would have to be replaced or recalibrated every six months. We did allow for cases where you have an extra sensor that is being used in dual roles (that is, a resident sensor that also functions as a high limit protection), then you can go ahead and do that. I think that that is something that is beneficial to the suppliers, in that we don’t have to go out and put a third sensor into a furnace or drill a hole to put our resident sensor in.
The one thing that we really want to emphasize with these resident sensors is that their position is to be verified during the installation process and when it’s replaced. When a resident sensor is in a fixed position, we want to make sure it is not moving. Typically, you see a compression fitting that is going to tie the thermocouple down and lock it into place. We want to make sure it is not moving between tests. So, now when you replace these things, you must verify the positioning when you put it in on a replacement basis.
Also, it’s always been the requirement to put the thermocouple in for the 90 days or 180 days, and leave it in there. We’re going to allow you to take it out between the tests, but only as long as it is verified after every single time it’s replaced. I’m not a big believer in that; just because someone from Quality doesn’t come out and verify it doesn’t mean that it could be in the wrong position. But we are allowing you to independently move this thing in and out between the test if you want; that is acceptable. You still have the same replacement periods as quarterly and 180 days depending on the sensor type. We did give a little leeway on that from the resident sensor standpoint. Again, we didn’t make a whole lot of changes on it. We just wanted to spell out the little bit of differences allowing for other types of sensors to be used, or have a dual purpose, I should say.
DG: Let’s move on to the second issue, and that is the alternate SAT process, which I know has sparked a lot of questions with the articles we’ve had on our website. We’ve always had people asking about what they can do, what they can’t do. Let’s talk about that.
AB: Sure. The previous revision in rev E was kind of this dark black hole of what the alternate SAT process was all about. Finally, it was more spelled out in what’s called the “PyrometryReference Guide.” That’s the document that NADCAP puts out, the “pyrometry for dummies,” so to speak. This is basically their interpretation of AMS2750. And then kind of evolved that into what’s called a “heat treat audit advisory.” There were different interpretations of this alternate SAT which were too conflicting to the suppliers. We said, “Let’s make it more clear-cut of what the expectation of this alternate SAT process is.”
First off, the process applies to load sensors that are used once, or for any other type of sensor control or recording sensors that are replaced at the same, or less frequent than the normal, SAT intervals. One of the things that was in the previous version, which we kept, is that the calibration must be performed from where you connect the sensor. Then, once you do that calibration, one of the following three options have to be met. Option 1 is that we take the sum of the sensor calibration error. That’s when you first complete calibration from the point of connection and run through the whole system, including the connections, the lead wire, and the instruments. Then, you document those results and algebraically add that to the correction factors or the errors of the wire either being used or replaced more frequently, and if the sum of those two correction factors are within the allowable SAT tolerance of AMS2750, you would have to document that. And that’s the first option; it’s basically a math function; it’s sitting at your desk and taking the calibration report of your process instrumentation, typically from the recording, and adding it to the wire that’s being used. If you fall within that certain table of AMS2750 for SAT tolerances, you’re good to go. It’s kind of a “desk SAT,” as they call it.
The other way of doing this is to use the appropriate sensor and instrument calibration correction factors. You can either program them into the system or apply it manually as allowed by the limits in AMS2750. Basically, you’re taking the correction factors for the instrumentation that you have calibrated and the sensors that you have calibration “certs” on, and programming that into your system. Again, as long as that meets within the applicable table of AMS2750, that is the second option that is allowed. Because you’re basically using the correction values from the calibration reports for your instruments and your thermocouples, you will always be within your SAT requirements.
The third option allows you to do a couple of things. For one, you can limit your instrumentation calibration error. A company comes in and does your calibrations, and the supplier says they don’t want any of their channels to be more than one degree out of calibration, so, you adjust the instrument calibration to be within that limit. Or, you can specify when you purchase thermocouples wire that you won’t take any thermocouple wire that is no more than two degrees out throughout the whole range you need them calibrated. In that instance, you will always be compliant to the requirements of the SAT tolerances. So, if you restrict the calibrations and you restrict the error on the thermocouples, then you will always meet that requirement. All you would have to do is show, for documentation purposes, the instrument calibration reports that say it is all within 1 degree and all of the wire certifications are within two degrees, and that will always meet the most stringent requirement for SAT tolerances. As long as that documentation is there, you will be able to show compliance to the requirement.
[blockquote author=”Andrew Bassett” style=”2″]“Before, there was no requirement of how to document all this, so we actually put in some hard requirements down on how to document the alternate SAT requirements.”[/blockquote]
Those are the more defined options you have. Before, if you gave it to 100 different people to read, and they said, “I don’t know what to do with this information.” Well, now we’ve put out what we actually meant and defined it a little further now.
DG: Great, so that covers the first two that we wanted to talk about – the overall of the resident SAT and now the alternate SAT – so let’s wrap up with this SAT waiver, which is obviously of interest.
AB: First, I want to jump back real quick into the alternate SAT. We finally added some documentation requirements. Before, there was no requirement of how to document all this, so we actually put in some hard requirements down on how to document the alternate SAT requirements. You have to list out the thermal processing equipment (you have to identify which furnace you’re doing this on), what is the sensor system that’s being tested, and what sensor or roll of wire that’s being replaced. You also have to identify the reason why you’re doing the SAT; for example, because you replaced the thermocouple after every run, something simple like that. If you’re doing the full calculation method, then you’d have to show all your calculated methods. We did finally put some teeth in to help you document this well.
DG: Now, the SAT waiver. Tell us about it.
AB: In all my years out in the field of pyrometry, I rarely found many suppliers that did this SAT waiver correctly. We didn’t change a lot of the basics of the requirements, but we did change some new requirements regarding how to gather your data to make sure that you do this correctly. We still require that if you’re using noble metal load thermocouples, which are the platinum based thermocouples, you replace and recalibrate them on a quarterly basis. If you have base metal load thermocouples, if they are expendable, they should still be just a single use. If they’re nonexpendable, sheath type thermocouples, they shall meet the requirements of Table 6 in AMS2750F, and that gives you guidelines of how often those need to be replaced.
If you have any kind of observations that are made and recorded on at least a weekly basis and which reveal any unexplainable difference between observable readings and readings of two recording sensors, this is where the change really occurred on those two additional sensors. We spelled out that these weekly readings have to be conducted at one production setpoint and measured within the five minutes at the end of the production soak period. What this weekly log is supposed to be doing is to compare one sensor against another sensor that you’ve identified.
Some people have used the control sensor as the one sensor and, let’s say, the high limit thermocouple as the second sensor. These have to stay within a two-degree relationship from the last successful survey, and so people were wondering when they were to take the weekly reading. We decided to spell this out a little bit further: this weekly reading must be done at production setpoint and measured within the minutes of the production soak period. In other words, you can let your thermocouples soak out for a period of time, during which you can complete your comparison check. These have to be within two degrees of the relationship determined at the most recent TUS temperature and at the nearest temperature tested during the most recent TUS.
For example, let’s say we do a survey at 1600 degrees and the control is reading 1600 degrees and my over temp is reading 1602. Next week, we come along and we’re running a job here at 1500 degrees and my control is reading 1500 degrees and my over temp is reading 1501, you’re good. You’re within that two-degree relationship. That’s where this two-degree relationship needs to occur.
But the one thing that we’ve done now is we’ve asserted that the two sensors have to be different types. Before, you’d have, let say, two type S thermocouples in your furnace; you can’t have two type S thermocouples now. You have to make a different thermocouple type for the relationship. This is more to catch any drifting of your thermocouples over time. For instance, if you had a type S thermocouple in your furnace as your control, you’re going to have to be limited to either a type B or type N thermocouple as that secondary sensor that you’re doing your relationship check with.
That’s what a big change is. Before people just used the two same sensors. What we were concerned about is – and let’s say those two thermocouples were made from the same lot of material – that there is a good chance that when the thermocouples start to drift, they’re going to drift in the same direction.
Again, we did put some similar restrictions on resident thermocouples. For the example I used, if you had type S control thermocouple, you’d be limited to type B or N, but we also allow for R as that extra thermocouple. But R and S are very similar in the chemical composition makeup, so we don’t allow an S to go against an R and vice versa, in that case. If you had a control thermocouple that was K, then really any other thermocouple that is allowed once you’re above 500 degrees you’re limited to the B, R, S, and N. Actually, these requirements are exactly the resident sensor requirements as well.
DG: Anything else on that SAT waiver?
AB: We do now have some documentation requirements, too. Again, before there were no requirements there. Now you have to list the equipment that you’re doing the waiver on, you have to identify the control sensor, what type of sensor it is, plus what the additional sensor is used for the sensor relationship test. You have to list out the date of when the control and the additional sensor to be used, when they were installed, and when they were replaced or recalibrated. You have to list out the run number and date, so that when you are completing the production cycle on a weekly, you have some kind of easy identifier to tell you that it was done on run #ABC123, and the date was 9/8/20, so we can go back to the records and verify it. Date and temperature of the recent TUS and the documentation, that weekly log, are necessary; we need to see that weekly log as well.
We finally put some teeth into the requirements of the SAT waiver. I don’t think it’s going to be a big change for a lot of the suppliers out there. They will have to change over that one sensor, but, for the most part, I think we tweaked it enough where we felt more comfortable, especially changing those two different sensors so that we didn’t have drift occurring at the same time. That was our biggest concern as a committee.
DG: So, you’re basically trying to ensure reliability and you’re going to actually test for what you’re testing for. That makes sense.
We talked briefly about the overall or resident SAT, the alternate SAT, and the waiver. If you, the listeners, have questions, be sure to email them into us and we can potentially get Andrew to respond to them. Send those to htt@heattreattoday.com. We’ll leave Andrew’s information at the end of each of these podcasts.
Andrew, I’ve got a final question for you, not dealing with any specific aspect of the revision, but just to give people a sense of the amount of time that folks in your shoes, people that have invested time or actually on the committee: How much time do you think you’ve invested in the rev F portion of AMS2750?
AB: It was a long process. To put it in perspective, we developed our sub team and had our first meeting back in October of 2017, during one of the NADCAP meetings. We were kind of on a fast-track to get this spec revised and put out there. It wasn’t actually released until June of 2020; so three year plus is a fast-track in the eyes of the AMS world. We did meet at least six or seven times a year, either during an AMEC meeting or during one of the NADCAP meetings, and we had numerous Webex calls. When we actually met face to face, they were good 8 – 10 hour sessions of hammering out the spec. Then, we would take it back to our own groups and muddle through what we discussed. It was a long period of time. I would hate to put an hour on it. I wish we’d gotten paid for that! Taking into account what our company is and what we do, we have to live, breathe and eat this spec, day in and day out, for our customers. I just wanted to be a part of the process of getting this documentation, so the world can understand the issues in pyrometry.
DG: I actually have one other question for you. You told us in the first episode how you got onto the committee. Are they always looking for people to participate on the committee, or do they carefully fence that and only invite in certain types?
AB: Anybody can be a member of AMEC. So anybody that wants to get involved with the revisions of any of these specifications, including the AMS2750, they’re more than welcome to show up at an AMEC meeting, get involved, and volunteer to get involved with the specifications. I remember my first meeting where the chairman said, “You’ve got to get on this 2750 team. And, oh by the way, we’re thinking about writing some other specs that we’re going to throw you under the bus for.” They’re looking for young blood to get involved with these specifications and be a part of it, so yes, anybody can get involved with these specifications.
DG: If you are listening and you’re one of those people that might be interested in participating in that, you can certainly get a hold of Andrew.
This was our second part in a three part series. Our last episode will be on temperature uniformity surveys, the issue of rounding, and quality assurance provisions. If you’d like to learn more or reach out to Andrew, you can go to www.atp-cal.com and look at their ‘about our team’ section in the main navigation bar. I’d also be happy to receive emails on behalf of Andrew. My email is doug@heattreattoday.com. Thanks for listening.
Doug Glenn,Heat Treat Today publisher and Heat Treat Radio host.
To find other Heat Treat Radio episodes, go to www.heattreattoday.com/radio and look in the list of Heat Treat Radio episodes listed.
